[T6] Advanced methods for neural end-to-end speech processing – unification, integration, and implementation -

PART4: Building End-to-End ASR System

Speaker: Shigeki Karita

NTT Communication Science Laboratories

15, September, 2019

Overview of tutorial

1. Introduction to End-to-End Speech Processing Watanabe

2. End-to-End Integration of Multiple Speech Applications Hori

coffee break

3. Building End-to-End TTS System (45 min.) Hayashi

4. Building End-to-End ASR System (45 min.) Karita

5. Conclusion and Future Research Directions Watanabe

Abstract

How to build end-to-end ASR systems using ESPnet.

materials

TOC

  1. Overview
  2. Installation
  3. Use ESPnet in Bash
  4. Use ESPnet in Python
  5. Extend ESPnet
  6. Summary

1. Overview

ESPnet provides bash recipes and python library for speech processing.

This part demonstrates

  1. bash recipes in ASR
  2. internals of python library

1.1 Python library overview

1.2 Bash recipe overview

ESPnet supports total 34+ ASR tasks including

  • Multilingual ASR: en, zh, ja, etc
  • Noise robust and far-field ASR
  • Multi-ch ASR: joint training with speech enhancement
  • Speech Translation: transfer learning from ASR and MT

For more detail: https://github.com/espnet/espnet/tree/master/egs

1.3 ASR Performance

On free corpora, ESPnet achieved:

  • Aishell (zh): CER test: 6.7%
  • Common Voice (en): WER test: 2.3%
  • LibriSpeech (en): WER test-clean: 2.6%, test-other 5.7%
  • TED-LIUM2 (en): WER test: 8.1%

Pretrained models are available

https://github.com/espnet/espnet#asr-results

2. Installation

ESPnet depends on Kaldi ASR toolkit and Warp-CTC.

You can install them with source compilation:

$ cd espnet/tools; make

2.1 Installation (Google colab)

In Google colab, we can use pre-compiled binaries for faster startup (3 min):

In [0]:
# OS setup
!cat /etc/os-release
!apt-get install -qq bc tree sox

# espnet setup
!git clone --depth 5 https://github.com/espnet/espnet
!pip install -q torch==1.1
!cd espnet; pip install -q -e .

# download pre-compiled warp-ctc and kaldi tools
!espnet/utils/download_from_google_drive.sh \
    "https://drive.google.com/open?id=13Y4tSygc8WtqzvAVGK_vRV9GlV7TRC0w" espnet/tools tar.gz > /dev/null
!cd espnet/tools/warp-ctc/pytorch_binding && \
    pip install -U dist/warpctc_pytorch-0.1.1-cp36-cp36m-linux_x86_64.whl

# make dummy activate
!mkdir -p espnet/tools/venv/bin && touch espnet/tools/venv/bin/activate
!echo "setup done."
NAME="Ubuntu"
VERSION="18.04.3 LTS (Bionic Beaver)"
ID=ubuntu
ID_LIKE=debian
PRETTY_NAME="Ubuntu 18.04.3 LTS"
VERSION_ID="18.04"
HOME_URL="https://www.ubuntu.com/"
SUPPORT_URL="https://help.ubuntu.com/"
BUG_REPORT_URL="https://bugs.launchpad.net/ubuntu/"
PRIVACY_POLICY_URL="https://www.ubuntu.com/legal/terms-and-policies/privacy-policy"
VERSION_CODENAME=bionic
UBUNTU_CODENAME=bionic
Selecting previously unselected package libopencore-amrnb0:amd64.
(Reading database ... 131183 files and directories currently installed.)
Preparing to unpack .../0-libopencore-amrnb0_0.1.3-2.1_amd64.deb ...
Unpacking libopencore-amrnb0:amd64 (0.1.3-2.1) ...
Selecting previously unselected package libopencore-amrwb0:amd64.
Preparing to unpack .../1-libopencore-amrwb0_0.1.3-2.1_amd64.deb ...
Unpacking libopencore-amrwb0:amd64 (0.1.3-2.1) ...
Selecting previously unselected package libmagic-mgc.
Preparing to unpack .../2-libmagic-mgc_1%3a5.32-2ubuntu0.2_amd64.deb ...
Unpacking libmagic-mgc (1:5.32-2ubuntu0.2) ...
Selecting previously unselected package libmagic1:amd64.
Preparing to unpack .../3-libmagic1_1%3a5.32-2ubuntu0.2_amd64.deb ...
Unpacking libmagic1:amd64 (1:5.32-2ubuntu0.2) ...
Selecting previously unselected package bc.
Preparing to unpack .../4-bc_1.07.1-2_amd64.deb ...
Unpacking bc (1.07.1-2) ...
Selecting previously unselected package libsox3:amd64.
Preparing to unpack .../5-libsox3_14.4.2-3ubuntu0.18.04.1_amd64.deb ...
Unpacking libsox3:amd64 (14.4.2-3ubuntu0.18.04.1) ...
Selecting previously unselected package libsox-fmt-alsa:amd64.
Preparing to unpack .../6-libsox-fmt-alsa_14.4.2-3ubuntu0.18.04.1_amd64.deb ...
Unpacking libsox-fmt-alsa:amd64 (14.4.2-3ubuntu0.18.04.1) ...
Selecting previously unselected package libsox-fmt-base:amd64.
Preparing to unpack .../7-libsox-fmt-base_14.4.2-3ubuntu0.18.04.1_amd64.deb ...
Unpacking libsox-fmt-base:amd64 (14.4.2-3ubuntu0.18.04.1) ...
Selecting previously unselected package sox.
Preparing to unpack .../8-sox_14.4.2-3ubuntu0.18.04.1_amd64.deb ...
Unpacking sox (14.4.2-3ubuntu0.18.04.1) ...
Selecting previously unselected package tree.
Preparing to unpack .../9-tree_1.7.0-5_amd64.deb ...
Unpacking tree (1.7.0-5) ...
Setting up tree (1.7.0-5) ...
Processing triggers for mime-support (3.60ubuntu1) ...
Setting up libmagic-mgc (1:5.32-2ubuntu0.2) ...
Setting up libmagic1:amd64 (1:5.32-2ubuntu0.2) ...
Processing triggers for libc-bin (2.27-3ubuntu1) ...
Setting up libopencore-amrnb0:amd64 (0.1.3-2.1) ...
Processing triggers for man-db (2.8.3-2ubuntu0.1) ...
Setting up bc (1.07.1-2) ...
Setting up libopencore-amrwb0:amd64 (0.1.3-2.1) ...
Setting up libsox3:amd64 (14.4.2-3ubuntu0.18.04.1) ...
Setting up libsox-fmt-base:amd64 (14.4.2-3ubuntu0.18.04.1) ...
Setting up libsox-fmt-alsa:amd64 (14.4.2-3ubuntu0.18.04.1) ...
Setting up sox (14.4.2-3ubuntu0.18.04.1) ...
Processing triggers for libc-bin (2.27-3ubuntu1) ...
Cloning into 'espnet'...
remote: Enumerating objects: 1287, done.
remote: Counting objects: 100% (1287/1287), done.
remote: Compressing objects: 100% (997/997), done.
remote: Total 1287 (delta 404), reused 619 (delta 164), pack-reused 0
Receiving objects: 100% (1287/1287), 1.67 MiB | 9.16 MiB/s, done.
Resolving deltas: 100% (404/404), done.
     |████████████████████████████████| 890kB 3.5MB/s 
     |████████████████████████████████| 1.6MB 46.3MB/s 
     |████████████████████████████████| 245kB 40.2MB/s 
     |████████████████████████████████| 174kB 45.6MB/s 
     |████████████████████████████████| 2.8MB 28.5MB/s 
     |████████████████████████████████| 225kB 32.7MB/s 
     |████████████████████████████████| 2.1MB 33.8MB/s 
     |████████████████████████████████| 13.1MB 32.5MB/s 
     |████████████████████████████████| 266kB 42.6MB/s 
     |████████████████████████████████| 1.0MB 36.5MB/s 
     |████████████████████████████████| 419kB 32.3MB/s 
     |████████████████████████████████| 10.5MB 31.2MB/s 
     |████████████████████████████████| 81kB 26.8MB/s 
     |████████████████████████████████| 512kB 42.0MB/s 
  Building wheel for chainer (setup.py) ... done
  Building wheel for librosa (setup.py) ... done
  Building wheel for kaldiio (setup.py) ... done
  Building wheel for configargparse (setup.py) ... done
  Building wheel for PyYAML (setup.py) ... done
  Building wheel for pysptk (setup.py) ... done
  Building wheel for torch-complex (setup.py) ... done
  Building wheel for pytorch-wpe (setup.py) ... done
  Building wheel for simplejson (setup.py) ... done
ERROR: pandas 0.25.1 has requirement python-dateutil>=2.6.1, but you'll have python-dateutil 2.5.3 which is incompatible.
ERROR: google-colab 1.0.0 has requirement pandas~=0.24.0, but you'll have pandas 0.25.1 which is incompatible.
ERROR: albumentations 0.1.12 has requirement imgaug<0.2.7,>=0.2.5, but you'll have imgaug 0.2.9 which is incompatible.
--2019-09-11 11:05:47--  https://drive.google.com/uc?export=download&id=13Y4tSygc8WtqzvAVGK_vRV9GlV7TRC0w
Resolving drive.google.com (drive.google.com)... 172.217.214.139, 172.217.214.102, 172.217.214.113, ...
Connecting to drive.google.com (drive.google.com)|172.217.214.139|:443... connected.
HTTP request sent, awaiting response... 200 OK
Length: unspecified [text/html]
Saving to: ‘espnet/tools/MUBK0o.tar.gz’

espnet/tools/MUBK0o     [ <=>                ]   3.22K  --.-KB/s    in 0s      

2019-09-11 11:05:48 (34.0 MB/s) - ‘espnet/tools/MUBK0o.tar.gz’ saved [3296]


gzip: stdin: not in gzip format
tar: Child returned status 1
tar: Error is not recoverable: exiting now
  % Total    % Received % Xferd  Average Speed   Time    Time     Time  Current
                                 Dload  Upload   Total   Spent    Left  Speed
100  3296    0  3296    0     0  26580      0 --:--:-- --:--:-- --:--:-- 26796
  % Total    % Received % Xferd  Average Speed   Time    Time     Time  Current
                                 Dload  Upload   Total   Spent    Left  Speed
100   388    0   388    0     0   3559      0 --:--:-- --:--:-- --:--:--  3592
100  662M    0  662M    0     0  81.2M      0 --:--:--  0:00:08 --:--:-- 92.6M
Processing ./dist/warpctc_pytorch-0.1.1-cp36-cp36m-linux_x86_64.whl
Installing collected packages: warpctc-pytorch
Successfully installed warpctc-pytorch-0.1.1
setup done.

3. Use ESPnet in Bash

  • espnet/egs/*/asr1/run.sh is an out-of-the-box recipe
  • It reproduces our reported results
  • It consists of common stages similar to our TTS recipe: image.png

3.1 Kaldi-style directory structure

Always we organize each recipe as egs/xxx/asr1/run.sh

The most important directories:

  • conf/: configurations for stages and machines (e.g., Local, SLURM, SGE)
  • data/: raw data prepared by Kaldi in Stage 0 - 1
  • dump/: python-friendly dataset (e.g., json, hdf5) in Stage 2
  • exp/: log files and saved model parameters in Stage 3 - 5
In [13]:
!tree -L 1 espnet/egs/librispeech/asr1
espnet/egs/librispeech/asr1
├── cmd.sh
├── conf
├── local
├── path.sh
├── RESULTS.md
├── run.sh
├── steps -> ../../../tools/kaldi/egs/wsj/s5/steps
└── utils -> ../../../tools/kaldi/egs/wsj/s5/utils

4 directories, 4 files

3.2 Data preparation (Stage 0 - 2)

For example, if you add --stop-stage 2, you can stop the script before neural network training.

These stages perform FBANK speech feature extraction, normalization, and text formatting as same as TTS recipe.

image.png

In [14]:
!cd espnet/egs/an4/asr1; ./run.sh  --ngpu 1 --stop-stage 2
stage -1: Data Download
local/download_and_untar.sh: an4 directory already exists in ./downloads
stage 0: Data preparation
stage 1: Feature Generation
steps/make_fbank_pitch.sh --cmd stdout.pl --nj 8 --write_utt2num_frames true data/test exp/make_fbank/test fbank
steps/make_fbank_pitch.sh: moving data/test/feats.scp to data/test/.backup
utils/validate_data_dir.sh: Successfully validated data-directory data/test
steps/make_fbank_pitch.sh: [info]: no segments file exists: assuming wav.scp indexed by utterance.
copy-feats --compress=true --write-num-frames=ark,t:exp/make_fbank/test/utt2num_frames.2 ark:- ark,scp:/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/fbank/raw_fbank_pitch_test.2.ark,/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/fbank/raw_fbank_pitch_test.2.scp 
paste-feats --length-tolerance=2 'ark:compute-fbank-feats  --verbose=2 --config=conf/fbank.conf scp,p:exp/make_fbank/test/wav.1.scp ark:- |' 'ark,s,cs:compute-kaldi-pitch-feats --verbose=2 --config=conf/pitch.conf scp,p:exp/make_fbank/test/wav.1.scp ark:- | process-kaldi-pitch-feats  ark:- ark:- |' ark:- 
paste-feats --length-tolerance=2 'ark:compute-fbank-feats  --verbose=2 --config=conf/fbank.conf scp,p:exp/make_fbank/test/wav.4.scp ark:- |' 'ark,s,cs:compute-kaldi-pitch-feats --verbose=2 --config=conf/pitch.conf scp,p:exp/make_fbank/test/wav.4.scp ark:- | process-kaldi-pitch-feats  ark:- ark:- |' ark:- 
copy-feats --compress=true --write-num-frames=ark,t:exp/make_fbank/test/utt2num_frames.4 ark:- ark,scp:/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/fbank/raw_fbank_pitch_test.4.ark,/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/fbank/raw_fbank_pitch_test.4.scp 
paste-feats --length-tolerance=2 'ark:compute-fbank-feats  --verbose=2 --config=conf/fbank.conf scp,p:exp/make_fbank/test/wav.2.scp ark:- |' 'ark,s,cs:compute-kaldi-pitch-feats --verbose=2 --config=conf/pitch.conf scp,p:exp/make_fbank/test/wav.2.scp ark:- | process-kaldi-pitch-feats  ark:- ark:- |' ark:- 
copy-feats --compress=true --write-num-frames=ark,t:exp/make_fbank/test/utt2num_frames.3 ark:- ark,scp:/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/fbank/raw_fbank_pitch_test.3.ark,/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/fbank/raw_fbank_pitch_test.3.scp 
paste-feats --length-tolerance=2 'ark:compute-fbank-feats  --verbose=2 --config=conf/fbank.conf scp,p:exp/make_fbank/test/wav.3.scp ark:- |' 'ark,s,cs:compute-kaldi-pitch-feats --verbose=2 --config=conf/pitch.conf scp,p:exp/make_fbank/test/wav.3.scp ark:- | process-kaldi-pitch-feats  ark:- ark:- |' ark:- 
copy-feats --compress=true --write-num-frames=ark,t:exp/make_fbank/test/utt2num_frames.1 ark:- ark,scp:/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/fbank/raw_fbank_pitch_test.1.ark,/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/fbank/raw_fbank_pitch_test.1.scp 
compute-fbank-feats --verbose=2 --config=conf/fbank.conf scp,p:exp/make_fbank/test/wav.3.scp ark:- 
compute-fbank-feats --verbose=2 --config=conf/fbank.conf scp,p:exp/make_fbank/test/wav.1.scp ark:- 
compute-fbank-feats --verbose=2 --config=conf/fbank.conf scp,p:exp/make_fbank/test/wav.4.scp ark:- 
compute-fbank-feats --verbose=2 --config=conf/fbank.conf scp,p:exp/make_fbank/test/wav.2.scp ark:- 
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key marh-cen7-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fvap-cen4-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fcaw-an406-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fjlp-an420-b
compute-kaldi-pitch-feats --verbose=2 --config=conf/pitch.conf scp,p:exp/make_fbank/test/wav.3.scp ark:- 
process-kaldi-pitch-feats ark:- ark:- 
compute-kaldi-pitch-feats --verbose=2 --config=conf/pitch.conf scp,p:exp/make_fbank/test/wav.4.scp ark:- 
process-kaldi-pitch-feats ark:- ark:- 
process-kaldi-pitch-feats ark:- ark:- 
compute-kaldi-pitch-feats --verbose=2 --config=conf/pitch.conf scp,p:exp/make_fbank/test/wav.2.scp ark:- 
process-kaldi-pitch-feats ark:- ark:- 
compute-kaldi-pitch-feats --verbose=2 --config=conf/pitch.conf scp,p:exp/make_fbank/test/wav.1.scp ark:- 
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key marh-cen8-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fvap-cen5-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fcaw-an407-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fvap-cen6-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fcaw-an408-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fcaw-an409-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fjlp-cen1-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fvap-cen7-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fjlp-cen2-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mdms2-an401-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fvap-cen8-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fcaw-an410-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fjlp-cen3-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key marh-an431-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mdms2-an402-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fcaw-cen1-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mdms2-an403-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fcaw-cen2-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fjlp-cen4-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mdms2-an404-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key marh-an432-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fcaw-cen3-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key marh-an433-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mdms2-an405-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fcaw-cen4-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mdms2-cen1-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fjlp-cen5-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fjlp-cen6-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key marh-an434-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:169) Processed 10 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key marh-an435-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:169) Processed 10 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fcaw-cen5-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fjlp-cen7-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key marh-cen1-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fjlp-cen8-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key marh-cen2-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:169) Processed 10 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fvap-an426-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fcaw-cen6-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key marh-cen3-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mdms2-cen2-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fvap-an427-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fcaw-cen7-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:169) Processed 10 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mdms2-cen3-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fcaw-cen8-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fvap-an428-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mdms2-cen4-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key marh-cen4-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fjlp-an416-b
LOG (compute-kaldi-pitch-feats[5.5.284-76bd]:main():compute-kaldi-pitch-feats.cc:110) Done 16 utterances, 0 with errors.
LOG (process-kaldi-pitch-feats[5.5.284-76bd]:main():process-kaldi-pitch-feats.cc:85) Post-processed pitch for 16 utterances.
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key marh-cen5-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fvap-an429-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mdms2-cen5-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mdms2-cen6-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key marh-cen6-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:173)  Done 16 out of 16 utterances.
LOG (paste-feats[5.5.284-76bd]:main():paste-feats.cc:158) Done 16 utts, errors on 0
LOG (copy-feats[5.5.284-76bd]:main():copy-feats.cc:143) Copied 16 feature matrices.
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fvap-an430-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fjlp-an417-b
LOG (compute-kaldi-pitch-feats[5.5.284-76bd]:main():compute-kaldi-pitch-feats.cc:110) Done 16 utterances, 0 with errors.
LOG (process-kaldi-pitch-feats[5.5.284-76bd]:main():process-kaldi-pitch-feats.cc:85) Post-processed pitch for 16 utterances.
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fjlp-an418-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mdms2-cen7-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fvap-cen1-b
LOG (compute-kaldi-pitch-feats[5.5.284-76bd]:main():compute-kaldi-pitch-feats.cc:110) Done 17 utterances, 0 with errors.
LOG (process-kaldi-pitch-feats[5.5.284-76bd]:main():process-kaldi-pitch-feats.cc:85) Post-processed pitch for 17 utterances.
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fjlp-an419-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:173)  Done 17 out of 17 utterances.
LOG (paste-feats[5.5.284-76bd]:main():paste-feats.cc:158) Done 17 utts, errors on 0
LOG (copy-feats[5.5.284-76bd]:main():copy-feats.cc:143) Copied 17 feature matrices.
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mdms2-cen8-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fvap-cen2-b
LOG (compute-kaldi-pitch-feats[5.5.284-76bd]:main():compute-kaldi-pitch-feats.cc:110) Done 17 utterances, 0 with errors.
LOG (process-kaldi-pitch-feats[5.5.284-76bd]:main():process-kaldi-pitch-feats.cc:85) Post-processed pitch for 17 utterances.
paste-feats --length-tolerance=2 'ark:compute-fbank-feats  --verbose=2 --config=conf/fbank.conf scp,p:exp/make_fbank/test/wav.5.scp ark:- |' 'ark,s,cs:compute-kaldi-pitch-feats --verbose=2 --config=conf/pitch.conf scp,p:exp/make_fbank/test/wav.5.scp ark:- | process-kaldi-pitch-feats  ark:- ark:- |' ark:- 
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key menk-an421-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:173)  Done 16 out of 16 utterances.
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fvap-cen3-b
LOG (paste-feats[5.5.284-76bd]:main():paste-feats.cc:158) Done 16 utts, errors on 0
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:173)  Done 17 out of 17 utterances.
LOG (paste-feats[5.5.284-76bd]:main():paste-feats.cc:158) Done 17 utts, errors on 0
LOG (copy-feats[5.5.284-76bd]:main():copy-feats.cc:143) Copied 16 feature matrices.
LOG (copy-feats[5.5.284-76bd]:main():copy-feats.cc:143) Copied 17 feature matrices.
copy-feats --compress=true --write-num-frames=ark,t:exp/make_fbank/test/utt2num_frames.5 ark:- ark,scp:/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/fbank/raw_fbank_pitch_test.5.ark,/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/fbank/raw_fbank_pitch_test.5.scp 
compute-fbank-feats --verbose=2 --config=conf/fbank.conf scp,p:exp/make_fbank/test/wav.5.scp ark:- 
copy-feats --compress=true --write-num-frames=ark,t:exp/make_fbank/test/utt2num_frames.6 ark:- ark,scp:/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/fbank/raw_fbank_pitch_test.6.ark,/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/fbank/raw_fbank_pitch_test.6.scp 
paste-feats --length-tolerance=2 'ark:compute-fbank-feats  --verbose=2 --config=conf/fbank.conf scp,p:exp/make_fbank/test/wav.6.scp ark:- |' 'ark,s,cs:compute-kaldi-pitch-feats --verbose=2 --config=conf/pitch.conf scp,p:exp/make_fbank/test/wav.6.scp ark:- | process-kaldi-pitch-feats  ark:- ark:- |' ark:- 
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key menk-an422-b
compute-fbank-feats --verbose=2 --config=conf/fbank.conf scp,p:exp/make_fbank/test/wav.6.scp ark:- 
process-kaldi-pitch-feats ark:- ark:- 
copy-feats --compress=true --write-num-frames=ark,t:exp/make_fbank/test/utt2num_frames.8 ark:- ark,scp:/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/fbank/raw_fbank_pitch_test.8.ark,/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/fbank/raw_fbank_pitch_test.8.scp 
compute-kaldi-pitch-feats --verbose=2 --config=conf/pitch.conf scp,p:exp/make_fbank/test/wav.5.scp ark:- 
paste-feats --length-tolerance=2 'ark:compute-fbank-feats  --verbose=2 --config=conf/fbank.conf scp,p:exp/make_fbank/test/wav.8.scp ark:- |' 'ark,s,cs:compute-kaldi-pitch-feats --verbose=2 --config=conf/pitch.conf scp,p:exp/make_fbank/test/wav.8.scp ark:- | process-kaldi-pitch-feats  ark:- ark:- |' ark:- 
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key miry-an400-b
compute-fbank-feats --verbose=2 --config=conf/fbank.conf scp,p:exp/make_fbank/test/wav.8.scp ark:- 
paste-feats --length-tolerance=2 'ark:compute-fbank-feats  --verbose=2 --config=conf/fbank.conf scp,p:exp/make_fbank/test/wav.7.scp ark:- |' 'ark,s,cs:compute-kaldi-pitch-feats --verbose=2 --config=conf/pitch.conf scp,p:exp/make_fbank/test/wav.7.scp ark:- | process-kaldi-pitch-feats  ark:- ark:- |' ark:- 
process-kaldi-pitch-feats ark:- ark:- 
compute-kaldi-pitch-feats --verbose=2 --config=conf/pitch.conf scp,p:exp/make_fbank/test/wav.6.scp ark:- 
copy-feats --compress=true --write-num-frames=ark,t:exp/make_fbank/test/utt2num_frames.7 ark:- ark,scp:/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/fbank/raw_fbank_pitch_test.7.ark,/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/fbank/raw_fbank_pitch_test.7.scp 
compute-fbank-feats --verbose=2 --config=conf/fbank.conf scp,p:exp/make_fbank/test/wav.7.scp ark:- 
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjwl-cen6-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjgm-cen3-b
process-kaldi-pitch-feats ark:- ark:- 
compute-kaldi-pitch-feats --verbose=2 --config=conf/pitch.conf scp,p:exp/make_fbank/test/wav.8.scp ark:- 
process-kaldi-pitch-feats ark:- ark:- 
compute-kaldi-pitch-feats --verbose=2 --config=conf/pitch.conf scp,p:exp/make_fbank/test/wav.7.scp ark:- 
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key menk-an423-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key miry-cen1-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key miry-cen2-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjwl-cen7-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key menk-an424-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjgm-cen4-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjwl-cen8-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key miry-cen3-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key menk-an425-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmxg-an441-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjgm-cen5-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmxg-an442-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmxg-an443-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key menk-cen1-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key miry-cen4-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjgm-cen6-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjgm-cen7-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key miry-cen5-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key menk-cen2-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmxg-an444-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key miry-cen6-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key menk-cen3-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjgm-cen8-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmxg-an445-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key miry-cen7-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjwl-an391-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjwl-an392-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmxg-cen1-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key menk-cen4-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key miry-cen8-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:169) Processed 10 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmxg-cen2-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:169) Processed 10 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjgm-an436-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key menk-cen5-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmxg-cen3-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjwl-an393-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjgm-an437-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:169) Processed 10 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key menk-cen6-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:169) Processed 10 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjwl-an394-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjgm-an438-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmxg-cen4-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjwl-an395-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjgm-an439-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjwl-cen1-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key menk-cen7-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjgm-an440-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjwl-cen2-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key menk-cen8-b
LOG (compute-kaldi-pitch-feats[5.5.284-76bd]:main():compute-kaldi-pitch-feats.cc:110) Done 16 utterances, 0 with errors.
LOG (process-kaldi-pitch-feats[5.5.284-76bd]:main():process-kaldi-pitch-feats.cc:85) Post-processed pitch for 16 utterances.
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjgm-cen1-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjwl-cen3-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjgm-cen2-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:173)  Done 16 out of 16 utterances.
LOG (paste-feats[5.5.284-76bd]:main():paste-feats.cc:158) Done 16 utts, errors on 0
LOG (copy-feats[5.5.284-76bd]:main():copy-feats.cc:143) Copied 16 feature matrices.
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key miry-an396-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmxg-cen5-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key miry-an397-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjwl-cen4-b
LOG (compute-kaldi-pitch-feats[5.5.284-76bd]:main():compute-kaldi-pitch-feats.cc:110) Done 16 utterances, 0 with errors.
LOG (process-kaldi-pitch-feats[5.5.284-76bd]:main():process-kaldi-pitch-feats.cc:85) Post-processed pitch for 16 utterances.
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmxg-cen6-b
LOG (compute-kaldi-pitch-feats[5.5.284-76bd]:main():compute-kaldi-pitch-feats.cc:110) Done 16 utterances, 0 with errors.
LOG (process-kaldi-pitch-feats[5.5.284-76bd]:main():process-kaldi-pitch-feats.cc:85) Post-processed pitch for 16 utterances.
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjwl-cen5-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:173)  Done 16 out of 16 utterances.
LOG (paste-feats[5.5.284-76bd]:main():paste-feats.cc:158) Done 16 utts, errors on 0
LOG (copy-feats[5.5.284-76bd]:main():copy-feats.cc:143) Copied 16 feature matrices.
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key miry-an398-b
LOG (compute-kaldi-pitch-feats[5.5.284-76bd]:main():compute-kaldi-pitch-feats.cc:110) Done 16 utterances, 0 with errors.
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmxg-cen7-b
LOG (process-kaldi-pitch-feats[5.5.284-76bd]:main():process-kaldi-pitch-feats.cc:85) Post-processed pitch for 16 utterances.
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmxg-cen8-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:173)  Done 16 out of 16 utterances.
LOG (paste-feats[5.5.284-76bd]:main():paste-feats.cc:158) Done 16 utts, errors on 0
LOG (copy-feats[5.5.284-76bd]:main():copy-feats.cc:143) Copied 16 feature matrices.
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key miry-an399-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:173)  Done 16 out of 16 utterances.
LOG (paste-feats[5.5.284-76bd]:main():paste-feats.cc:158) Done 16 utts, errors on 0
LOG (copy-feats[5.5.284-76bd]:main():copy-feats.cc:143) Copied 16 feature matrices.
Succeeded creating filterbank & pitch features for test
fix_data_dir.sh: kept all 130 utterances.
fix_data_dir.sh: old files are kept in data/test/.backup
steps/make_fbank_pitch.sh --cmd stdout.pl --nj 8 --write_utt2num_frames true data/train exp/make_fbank/train fbank
steps/make_fbank_pitch.sh: moving data/train/feats.scp to data/train/.backup
utils/validate_data_dir.sh: Successfully validated data-directory data/train
steps/make_fbank_pitch.sh: [info]: no segments file exists: assuming wav.scp indexed by utterance.
paste-feats --length-tolerance=2 'ark:compute-fbank-feats  --verbose=2 --config=conf/fbank.conf scp,p:exp/make_fbank/train/wav.2.scp ark:- |' 'ark,s,cs:compute-kaldi-pitch-feats --verbose=2 --config=conf/pitch.conf scp,p:exp/make_fbank/train/wav.2.scp ark:- | process-kaldi-pitch-feats  ark:- ark:- |' ark:- 
paste-feats --length-tolerance=2 'ark:compute-fbank-feats  --verbose=2 --config=conf/fbank.conf scp,p:exp/make_fbank/train/wav.4.scp ark:- |' 'ark,s,cs:compute-kaldi-pitch-feats --verbose=2 --config=conf/pitch.conf scp,p:exp/make_fbank/train/wav.4.scp ark:- | process-kaldi-pitch-feats  ark:- ark:- |' ark:- 
copy-feats --compress=true --write-num-frames=ark,t:exp/make_fbank/train/utt2num_frames.4 ark:- ark,scp:/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/fbank/raw_fbank_pitch_train.4.ark,/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/fbank/raw_fbank_pitch_train.4.scp 
copy-feats --compress=true --write-num-frames=ark,t:exp/make_fbank/train/utt2num_frames.2 ark:- ark,scp:/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/fbank/raw_fbank_pitch_train.2.ark,/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/fbank/raw_fbank_pitch_train.2.scp 
paste-feats --length-tolerance=2 'ark:compute-fbank-feats  --verbose=2 --config=conf/fbank.conf scp,p:exp/make_fbank/train/wav.3.scp ark:- |' 'ark,s,cs:compute-kaldi-pitch-feats --verbose=2 --config=conf/pitch.conf scp,p:exp/make_fbank/train/wav.3.scp ark:- | process-kaldi-pitch-feats  ark:- ark:- |' ark:- 
paste-feats --length-tolerance=2 'ark:compute-fbank-feats  --verbose=2 --config=conf/fbank.conf scp,p:exp/make_fbank/train/wav.1.scp ark:- |' 'ark,s,cs:compute-kaldi-pitch-feats --verbose=2 --config=conf/pitch.conf scp,p:exp/make_fbank/train/wav.1.scp ark:- | process-kaldi-pitch-feats  ark:- ark:- |' ark:- 
copy-feats --compress=true --write-num-frames=ark,t:exp/make_fbank/train/utt2num_frames.1 ark:- ark,scp:/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/fbank/raw_fbank_pitch_train.1.ark,/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/fbank/raw_fbank_pitch_train.1.scp 
compute-fbank-feats --verbose=2 --config=conf/fbank.conf scp,p:exp/make_fbank/train/wav.1.scp ark:- 
copy-feats --compress=true --write-num-frames=ark,t:exp/make_fbank/train/utt2num_frames.3 ark:- ark,scp:/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/fbank/raw_fbank_pitch_train.3.ark,/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/fbank/raw_fbank_pitch_train.3.scp 
compute-fbank-feats --verbose=2 --config=conf/fbank.conf scp,p:exp/make_fbank/train/wav.4.scp ark:- 
compute-fbank-feats --verbose=2 --config=conf/fbank.conf scp,p:exp/make_fbank/train/wav.3.scp ark:- 
compute-fbank-feats --verbose=2 --config=conf/fbank.conf scp,p:exp/make_fbank/train/wav.2.scp ark:- 
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fash-an251-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key ftal-cen4-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mcrt-cen8-b
compute-kaldi-pitch-feats --verbose=2 --config=conf/pitch.conf scp,p:exp/make_fbank/train/wav.1.scp ark:- 
process-kaldi-pitch-feats ark:- ark:- 
process-kaldi-pitch-feats ark:- ark:- 
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fkdo-cen2-b
process-kaldi-pitch-feats ark:- ark:- 
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fash-an253-b
compute-kaldi-pitch-feats --verbose=2 --config=conf/pitch.conf scp,p:exp/make_fbank/train/wav.4.scp ark:- 
compute-kaldi-pitch-feats --verbose=2 --config=conf/pitch.conf scp,p:exp/make_fbank/train/wav.3.scp ark:- 
compute-kaldi-pitch-feats --verbose=2 --config=conf/pitch.conf scp,p:exp/make_fbank/train/wav.2.scp ark:- 
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fash-an254-b
process-kaldi-pitch-feats ark:- ark:- 
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fkdo-cen3-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mcsc-an231-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fash-an255-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mcsc-an232-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mcsc-an233-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key ftal-cen5-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fkdo-cen4-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key ftal-cen6-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key ftal-cen7-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fkdo-cen5-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fash-cen1-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key ftal-cen8-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fash-cen2-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mcsc-an234-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fkdo-cen6-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mcsc-an235-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fash-cen4-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fkdo-cen7-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key ftmj-an211-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fkdo-cen8-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mcsc-cen1-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fash-cen5-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mcsc-cen2-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fash-cen7-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key ftmj-an212-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:169) Processed 10 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fbbh-an86-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mcsc-cen3-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fbbh-an87-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:169) Processed 10 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mcsc-cen4-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key flmm2-an61-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mcsc-cen5-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key ftmj-an213-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key flmm2-an62-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:169) Processed 10 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key flmm2-an63-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mcsc-cen6-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key ftmj-an214-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:169) Processed 10 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key ftmj-an215-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key flmm2-an64-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key ftmj-cen1-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mcsc-cen7-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fbbh-an88-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key flmm2-an65-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key ftmj-cen2-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mcsc-cen8-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fbbh-an89-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fbbh-an90-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key flmm2-cen1-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key ftmj-cen3-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mdcs-an241-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key flmm2-cen2-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mdcs-an242-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key flmm2-cen3-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key ftmj-cen4-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fbbh-cen1-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mdcs-an243-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key ftmj-cen5-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fbbh-cen2-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key ftmj-cen6-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key flmm2-cen4-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key ftmj-cen7-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mdcs-an244-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fbbh-cen3-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key flmm2-cen6-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mdcs-an245-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key ftmj-cen8-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fbbh-cen4-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fwxs-an221-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:169) Processed 20 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mdcs-cen1-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fbbh-cen5-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mdcs-cen2-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:169) Processed 20 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fwxs-an222-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:169) Processed 20 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fbbh-cen6-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fwxs-an223-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fwxs-an224-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key flmm2-cen7-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fbbh-cen7-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fwxs-an225-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fbbh-cen8-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key flmm2-cen8-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mdcs-cen3-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fwxs-cen1-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:169) Processed 20 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key flrp-an21-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fclc-an146-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mdcs-cen4-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fclc-an147-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fwxs-cen2-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mdcs-cen5-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fwxs-cen3-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key flrp-an2121-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key flrp-an22-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mdcs-cen6-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fwxs-cen4-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mdcs-cen7-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fwxs-cen5-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mdcs-cen8-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key flrp-an23-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mdcs2-an111-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fclc-an148-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fwxs-cen6-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key flrp-an24-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fclc-an149-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mdcs2-an112-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:169) Processed 30 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fwxs-cen7-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key flrp-an25-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fclc-an150-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key flrp-cen1-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mblb-an371-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key flrp-cen2-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fclc-cen1-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:169) Processed 30 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mdcs2-an113-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fclc-cen2-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:169) Processed 30 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fclc-cen3-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key flrp-cen3-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mdcs2-an114-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mblb-an372-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fclc-cen4-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mdcs2-an115-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key flrp-cen4-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mblb-an373-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:169) Processed 30 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key flrp-cen5-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key flrp-cen6-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mblb-an374-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fclc-cen5-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mdcs2-cen1-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mblb-an375-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key flrp-cen7-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mdcs2-cen2-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fclc-cen6-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mdcs2-cen3-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fclc-cen7-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fclc-cen8-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mblb-cen1-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key flrp-cen8-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mdcs2-cen4-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fejs-an36-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fmjc-an116-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mdcs2-cen5-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mblb-cen2-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mblb-cen3-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fmjc-an117-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mdcs2-cen6-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fejs-an37-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fejs-an38-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fmjc-an118-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mblb-cen4-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mdcs2-cen7-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fmjc-an119-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:169) Processed 40 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mblb-cen5-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:169) Processed 40 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mdcs2-cen8-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fejs-an39-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fmjc-an120-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mblb-cen6-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:169) Processed 40 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fejs-an40-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mdmc-an206-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fmjc-cen1-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:169) Processed 40 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fmjc-cen2-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mblb-cen7-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mdmc-an207-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fmjc-cen3-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mblb-cen8-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fejs-cen1-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mblw-an1-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fejs-cen2-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mdmc-an208-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fmjc-cen4-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mdmc-an209-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fejs-cen3-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mblw-an2-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mblw-an3-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fmjc-cen5-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mblw-an4-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fmjc-cen6-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mdmc-an210-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fejs-cen4-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fmjc-cen7-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mblw-an5-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mdmc-cen1-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fmjc-cen8-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mblw-cen1-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fmjd-an191-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:169) Processed 50 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mblw-cen2-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fejs-cen5-b
VLOG[2] (compute-kaldi-pitch-feats[5.5.284-76bd]:main():compute-kaldi-pitch-feats.cc:107) Processed 50 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fejs-cen6-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mblw-cen3-b
VLOG[2] (compute-kaldi-pitch-feats[5.5.284-76bd]:main():compute-kaldi-pitch-feats.cc:107) Processed 50 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fejs-cen7-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mblw-cen4-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fmjd-an192-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mdmc-cen2-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mblw-cen5-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fejs-cen8-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key ffmm-an291-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mdmc-cen3-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mblw-cen6-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fmjd-an193-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:169) Processed 50 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key ffmm-an292-b
VLOG[2] (compute-kaldi-pitch-feats[5.5.284-76bd]:main():compute-kaldi-pitch-feats.cc:107) Processed 50 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mblw-cen7-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mdmc-cen4-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:169) Processed 50 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fmjd-an194-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mblw-cen8-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key ffmm-an293-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mbmg-an266-b
VLOG[2] (compute-kaldi-pitch-feats[5.5.284-76bd]:main():compute-kaldi-pitch-feats.cc:107) Processed 50 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fmjd-an195-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:169) Processed 50 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mdmc-cen5-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mdmc-cen6-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fmjd-cen1-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key ffmm-an294-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mdmc-cen7-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fmjd-cen2-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key ffmm-an295-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mbmg-an267-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fmjd-cen3-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mdmc-cen8-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key ffmm-cen1-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mdxn-an106-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key ffmm-cen2-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key ffmm-cen3-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mbmg-an268-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fmjd-cen4-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mdxn-an107-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key ffmm-cen4-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mdxn-an108-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:169) Processed 60 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mbmg-an269-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fmjd-cen5-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mdxn-an109-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key ffmm-cen5-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mbmg-an270-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key ffmm-cen6-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mdxn-an110-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mbmg-cen1-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:169) Processed 60 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key ffmm-cen7-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mdxn-cen1-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fmjd-cen6-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mbmg-cen2-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fmjd-cen7-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key ffmm-cen8-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mbmg-cen3-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fjam-an76-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fmjd-cen8-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:169) Processed 60 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mdxn-cen2-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mbmg-cen4-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:169) Processed 60 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fnsv-an181-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mdxn-cen3-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fnsv-an182-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mdxn-cen4-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mbmg-cen5-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fjam-an77-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mdxn-cen5-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mdxn-cen6-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mbmg-cen6-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fjam-an78-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fnsv-an183-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mbmg-cen7-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fnsv-an184-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fnsv-an185-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mbmg-cen8-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mdxn-cen7-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fjam-an79-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:169) Processed 70 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mcel-an201-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mdxn-cen8-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fnsv-cen1-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fjam-an80-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mdxs-an26-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fjam-cen1-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mcel-an202-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fnsv-cen2-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fnsv-cen3-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fjam-cen2-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mcel-an203-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fjam-cen3-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fnsv-cen4-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mcel-an204-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mdxs-an27-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mdxs-an28-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fnsv-cen5-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:169) Processed 70 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mdxs-an29-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:169) Processed 70 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fjam-cen4-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mcel-an205-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mcel-cen1-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mdxs-an30-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fjam-cen5-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mcel-cen2-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mcel-cen3-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mcel-cen4-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mdxs-cen1-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:169) Processed 70 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fnsv-cen6-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mdxs-cen2-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fjam-cen6-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mcel-cen5-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mdxs-cen3-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fnsv-cen7-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:169) Processed 80 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mcel-cen6-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fjam-cen7-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mcel-cen8-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fjam-cen8-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mdxs-cen4-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fjdn-an121-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fnsv-cen8-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fjdn-an122-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fjdn-an123-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mcen-an126-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mdxs-cen5-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mcen-an127-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fjdn-an124-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fplp-an91-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mdxs-cen6-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mcen-an128-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fjdn-an125-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mcen-an129-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mdxs-cen7-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:169) Processed 80 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fjdn-cen1-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fplp-an92-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fplp-an93-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mcen-an130-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fjdn-cen2-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mdxs-cen8-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fjdn-cen3-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mcen-cen1-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:169) Processed 80 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key meab-an136-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fjdn-cen4-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key meab-an137-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fplp-an94-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mcen-cen2-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fjdn-cen5-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key meab-an138-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fjdn-cen6-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mcen-cen3-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key meab-an139-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:169) Processed 90 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mcen-cen4-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fjdn-cen7-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fplp-an95-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mcen-cen5-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fjdn-cen8-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mcen-cen6-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key meab-an140-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fplp-cen1-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fjmd-an10-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mcen-cen7-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fjmd-an6-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key meab-cen1-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:169) Processed 90 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fjmd-an7-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fplp-cen2-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mcen-cen8-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fjmd-an8-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fjmd-an9-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mcfl-an261-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key meab-cen2-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:169) Processed 80 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fplp-cen3-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fjmd-cen1-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key meab-cen3-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key meab-cen4-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fjmd-cen2-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mcfl-an262-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fplp-cen4-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fjmd-cen3-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key meab-cen5-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fjmd-cen4-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mcfl-an263-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fjmd-cen5-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mcfl-an264-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:169) Processed 90 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key meab-cen6-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fjmd-cen6-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mcfl-an265-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fplp-cen5-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fjmd-cen7-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:169) Processed 100 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mcfl-cen1-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fplp-cen6-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key meab-cen7-b
VLOG[2] (compute-kaldi-pitch-feats[5.5.284-76bd]:main():compute-kaldi-pitch-feats.cc:107) Processed 100 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mcfl-cen2-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key meab-cen8-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fplp-cen7-b
VLOG[2] (compute-kaldi-pitch-feats[5.5.284-76bd]:main():compute-kaldi-pitch-feats.cc:107) Processed 100 utterances
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:169) Processed 100 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fjmd-cen8-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key meht-an66-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fplp-cen8-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mcfl-cen3-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fkai-an311-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key meht-an67-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key meht-an68-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key meht-an69-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fkai-an312-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mcfl-cen4-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key meht-an70-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fkai-an313-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fsaf2-an296-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mcfl-cen5-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fkai-an314-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key meht-cen1-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mcfl-cen6-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fkai-an315-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key meht-cen2-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fsaf2-an297-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fsaf2-an298-b
VLOG[2] (compute-kaldi-pitch-feats[5.5.284-76bd]:main():compute-kaldi-pitch-feats.cc:107) Processed 100 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fkai-cen1-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:169) Processed 100 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key meht-cen3-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fsaf2-an299-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fkai-cen2-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mcfl-cen7-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key meht-cen4-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fkai-cen3-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:169) Processed 90 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fsaf2-an300-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mcfl-cen8-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mcrt-an141-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fkai-cen4-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key meht-cen5-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mcrt-an142-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key meht-cen6-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:169) Processed 110 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mcrt-an143-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fsaf2-cen1-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:169) Processed 110 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fkai-cen5-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mcrt-an144-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fkai-cen6-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mcrt-an145-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key meht-cen7-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key meht-cen8-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mema-an286-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fkai-cen7-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mcrt-cen1-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fsaf2-cen2-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mema-an287-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fsaf2-cen3-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mcrt-cen2-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mema-an288-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fkai-cen8-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mema-an289-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fsaf2-cen4-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mcrt-cen3-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fkdo-an131-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fkdo-an132-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:169) Processed 110 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mema-an290-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fsaf2-cen5-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mema-cen1-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mcrt-cen4-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mema-cen2-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fsaf2-cen6-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fsaf2-cen7-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fkdo-an133-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mcrt-cen5-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fkdo-an134-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fsaf2-cen8-b
LOG (compute-kaldi-pitch-feats[5.5.284-76bd]:main():compute-kaldi-pitch-feats.cc:110) Done 119 utterances, 0 with errors.
LOG (process-kaldi-pitch-feats[5.5.284-76bd]:main():process-kaldi-pitch-feats.cc:85) Post-processed pitch for 119 utterances.
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mema-cen3-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mcrt-cen6-b
LOG (compute-kaldi-pitch-feats[5.5.284-76bd]:main():compute-kaldi-pitch-feats.cc:110) Done 119 utterances, 0 with errors.
LOG (process-kaldi-pitch-feats[5.5.284-76bd]:main():process-kaldi-pitch-feats.cc:85) Post-processed pitch for 119 utterances.
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fkdo-an135-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fkdo-cen1-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:173)  Done 119 out of 119 utterances.
LOG (paste-feats[5.5.284-76bd]:main():paste-feats.cc:158) Done 119 utts, errors on 0
LOG (copy-feats[5.5.284-76bd]:main():copy-feats.cc:143) Copied 119 feature matrices.
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mcrt-cen7-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:173)  Done 119 out of 119 utterances.
LOG (paste-feats[5.5.284-76bd]:main():paste-feats.cc:158) Done 119 utts, errors on 0
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fsrb-an166-b
LOG (copy-feats[5.5.284-76bd]:main():copy-feats.cc:143) Copied 119 feature matrices.
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:169) Processed 100 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fsrb-an167-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mema-cen4-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mema-cen5-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mema-cen6-b
paste-feats --length-tolerance=2 'ark:compute-fbank-feats  --verbose=2 --config=conf/fbank.conf scp,p:exp/make_fbank/train/wav.6.scp ark:- |' 'ark,s,cs:compute-kaldi-pitch-feats --verbose=2 --config=conf/pitch.conf scp,p:exp/make_fbank/train/wav.6.scp ark:- | process-kaldi-pitch-feats  ark:- ark:- |' ark:- 
copy-feats --compress=true --write-num-frames=ark,t:exp/make_fbank/train/utt2num_frames.6 ark:- ark,scp:/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/fbank/raw_fbank_pitch_train.6.ark,/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/fbank/raw_fbank_pitch_train.6.scp 
VLOG[2] (compute-kaldi-pitch-feats[5.5.284-76bd]:main():compute-kaldi-pitch-feats.cc:107) Processed 100 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fsrb-an168-b
compute-fbank-feats --verbose=2 --config=conf/fbank.conf scp,p:exp/make_fbank/train/wav.6.scp ark:- 
paste-feats --length-tolerance=2 'ark:compute-fbank-feats  --verbose=2 --config=conf/fbank.conf scp,p:exp/make_fbank/train/wav.5.scp ark:- |' 'ark,s,cs:compute-kaldi-pitch-feats --verbose=2 --config=conf/pitch.conf scp,p:exp/make_fbank/train/wav.5.scp ark:- | process-kaldi-pitch-feats  ark:- ark:- |' ark:- 
copy-feats --compress=true --write-num-frames=ark,t:exp/make_fbank/train/utt2num_frames.5 ark:- ark,scp:/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/fbank/raw_fbank_pitch_train.5.ark,/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/fbank/raw_fbank_pitch_train.5.scp 
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjjs2-an180-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mema-cen7-b
compute-kaldi-pitch-feats --verbose=2 --config=conf/pitch.conf scp,p:exp/make_fbank/train/wav.6.scp ark:- 
process-kaldi-pitch-feats ark:- ark:- 
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fsrb-an169-b
compute-fbank-feats --verbose=2 --config=conf/fbank.conf scp,p:exp/make_fbank/train/wav.5.scp ark:- 
LOG (compute-kaldi-pitch-feats[5.5.284-76bd]:main():compute-kaldi-pitch-feats.cc:110) Done 119 utterances, 0 with errors.
LOG (process-kaldi-pitch-feats[5.5.284-76bd]:main():process-kaldi-pitch-feats.cc:85) Post-processed pitch for 119 utterances.
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mema-cen8-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mewl-an256-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:173)  Done 119 out of 119 utterances.
LOG (paste-feats[5.5.284-76bd]:main():paste-feats.cc:158) Done 119 utts, errors on 0
LOG (copy-feats[5.5.284-76bd]:main():copy-feats.cc:143) Copied 119 feature matrices.
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fsrb-an170-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mewl-an257-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fsrb-cen1-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjjs2-cen1-b
compute-kaldi-pitch-feats --verbose=2 --config=conf/pitch.conf scp,p:exp/make_fbank/train/wav.5.scp ark:- 
process-kaldi-pitch-feats ark:- ark:- 
copy-feats --compress=true --write-num-frames=ark,t:exp/make_fbank/train/utt2num_frames.7 ark:- ark,scp:/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/fbank/raw_fbank_pitch_train.7.ark,/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/fbank/raw_fbank_pitch_train.7.scp 
paste-feats --length-tolerance=2 'ark:compute-fbank-feats  --verbose=2 --config=conf/fbank.conf scp,p:exp/make_fbank/train/wav.7.scp ark:- |' 'ark,s,cs:compute-kaldi-pitch-feats --verbose=2 --config=conf/pitch.conf scp,p:exp/make_fbank/train/wav.7.scp ark:- | process-kaldi-pitch-feats  ark:- ark:- |' ark:- 
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fsrb-cen2-b
compute-fbank-feats --verbose=2 --config=conf/fbank.conf scp,p:exp/make_fbank/train/wav.7.scp ark:- 
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fsrb-cen3-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmtm-cen2-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmtm-cen3-b
process-kaldi-pitch-feats ark:- ark:- 
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjjs2-cen2-b
compute-kaldi-pitch-feats --verbose=2 --config=conf/pitch.conf scp,p:exp/make_fbank/train/wav.7.scp ark:- 
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjjs2-cen4-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mewl-an258-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mewl-an259-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fsrb-cen4-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjjs2-cen5-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fsrb-cen5-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mewl-an260-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmtm-cen4-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fsrb-cen6-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mewl-cen1-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmtm-cen5-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjjs2-cen6-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:169) Processed 110 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fsrb-cen7-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mewl-cen2-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjjs2-cen7-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mewl-cen3-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmtm-cen6-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key fsrb-cen8-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key ftal-an326-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmtm-cen7-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mewl-cen4-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjjs2-cen8-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmtm-cen8-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mewl-cen5-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key ftal-an327-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mkdb-an316-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key ftal-an328-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:169) Processed 10 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mewl-cen6-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mnfe-an346-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:169) Processed 10 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mkdb-an317-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key ftal-an329-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mkdb-an318-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mewl-cen7-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mnfe-an347-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mewl-cen8-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mkdb-an319-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key ftal-an330-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mkdb-an320-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mfaa-an161-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key ftal-cen1-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mkdb-cen1-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mkdb-cen2-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key ftal-cen2-b
LOG (compute-kaldi-pitch-feats[5.5.284-76bd]:main():compute-kaldi-pitch-feats.cc:110) Done 119 utterances, 0 with errors.
LOG (process-kaldi-pitch-feats[5.5.284-76bd]:main():process-kaldi-pitch-feats.cc:85) Post-processed pitch for 119 utterances.
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key ftal-cen3-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:173)  Done 119 out of 119 utterances.
LOG (paste-feats[5.5.284-76bd]:main():paste-feats.cc:158) Done 119 utts, errors on 0
LOG (copy-feats[5.5.284-76bd]:main():copy-feats.cc:143) Copied 119 feature matrices.
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:169) Processed 10 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mnfe-an348-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mfaa-an162-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mkdb-cen3-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mfaa-an163-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mnfe-an349-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mfaa-an164-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mnfe-an350-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mkdb-cen4-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mfaa-an165-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mnfe-cen1-b
paste-feats --length-tolerance=2 'ark:compute-fbank-feats  --verbose=2 --config=conf/fbank.conf scp,p:exp/make_fbank/train/wav.8.scp ark:- |' 'ark,s,cs:compute-kaldi-pitch-feats --verbose=2 --config=conf/pitch.conf scp,p:exp/make_fbank/train/wav.8.scp ark:- | process-kaldi-pitch-feats  ark:- ark:- |' ark:- 
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mfaa-cen1-b
copy-feats --compress=true --write-num-frames=ark,t:exp/make_fbank/train/utt2num_frames.8 ark:- ark,scp:/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/fbank/raw_fbank_pitch_train.8.ark,/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/fbank/raw_fbank_pitch_train.8.scp 
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mkdb-cen5-b
compute-fbank-feats --verbose=2 --config=conf/fbank.conf scp,p:exp/make_fbank/train/wav.8.scp ark:- 
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mnfe-cen2-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mkdb-cen6-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mfaa-cen2-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mnfe-cen3-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:169) Processed 20 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mkdb-cen7-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key msjm-cen4-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:169) Processed 20 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mfaa-cen3-b
compute-kaldi-pitch-feats --verbose=2 --config=conf/pitch.conf scp,p:exp/make_fbank/train/wav.8.scp ark:- 
process-kaldi-pitch-feats ark:- ark:- 
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mfaa-cen4-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mkdb-cen8-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mnfe-cen4-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mkem-an186-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mfaa-cen5-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mfaa-cen6-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mkem-an187-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mnfe-cen5-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mkem-an188-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key msjm-cen5-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mfaa-cen7-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mnfe-cen6-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key msjm-cen6-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mfaa-cen8-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mkem-an189-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key msjm-cen7-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mgah-an196-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mnfe-cen7-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mkem-an190-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:169) Processed 20 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mnfe-cen8-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key msjm-cen8-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mgah-an197-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mkem-cen1-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mgah-an198-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mnjl-an81-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key msjr-an351-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mkem-cen2-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mgah-an199-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key msjr-an352-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:169) Processed 30 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mgah-an200-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key msjr-an353-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mnjl-an82-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mnjl-an83-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key msjr-an354-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mkem-cen4-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:169) Processed 10 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key msjr-an355-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mgah-cen1-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mnjl-an84-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mgah-cen2-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mgah-cen3-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:169) Processed 30 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mkem-cen5-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mnjl-an85-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key msjr-cen1-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mgah-cen4-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mnjl-cen1-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mkem-cen6-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mkem-cen7-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mnjl-cen2-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mkem-cen8-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mgah-cen5-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key msjr-cen2-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmaf-an51-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mnjl-cen3-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key msjr-cen3-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmaf-an52-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key msjr-cen4-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mnjl-cen5-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mgah-cen6-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmaf-an53-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key msjr-cen5-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:169) Processed 30 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mnjl-cen6-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmaf-an54-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mgah-cen7-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mnjl-cen7-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key msjr-cen6-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmaf-an55-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mgah-cen8-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjbh-an246-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:169) Processed 40 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjbh-an247-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mnjl-cen8-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key msjr-cen7-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key msjr-cen8-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mskh-an100-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mrab-an71-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mrab-an72-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:169) Processed 20 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mskh-an96-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmaf-cen1-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjbh-an248-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mskh-an97-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mrab-an73-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjbh-an249-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:169) Processed 40 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmaf-cen2-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mskh-an98-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjbh-an250-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmaf-cen3-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmaf-cen4-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mskh-an99-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mrab-an74-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mskh-cen1-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjbh-cen1-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mskh-cen2-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mrab-an75-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmaf-cen5-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mrab-cen1-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmaf-cen6-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjbh-cen2-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjbh-cen3-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mrab-cen2-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:169) Processed 40 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mrab-cen3-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mskh-cen3-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmaf-cen7-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjbh-cen4-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mskh-cen4-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjbh-cen5-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mrab-cen4-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmaf-cen8-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjbh-cen6-b
VLOG[2] (compute-kaldi-pitch-feats[5.5.284-76bd]:main():compute-kaldi-pitch-feats.cc:107) Processed 50 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mrab-cen5-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmal-an321-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mskh-cen5-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:169) Processed 50 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjbh-cen7-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mrab-cen6-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmal-an322-b
VLOG[2] (compute-kaldi-pitch-feats[5.5.284-76bd]:main():compute-kaldi-pitch-feats.cc:107) Processed 50 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmal-an323-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mskh-cen6-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjbh-cen8-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mrab-cen7-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:169) Processed 30 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mskh-cen7-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:169) Processed 50 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmal-an324-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjda-an171-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mrab-cen8-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmal-an325-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjda-an172-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmal-cen1-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mrcb-an101-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mskh-cen8-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmal-cen2-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mrcb-an102-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmal-cen3-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjda-an173-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key msmn-an276-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key msmn-an277-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjda-an174-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmal-cen4-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjda-an175-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mrcb-an103-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mrcb-an104-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmal-cen5-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key msmn-an278-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmal-cen6-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjda-cen1-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjda-cen2-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:169) Processed 50 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mrcb-an105-b
VLOG[2] (compute-kaldi-pitch-feats[5.5.284-76bd]:main():compute-kaldi-pitch-feats.cc:107) Processed 50 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmal-cen7-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjda-cen3-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmal-cen8-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:169) Processed 60 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmap-an46-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key msmn-an279-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mrcb-cen1-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmap-an47-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key msmn-cen4-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mrcb-cen2-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmap-an48-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key msmn-cen5-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:169) Processed 60 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjda-cen4-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mrcb-cen3-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmap-an49-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key msmn-cen6-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmap-an50-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mrcb-cen4-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjda-cen5-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key msmn-cen7-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjda-cen6-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mrcb-cen5-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:169) Processed 40 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key msmn-cen8-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mrcb-cen6-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmap-cen1-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key msrb-an331-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjda-cen7-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mrcb-cen7-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmap-cen2-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmap-cen3-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjda-cen8-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmap-cen4-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key msrb-an332-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mrcb-cen8-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjdr-an16-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjdr-an18-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmap-cen5-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mrjc2-an236-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key msrb-an333-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:169) Processed 70 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmap-cen6-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjdr-an19-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:169) Processed 60 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mrjc2-an237-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key msrb-an334-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmap-cen7-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key msrb-an335-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjdr-an20-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmap-cen8-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mrjc2-an238-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key msrb-cen1-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmdg-an336-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjdr-cen1-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mrjc2-an239-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mrjc2-an240-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key msrb-cen2-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmdg-an337-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:169) Processed 70 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjdr-cen2-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mrjc2-cen1-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key msrb-cen3-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmdg-an338-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjdr-cen3-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmdg-an339-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmdg-an340-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mrjc2-cen2-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key msrb-cen4-b
VLOG[2] (compute-kaldi-pitch-feats[5.5.284-76bd]:main():compute-kaldi-pitch-feats.cc:107) Processed 50 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjdr-cen4-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mrjc2-cen3-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjdr-cen5-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmdg-cen1-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjdr-cen6-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:169) Processed 50 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key msrb-cen5-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmdg-cen2-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:169) Processed 80 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmdg-cen3-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjdr-cen7-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key msrb-cen6-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mrjc2-cen4-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjdr-cen8-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjes-an156-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key msrb-cen7-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmdg-cen4-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key msrb-cen8-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjes-an157-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mrjc2-cen5-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmdg-cen5-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mtcv-an301-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mrjc2-cen6-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmdg-cen6-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:169) Processed 70 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mrjc2-cen7-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mtcv-an302-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjes-an158-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmdg-cen7-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mrjc2-cen8-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mtcv-an303-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mrmg-an281-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmdg-cen8-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:169) Processed 80 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjes-an159-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mtcv-an304-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjes-an160-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mtcv-an305-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mrmg-an282-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmkw-an361-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjes-cen1-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmkw-an362-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mrmg-an283-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mtcv-cen1-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mrmg-an284-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:169) Processed 60 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mtcv-cen2-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmkw-an363-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjes-cen2-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mrmg-an285-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mtcv-cen3-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjes-cen3-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmkw-an364-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mtcv-cen4-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mrmg-cen1-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjes-cen4-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mtcv-cen5-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:169) Processed 90 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmkw-an365-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mtcv-cen6-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mrmg-cen2-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjes-cen5-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mrmg-cen3-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjes-cen6-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:169) Processed 80 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mrmg-cen4-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mtcv-cen7-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmkw-an59-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mrmg-cen5-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjes-cen7-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mtcv-cen8-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmkw-cen1-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mtje-an31-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjes-cen8-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mtje-an32-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:169) Processed 90 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjgk-an216-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mrmg-cen6-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmkw-cen2-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mrmg-cen7-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjgk-an217-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mtje-an33-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjgk-an218-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:169) Processed 70 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mtje-an34-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mrmg-cen8-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjgk-an219-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mscg2-an356-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmkw-cen3-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjgk-an220-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mscg2-an357-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mtje-an35-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjgk-cen1-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmkw-cen4-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mscg2-an358-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjgk-cen2-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mscg2-an359-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmkw-cen5-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjgk-cen4-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmkw-cen6-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mscg2-an360-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mtje-cen1-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjgk-cen5-b
VLOG[2] (compute-kaldi-pitch-feats[5.5.284-76bd]:main():compute-kaldi-pitch-feats.cc:107) Processed 100 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmkw-cen7-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mtje-cen2-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjgk-cen6-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mtje-cen3-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:169) Processed 90 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mscg2-cen1-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmkw-cen8-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:169) Processed 100 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjgk-cen7-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjgk-cen8-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mscg2-cen2-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjhp-an56-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:169) Processed 100 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmsh-an381-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjhp-an57-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mtje-cen4-b
VLOG[2] (compute-kaldi-pitch-feats[5.5.284-76bd]:main():compute-kaldi-pitch-feats.cc:107) Processed 100 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mscg2-cen3-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjhp-an58-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmsh-an382-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjhp-an59-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mscg2-cen4-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mtje-cen5-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjhp-an60-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mtje-cen6-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmsh-an383-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mscg2-cen5-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjhp-cen1-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mtje-cen7-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmsh-an384-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mtje-cen8-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmsh-an385-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjhp-cen2-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:169) Processed 80 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mtos-an366-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mscg2-cen6-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjhp-cen3-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmsh-cen1-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mscg2-cen7-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmsh-cen2-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mscg2-cen8-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:169) Processed 110 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjhp-cen4-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key msct-an341-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjhp-cen5-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key msct-an342-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmsh-cen3-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mtos-an367-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjhp-cen6-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:169) Processed 100 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key msct-an343-b
VLOG[2] (compute-kaldi-pitch-feats[5.5.284-76bd]:main():compute-kaldi-pitch-feats.cc:107) Processed 100 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmsh-cen4-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key msct-an344-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjhp-cen7-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmsh-cen5-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mtos-an368-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:169) Processed 110 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmsh-cen6-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key msct-an345-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mtos-an369-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjhp-cen8-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mtos-an370-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmsh-cen7-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjjs2-an176-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key msct-cen1-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mtos-cen1-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmsh-cen8-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjjs2-an177-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key msct-cen2-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmtm-an386-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mtos-cen2-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key msct-cen3-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjjs2-an178-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mtos-cen3-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmtm-an387-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key msct-cen4-b
LOG (compute-kaldi-pitch-feats[5.5.284-76bd]:main():compute-kaldi-pitch-feats.cc:110) Done 118 utterances, 0 with errors.
LOG (process-kaldi-pitch-feats[5.5.284-76bd]:main():process-kaldi-pitch-feats.cc:85) Post-processed pitch for 118 utterances.
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mtos-cen4-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmtm-an388-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mjjs2-an179-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:173)  Done 118 out of 118 utterances.
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mtos-cen5-b
LOG (paste-feats[5.5.284-76bd]:main():paste-feats.cc:158) Done 118 utts, errors on 0
LOG (copy-feats[5.5.284-76bd]:main():copy-feats.cc:143) Copied 118 feature matrices.
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key msct-cen5-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmtm-an389-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:169) Processed 90 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mtos-cen6-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mtos-cen7-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key msct-cen6-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmtm-an390-b
LOG (compute-kaldi-pitch-feats[5.5.284-76bd]:main():compute-kaldi-pitch-feats.cc:110) Done 118 utterances, 0 with errors.
LOG (process-kaldi-pitch-feats[5.5.284-76bd]:main():process-kaldi-pitch-feats.cc:85) Post-processed pitch for 118 utterances.
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mtos-cen8-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mmtm-cen1-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:173)  Done 118 out of 118 utterances.
LOG (paste-feats[5.5.284-76bd]:main():paste-feats.cc:158) Done 118 utts, errors on 0
LOG (copy-feats[5.5.284-76bd]:main():copy-feats.cc:143) Copied 118 feature matrices.
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key msct-cen7-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mtxj-an376-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:169) Processed 110 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key msct-cen8-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key msjm-an226-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mtxj-an377-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mtxj-an378-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mtxj-an379-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key msjm-an227-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key msjm-an228-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mtxj-an380-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key msjm-an229-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key msjm-an230-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mtxj-cen1-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mtxj-cen2-b
LOG (compute-kaldi-pitch-feats[5.5.284-76bd]:main():compute-kaldi-pitch-feats.cc:110) Done 118 utterances, 0 with errors.
LOG (process-kaldi-pitch-feats[5.5.284-76bd]:main():process-kaldi-pitch-feats.cc:85) Post-processed pitch for 118 utterances.
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key msjm-cen1-b
VLOG[2] (compute-kaldi-pitch-feats[5.5.284-76bd]:main():compute-kaldi-pitch-feats.cc:107) Processed 100 utterances
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:169) Processed 100 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mtxj-cen3-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key msjm-cen2-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key msjm-cen3-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:173)  Done 118 out of 118 utterances.
LOG (paste-feats[5.5.284-76bd]:main():paste-feats.cc:158) Done 118 utts, errors on 0
LOG (copy-feats[5.5.284-76bd]:main():copy-feats.cc:143) Copied 118 feature matrices.
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mtxj-cen4-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mtxj-cen5-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mtxj-cen6-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mtxj-cen7-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mtxj-cen8-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mwhw-an151-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mwhw-an152-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mwhw-an153-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mwhw-an154-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:169) Processed 110 utterances
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mwhw-an155-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mwhw-cen1-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mwhw-cen2-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mwhw-cen3-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mwhw-cen4-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mwhw-cen5-b
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mwhw-cen6-b
LOG (compute-kaldi-pitch-feats[5.5.284-76bd]:main():compute-kaldi-pitch-feats.cc:110) Done 118 utterances, 0 with errors.
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mwhw-cen7-b
LOG (process-kaldi-pitch-feats[5.5.284-76bd]:main():process-kaldi-pitch-feats.cc:85) Post-processed pitch for 118 utterances.
VLOG[2] (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:170) Processed features for key mwhw-cen8-b
LOG (compute-fbank-feats[5.5.284-76bd]:main():compute-fbank-feats.cc:173)  Done 118 out of 118 utterances.
LOG (paste-feats[5.5.284-76bd]:main():paste-feats.cc:158) Done 118 utts, errors on 0
LOG (copy-feats[5.5.284-76bd]:main():copy-feats.cc:143) Copied 118 feature matrices.
Succeeded creating filterbank & pitch features for train
fix_data_dir.sh: kept all 948 utterances.
fix_data_dir.sh: old files are kept in data/train/.backup
utils/subset_data_dir.sh: reducing #utt from 948 to 100
utils/subset_data_dir.sh: reducing #utt from 948 to 848
compute-cmvn-stats scp:data/train_nodev/feats.scp data/train_nodev/cmvn.ark 
LOG (compute-cmvn-stats[5.5.284-76bd]:main():compute-cmvn-stats.cc:168) Wrote global CMVN stats to data/train_nodev/cmvn.ark
LOG (compute-cmvn-stats[5.5.284-76bd]:main():compute-cmvn-stats.cc:171) Done accumulating CMVN stats for 848 utterances; 0 had errors.
/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/../../../utils/dump.sh --cmd stdout.pl --nj 8 --do_delta false data/train_nodev/feats.scp data/train_nodev/cmvn.ark exp/dump_feats/train dump/train_nodev/deltafalse
apply-cmvn --norm-vars=true data/train_nodev/cmvn.ark scp:exp/dump_feats/train/feats.1.scp ark:- 
apply-cmvn --norm-vars=true data/train_nodev/cmvn.ark scp:exp/dump_feats/train/feats.4.scp ark:- 
apply-cmvn --norm-vars=true data/train_nodev/cmvn.ark scp:exp/dump_feats/train/feats.2.scp ark:- 
apply-cmvn --norm-vars=true data/train_nodev/cmvn.ark scp:exp/dump_feats/train/feats.3.scp ark:- 
LOG (apply-cmvn[5.5.284-76bd]:main():apply-cmvn.cc:159) Applied cepstral mean and variance normalization to 106 utterances, errors on 0
LOG (apply-cmvn[5.5.284-76bd]:main():apply-cmvn.cc:159) Applied cepstral mean and variance normalization to 106 utterances, errors on 0
LOG (apply-cmvn[5.5.284-76bd]:main():apply-cmvn.cc:159) Applied cepstral mean and variance normalization to 106 utterances, errors on 0
apply-cmvn --norm-vars=true data/train_nodev/cmvn.ark scp:exp/dump_feats/train/feats.5.scp ark:- 
LOG (apply-cmvn[5.5.284-76bd]:main():apply-cmvn.cc:159) Applied cepstral mean and variance normalization to 106 utterances, errors on 0
apply-cmvn --norm-vars=true data/train_nodev/cmvn.ark scp:exp/dump_feats/train/feats.6.scp ark:- 
apply-cmvn --norm-vars=true data/train_nodev/cmvn.ark scp:exp/dump_feats/train/feats.8.scp ark:- 
apply-cmvn --norm-vars=true data/train_nodev/cmvn.ark scp:exp/dump_feats/train/feats.7.scp ark:- 
LOG (apply-cmvn[5.5.284-76bd]:main():apply-cmvn.cc:159) Applied cepstral mean and variance normalization to 106 utterances, errors on 0
LOG (apply-cmvn[5.5.284-76bd]:main():apply-cmvn.cc:159) Applied cepstral mean and variance normalization to 106 utterances, errors on 0
LOG (apply-cmvn[5.5.284-76bd]:main():apply-cmvn.cc:159) Applied cepstral mean and variance normalization to 106 utterances, errors on 0
LOG (apply-cmvn[5.5.284-76bd]:main():apply-cmvn.cc:159) Applied cepstral mean and variance normalization to 106 utterances, errors on 0
/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/../../../utils/dump.sh --cmd stdout.pl --nj 8 --do_delta false data/train_dev/feats.scp data/train_nodev/cmvn.ark exp/dump_feats/dev dump/train_dev/deltafalse
apply-cmvn --norm-vars=true data/train_nodev/cmvn.ark scp:exp/dump_feats/dev/feats.3.scp ark:- 
apply-cmvn --norm-vars=true data/train_nodev/cmvn.ark scp:exp/dump_feats/dev/feats.1.scp ark:- 
apply-cmvn --norm-vars=true data/train_nodev/cmvn.ark scp:exp/dump_feats/dev/feats.4.scp ark:- 
apply-cmvn --norm-vars=true data/train_nodev/cmvn.ark scp:exp/dump_feats/dev/feats.2.scp ark:- 
LOG (apply-cmvn[5.5.284-76bd]:main():apply-cmvn.cc:159) Applied cepstral mean and variance normalization to 13 utterances, errors on 0
LOG (apply-cmvn[5.5.284-76bd]:main():apply-cmvn.cc:159) Applied cepstral mean and variance normalization to 13 utterances, errors on 0
LOG (apply-cmvn[5.5.284-76bd]:main():apply-cmvn.cc:159) Applied cepstral mean and variance normalization to 13 utterances, errors on 0
apply-cmvn --norm-vars=true data/train_nodev/cmvn.ark scp:exp/dump_feats/dev/feats.5.scp ark:- 
apply-cmvn --norm-vars=true data/train_nodev/cmvn.ark scp:exp/dump_feats/dev/feats.6.scp ark:- 
apply-cmvn --norm-vars=true data/train_nodev/cmvn.ark scp:exp/dump_feats/dev/feats.7.scp ark:- 
LOG (apply-cmvn[5.5.284-76bd]:main():apply-cmvn.cc:159) Applied cepstral mean and variance normalization to 13 utterances, errors on 0
apply-cmvn --norm-vars=true data/train_nodev/cmvn.ark scp:exp/dump_feats/dev/feats.8.scp ark:- 
LOG (apply-cmvn[5.5.284-76bd]:main():apply-cmvn.cc:159) Applied cepstral mean and variance normalization to 12 utterances, errors on 0
LOG (apply-cmvn[5.5.284-76bd]:main():apply-cmvn.cc:159) Applied cepstral mean and variance normalization to 12 utterances, errors on 0
LOG (apply-cmvn[5.5.284-76bd]:main():apply-cmvn.cc:159) Applied cepstral mean and variance normalization to 12 utterances, errors on 0
LOG (apply-cmvn[5.5.284-76bd]:main():apply-cmvn.cc:159) Applied cepstral mean and variance normalization to 12 utterances, errors on 0
/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/../../../utils/dump.sh --cmd stdout.pl --nj 8 --do_delta false data/train_dev/feats.scp data/train_nodev/cmvn.ark exp/dump_feats/recog/train_dev dump/train_dev/deltafalse
apply-cmvn --norm-vars=true data/train_nodev/cmvn.ark scp:exp/dump_feats/recog/train_dev/feats.1.scp ark:- 
apply-cmvn --norm-vars=true data/train_nodev/cmvn.ark scp:exp/dump_feats/recog/train_dev/feats.3.scp ark:- 
apply-cmvn --norm-vars=true data/train_nodev/cmvn.ark scp:exp/dump_feats/recog/train_dev/feats.2.scp ark:- 
apply-cmvn --norm-vars=true data/train_nodev/cmvn.ark scp:exp/dump_feats/recog/train_dev/feats.4.scp ark:- 
LOG (apply-cmvn[5.5.284-76bd]:main():apply-cmvn.cc:159) Applied cepstral mean and variance normalization to 13 utterances, errors on 0
LOG (apply-cmvn[5.5.284-76bd]:main():apply-cmvn.cc:159) Applied cepstral mean and variance normalization to 13 utterances, errors on 0
LOG (apply-cmvn[5.5.284-76bd]:main():apply-cmvn.cc:159) Applied cepstral mean and variance normalization to 13 utterances, errors on 0
apply-cmvn --norm-vars=true data/train_nodev/cmvn.ark scp:exp/dump_feats/recog/train_dev/feats.5.scp ark:- 
LOG (apply-cmvn[5.5.284-76bd]:main():apply-cmvn.cc:159) Applied cepstral mean and variance normalization to 13 utterances, errors on 0
apply-cmvn --norm-vars=true data/train_nodev/cmvn.ark scp:exp/dump_feats/recog/train_dev/feats.6.scp ark:- 
apply-cmvn --norm-vars=true data/train_nodev/cmvn.ark scp:exp/dump_feats/recog/train_dev/feats.7.scp ark:- 
apply-cmvn --norm-vars=true data/train_nodev/cmvn.ark scp:exp/dump_feats/recog/train_dev/feats.8.scp ark:- 
LOG (apply-cmvn[5.5.284-76bd]:main():apply-cmvn.cc:159) Applied cepstral mean and variance normalization to 12 utterances, errors on 0
LOG (apply-cmvn[5.5.284-76bd]:main():apply-cmvn.cc:159) Applied cepstral mean and variance normalization to 12 utterances, errors on 0
LOG (apply-cmvn[5.5.284-76bd]:main():apply-cmvn.cc:159) Applied cepstral mean and variance normalization to 12 utterances, errors on 0
LOG (apply-cmvn[5.5.284-76bd]:main():apply-cmvn.cc:159) Applied cepstral mean and variance normalization to 12 utterances, errors on 0
/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/../../../utils/dump.sh --cmd stdout.pl --nj 8 --do_delta false data/test/feats.scp data/train_nodev/cmvn.ark exp/dump_feats/recog/test dump/test/deltafalse
apply-cmvn --norm-vars=true data/train_nodev/cmvn.ark scp:exp/dump_feats/recog/test/feats.4.scp ark:- 
apply-cmvn --norm-vars=true data/train_nodev/cmvn.ark scp:exp/dump_feats/recog/test/feats.1.scp ark:- 
apply-cmvn --norm-vars=true data/train_nodev/cmvn.ark scp:exp/dump_feats/recog/test/feats.3.scp ark:- 
apply-cmvn --norm-vars=true data/train_nodev/cmvn.ark scp:exp/dump_feats/recog/test/feats.2.scp ark:- 
LOG (apply-cmvn[5.5.284-76bd]:main():apply-cmvn.cc:159) Applied cepstral mean and variance normalization to 16 utterances, errors on 0
LOG (apply-cmvn[5.5.284-76bd]:main():apply-cmvn.cc:159) Applied cepstral mean and variance normalization to 16 utterances, errors on 0
LOG (apply-cmvn[5.5.284-76bd]:main():apply-cmvn.cc:159) Applied cepstral mean and variance normalization to 17 utterances, errors on 0
LOG (apply-cmvn[5.5.284-76bd]:main():apply-cmvn.cc:159) Applied cepstral mean and variance normalization to 17 utterances, errors on 0
apply-cmvn --norm-vars=true data/train_nodev/cmvn.ark scp:exp/dump_feats/recog/test/feats.5.scp ark:- 
apply-cmvn --norm-vars=true data/train_nodev/cmvn.ark scp:exp/dump_feats/recog/test/feats.6.scp ark:- 
apply-cmvn --norm-vars=true data/train_nodev/cmvn.ark scp:exp/dump_feats/recog/test/feats.7.scp ark:- 
apply-cmvn --norm-vars=true data/train_nodev/cmvn.ark scp:exp/dump_feats/recog/test/feats.8.scp ark:- 
LOG (apply-cmvn[5.5.284-76bd]:main():apply-cmvn.cc:159) Applied cepstral mean and variance normalization to 16 utterances, errors on 0
LOG (apply-cmvn[5.5.284-76bd]:main():apply-cmvn.cc:159) Applied cepstral mean and variance normalization to 16 utterances, errors on 0
LOG (apply-cmvn[5.5.284-76bd]:main():apply-cmvn.cc:159) Applied cepstral mean and variance normalization to 16 utterances, errors on 0
LOG (apply-cmvn[5.5.284-76bd]:main():apply-cmvn.cc:159) Applied cepstral mean and variance normalization to 16 utterances, errors on 0
dictionary: data/lang_1char/train_nodev_units.txt
stage 2: Dictionary and Json Data Preparation
28 data/lang_1char/train_nodev_units.txt
/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/../../../utils/data2json.sh --feat dump/train_nodev/deltafalse/feats.scp data/train_nodev data/lang_1char/train_nodev_units.txt
/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/../../../utils/feat_to_shape.sh --cmd run.pl --nj 1 --filetype  --preprocess-conf  --verbose 0 dump/train_nodev/deltafalse/feats.scp data/train_nodev/tmp-QK1cG/input_1/shape.scp
/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/../../../utils/data2json.sh --feat dump/train_dev/deltafalse/feats.scp data/train_dev data/lang_1char/train_nodev_units.txt
/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/../../../utils/feat_to_shape.sh --cmd run.pl --nj 1 --filetype  --preprocess-conf  --verbose 0 dump/train_dev/deltafalse/feats.scp data/train_dev/tmp-i67z3/input_1/shape.scp
/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/../../../utils/data2json.sh --feat dump/train_dev/deltafalse/feats.scp data/train_dev data/lang_1char/train_nodev_units.txt
/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/../../../utils/feat_to_shape.sh --cmd run.pl --nj 1 --filetype  --preprocess-conf  --verbose 0 dump/train_dev/deltafalse/feats.scp data/train_dev/tmp-Xxtop/input_1/shape.scp
/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/../../../utils/data2json.sh --feat dump/test/deltafalse/feats.scp data/test data/lang_1char/train_nodev_units.txt
/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/../../../utils/feat_to_shape.sh --cmd run.pl --nj 1 --filetype  --preprocess-conf  --verbose 0 dump/test/deltafalse/feats.scp data/test/tmp-vY6cG/input_1/shape.scp

JSON dataset

Stage 2 dumps dataset of the speech feature and trascription pairs into JSON as same as TTS recipe

In [17]:
!cat espnet/egs/an4/asr1/dump/train_dev/deltafalse/data.json
{
    "utts": {
        "fash-an251-b": {
            "input": [
                {
                    "feat": "/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/dump/train_dev/deltafalse/feats.1.ark:13",
                    "name": "input1",
                    "shape": [
                        98,
                        83
                    ]
                }
            ],
            "output": [
                {
                    "name": "target1",
                    "shape": [
                        3,
                        30
                    ],
                    "text": "YES",
                    "token": "Y E S",
                    "tokenid": "27 7 21"
                }
            ],
            "utt2spk": "fash"
        },
        "fash-an253-b": {
            "input": [
                {
                    "feat": "/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/dump/train_dev/deltafalse/feats.1.ark:8845",
                    "name": "input1",
                    "shape": [
                        68,
                        83
                    ]
                }
            ],
            "output": [
                {
                    "name": "target1",
                    "shape": [
                        2,
                        30
                    ],
                    "text": "GO",
                    "token": "G O",
                    "tokenid": "9 17"
                }
            ],
            "utt2spk": "fash"
        },
        "fash-an254-b": {
            "input": [
                {
                    "feat": "/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/dump/train_dev/deltafalse/feats.1.ark:15187",
                    "name": "input1",
                    "shape": [
                        88,
                        83
                    ]
                }
            ],
            "output": [
                {
                    "name": "target1",
                    "shape": [
                        3,
                        30
                    ],
                    "text": "YES",
                    "token": "Y E S",
                    "tokenid": "27 7 21"
                }
            ],
            "utt2spk": "fash"
        },
        "fash-an255-b": {
            "input": [
                {
                    "feat": "/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/dump/train_dev/deltafalse/feats.1.ark:23189",
                    "name": "input1",
                    "shape": [
                        258,
                        83
                    ]
                }
            ],
            "output": [
                {
                    "name": "target1",
                    "shape": [
                        13,
                        30
                    ],
                    "text": "U M N Y H SIX",
                    "token": "U <space> M <space> N <space> Y <space> H <space> S I X",
                    "tokenid": "23 2 15 2 16 2 27 2 10 2 21 11 26"
                }
            ],
            "utt2spk": "fash"
        },
        "fash-cen1-b": {
            "input": [
                {
                    "feat": "/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/dump/train_dev/deltafalse/feats.1.ark:45300",
                    "name": "input1",
                    "shape": [
                        348,
                        83
                    ]
                }
            ],
            "output": [
                {
                    "name": "target1",
                    "shape": [
                        11,
                        30
                    ],
                    "text": "H I N I C H",
                    "token": "H <space> I <space> N <space> I <space> C <space> H",
                    "tokenid": "10 2 11 2 16 2 11 2 5 2 10"
                }
            ],
            "utt2spk": "fash"
        },
        "fash-cen2-b": {
            "input": [
                {
                    "feat": "/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/dump/train_dev/deltafalse/feats.1.ark:74881",
                    "name": "input1",
                    "shape": [
                        128,
                        83
                    ]
                }
            ],
            "output": [
                {
                    "name": "target1",
                    "shape": [
                        5,
                        30
                    ],
                    "text": "A M Y",
                    "token": "A <space> M <space> Y",
                    "tokenid": "3 2 15 2 27"
                }
            ],
            "utt2spk": "fash"
        },
        "fash-cen4-b": {
            "input": [
                {
                    "feat": "/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/dump/train_dev/deltafalse/feats.1.ark:86202",
                    "name": "input1",
                    "shape": [
                        358,
                        83
                    ]
                }
            ],
            "output": [
                {
                    "name": "target1",
                    "shape": [
                        15,
                        30
                    ],
                    "text": "M O R E W O O D",
                    "token": "M <space> O <space> R <space> E <space> W <space> O <space> O <space> D",
                    "tokenid": "15 2 17 2 20 2 7 2 25 2 17 2 17 2 6"
                }
            ],
            "utt2spk": "fash"
        },
        "fash-cen5-b": {
            "input": [
                {
                    "feat": "/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/dump/train_dev/deltafalse/feats.1.ark:116613",
                    "name": "input1",
                    "shape": [
                        448,
                        83
                    ]
                }
            ],
            "output": [
                {
                    "name": "target1",
                    "shape": [
                        19,
                        30
                    ],
                    "text": "P I T T S B U R G H",
                    "token": "P <space> I <space> T <space> T <space> S <space> B <space> U <space> R <space> G <space> H",
                    "tokenid": "18 2 11 2 22 2 22 2 21 2 4 2 23 2 20 2 9 2 10"
                }
            ],
            "utt2spk": "fash"
        },
        "fash-cen7-b": {
            "input": [
                {
                    "feat": "/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/dump/train_dev/deltafalse/feats.1.ark:154494",
                    "name": "input1",
                    "shape": [
                        248,
                        83
                    ]
                }
            ],
            "output": [
                {
                    "name": "target1",
                    "shape": [
                        33,
                        30
                    ],
                    "text": "TWO SIX EIGHT FOUR FOUR ONE EIGHT",
                    "token": "T W O <space> S I X <space> E I G H T <space> F O U R <space> F O U R <space> O N E <space> E I G H T",
                    "tokenid": "22 25 17 2 21 11 26 2 7 11 9 10 22 2 8 17 23 20 2 8 17 23 20 2 17 16 7 2 7 11 9 10 22"
                }
            ],
            "utt2spk": "fash"
        },
        "fbbh-an86-b": {
            "input": [
                {
                    "feat": "/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/dump/train_dev/deltafalse/feats.1.ark:175775",
                    "name": "input1",
                    "shape": [
                        328,
                        83
                    ]
                }
            ],
            "output": [
                {
                    "name": "target1",
                    "shape": [
                        15,
                        30
                    ],
                    "text": "C Z D Z W EIGHT",
                    "token": "C <space> Z <space> D <space> Z <space> W <space> E I G H T",
                    "tokenid": "5 2 28 2 6 2 28 2 25 2 7 11 9 10 22"
                }
            ],
            "utt2spk": "fbbh"
        },
        "fbbh-an87-b": {
            "input": [
                {
                    "feat": "/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/dump/train_dev/deltafalse/feats.1.ark:203696",
                    "name": "input1",
                    "shape": [
                        168,
                        83
                    ]
                }
            ],
            "output": [
                {
                    "name": "target1",
                    "shape": [
                        18,
                        30
                    ],
                    "text": "ENTER SIX TWO FOUR",
                    "token": "E N T E R <space> S I X <space> T W O <space> F O U R",
                    "tokenid": "7 16 22 7 20 2 21 11 26 2 22 25 17 2 8 17 23 20"
                }
            ],
            "utt2spk": "fbbh"
        },
        "fbbh-an88-b": {
            "input": [
                {
                    "feat": "/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/dump/train_dev/deltafalse/feats.1.ark:218337",
                    "name": "input1",
                    "shape": [
                        468,
                        83
                    ]
                }
            ],
            "output": [
                {
                    "name": "target1",
                    "shape": [
                        25,
                        30
                    ],
                    "text": "ERASE O T H F I FIVE ZERO",
                    "token": "E R A S E <space> O <space> T <space> H <space> F <space> I <space> F I V E <space> Z E R O",
                    "tokenid": "7 20 3 21 7 2 17 2 22 2 10 2 8 2 11 2 8 11 24 7 2 28 7 20 17"
                }
            ],
            "utt2spk": "fbbh"
        },
        "fbbh-an89-b": {
            "input": [
                {
                    "feat": "/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/dump/train_dev/deltafalse/feats.1.ark:257878",
                    "name": "input1",
                    "shape": [
                        638,
                        83
                    ]
                }
            ],
            "output": [
                {
                    "name": "target1",
                    "shape": [
                        40,
                        30
                    ],
                    "text": "RUBOUT T G J W B SEVENTY NINE FIFTY NINE",
                    "token": "R U B O U T <space> T <space> G <space> J <space> W <space> B <space> S E V E N T Y <space> N I N E <space> F I F T Y <space> N I N E",
                    "tokenid": "20 23 4 17 23 22 2 22 2 9 2 12 2 25 2 4 2 21 7 24 7 16 22 27 2 16 11 16 7 2 8 11 8 22 27 2 16 11 16 7"
                }
            ],
            "utt2spk": "fbbh"
        },
        "fbbh-an90-b": {
            "input": [
                {
                    "feat": "/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/dump/train_dev/deltafalse/feats.2.ark:12",
                    "name": "input1",
                    "shape": [
                        98,
                        83
                    ]
                }
            ],
            "output": [
                {
                    "name": "target1",
                    "shape": [
                        2,
                        30
                    ],
                    "text": "NO",
                    "token": "N O",
                    "tokenid": "16 17"
                }
            ],
            "utt2spk": "fbbh"
        },
        "fbbh-cen1-b": {
            "input": [
                {
                    "feat": "/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/dump/train_dev/deltafalse/feats.2.ark:8843",
                    "name": "input1",
                    "shape": [
                        238,
                        83
                    ]
                }
            ],
            "output": [
                {
                    "name": "target1",
                    "shape": [
                        11,
                        30
                    ],
                    "text": "H O W E L L",
                    "token": "H <space> O <space> W <space> E <space> L <space> L",
                    "tokenid": "10 2 17 2 25 2 7 2 14 2 14"
                }
            ],
            "utt2spk": "fbbh"
        },
        "fbbh-cen2-b": {
            "input": [
                {
                    "feat": "/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/dump/train_dev/deltafalse/feats.2.ark:29294",
                    "name": "input1",
                    "shape": [
                        298,
                        83
                    ]
                }
            ],
            "output": [
                {
                    "name": "target1",
                    "shape": [
                        13,
                        30
                    ],
                    "text": "B E V E R L Y",
                    "token": "B <space> E <space> V <space> E <space> R <space> L <space> Y",
                    "tokenid": "4 2 7 2 24 2 7 2 20 2 14 2 27"
                }
            ],
            "utt2spk": "fbbh"
        },
        "fbbh-cen3-b": {
            "input": [
                {
                    "feat": "/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/dump/train_dev/deltafalse/feats.2.ark:54725",
                    "name": "input1",
                    "shape": [
                        228,
                        83
                    ]
                }
            ],
            "output": [
                {
                    "name": "target1",
                    "shape": [
                        19,
                        30
                    ],
                    "text": "FIFTY ONE FIFTY SIX",
                    "token": "F I F T Y <space> O N E <space> F I F T Y <space> S I X",
                    "tokenid": "8 11 8 22 27 2 17 16 7 2 8 11 8 22 27 2 21 11 26"
                }
            ],
            "utt2spk": "fbbh"
        },
        "fbbh-cen4-b": {
            "input": [
                {
                    "feat": "/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/dump/train_dev/deltafalse/feats.2.ark:74346",
                    "name": "input1",
                    "shape": [
                        338,
                        83
                    ]
                }
            ],
            "output": [
                {
                    "name": "target1",
                    "shape": [
                        11,
                        30
                    ],
                    "text": "P R I N C E",
                    "token": "P <space> R <space> I <space> N <space> C <space> E",
                    "tokenid": "18 2 20 2 11 2 16 2 5 2 7"
                }
            ],
            "utt2spk": "fbbh"
        },
        "fbbh-cen5-b": {
            "input": [
                {
                    "feat": "/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/dump/train_dev/deltafalse/feats.2.ark:103097",
                    "name": "input1",
                    "shape": [
                        288,
                        83
                    ]
                }
            ],
            "output": [
                {
                    "name": "target1",
                    "shape": [
                        15,
                        30
                    ],
                    "text": "G I B S O N I A",
                    "token": "G <space> I <space> B <space> S <space> O <space> N <space> I <space> A",
                    "tokenid": "9 2 11 2 4 2 21 2 17 2 16 2 11 2 3"
                }
            ],
            "utt2spk": "fbbh"
        },
        "fbbh-cen6-b": {
            "input": [
                {
                    "feat": "/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/dump/train_dev/deltafalse/feats.2.ark:127698",
                    "name": "input1",
                    "shape": [
                        238,
                        83
                    ]
                }
            ],
            "output": [
                {
                    "name": "target1",
                    "shape": [
                        21,
                        30
                    ],
                    "text": "ONE FIVE OH FOUR FOUR",
                    "token": "O N E <space> F I V E <space> O H <space> F O U R <space> F O U R",
                    "tokenid": "17 16 7 2 8 11 24 7 2 17 10 2 8 17 23 20 2 8 17 23 20"
                }
            ],
            "utt2spk": "fbbh"
        },
        "fbbh-cen7-b": {
            "input": [
                {
                    "feat": "/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/dump/train_dev/deltafalse/feats.2.ark:148149",
                    "name": "input1",
                    "shape": [
                        308,
                        83
                    ]
                }
            ],
            "output": [
                {
                    "name": "target1",
                    "shape": [
                        33,
                        30
                    ],
                    "text": "FOUR FOUR THREE SIX THREE ONE TWO",
                    "token": "F O U R <space> F O U R <space> T H R E E <space> S I X <space> T H R E E <space> O N E <space> T W O",
                    "tokenid": "8 17 23 20 2 8 17 23 20 2 22 10 20 7 7 2 21 11 26 2 22 10 20 7 7 2 17 16 7 2 22 25 17"
                }
            ],
            "utt2spk": "fbbh"
        },
        "fbbh-cen8-b": {
            "input": [
                {
                    "feat": "/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/dump/train_dev/deltafalse/feats.2.ark:174410",
                    "name": "input1",
                    "shape": [
                        278,
                        83
                    ]
                }
            ],
            "output": [
                {
                    "name": "target1",
                    "shape": [
                        33,
                        30
                    ],
                    "text": "MARCH THIRD NINETEEN TWENTY EIGHT",
                    "token": "M A R C H <space> T H I R D <space> N I N E T E E N <space> T W E N T Y <space> E I G H T",
                    "tokenid": "15 3 20 5 10 2 22 10 11 20 6 2 16 11 16 7 22 7 7 16 2 22 25 7 16 22 27 2 7 11 9 10 22"
                }
            ],
            "utt2spk": "fbbh"
        },
        "fclc-an146-b": {
            "input": [
                {
                    "feat": "/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/dump/train_dev/deltafalse/feats.2.ark:198182",
                    "name": "input1",
                    "shape": [
                        338,
                        83
                    ]
                }
            ],
            "output": [
                {
                    "name": "target1",
                    "shape": [
                        24,
                        30
                    ],
                    "text": "N L N S ONE SEVENTY FIVE",
                    "token": "N <space> L <space> N <space> S <space> O N E <space> S E V E N T Y <space> F I V E",
                    "tokenid": "16 2 14 2 16 2 21 2 17 16 7 2 21 7 24 7 16 22 27 2 8 11 24 7"
                }
            ],
            "utt2spk": "fclc"
        },
        "fclc-an147-b": {
            "input": [
                {
                    "feat": "/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/dump/train_dev/deltafalse/feats.2.ark:226934",
                    "name": "input1",
                    "shape": [
                        288,
                        83
                    ]
                }
            ],
            "output": [
                {
                    "name": "target1",
                    "shape": [
                        11,
                        30
                    ],
                    "text": "Q E E A SIX",
                    "token": "Q <space> E <space> E <space> A <space> S I X",
                    "tokenid": "19 2 7 2 7 2 3 2 21 11 26"
                }
            ],
            "utt2spk": "fclc"
        },
        "fclc-an148-b": {
            "input": [
                {
                    "feat": "/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/dump/train_dev/deltafalse/feats.2.ark:251536",
                    "name": "input1",
                    "shape": [
                        398,
                        83
                    ]
                }
            ],
            "output": [
                {
                    "name": "target1",
                    "shape": [
                        27,
                        30
                    ],
                    "text": "ERASE O J T K S THIRTY FIVE",
                    "token": "E R A S E <space> O <space> J <space> T <space> K <space> S <space> T H I R T Y <space> F I V E",
                    "tokenid": "7 20 3 21 7 2 17 2 12 2 22 2 13 2 21 2 22 10 11 20 22 27 2 8 11 24 7"
                }
            ],
            "utt2spk": "fclc"
        },
        "fclc-an149-b": {
            "input": [
                {
                    "feat": "/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/dump/train_dev/deltafalse/feats.2.ark:285268",
                    "name": "input1",
                    "shape": [
                        228,
                        83
                    ]
                }
            ],
            "output": [
                {
                    "name": "target1",
                    "shape": [
                        22,
                        30
                    ],
                    "text": "ENTER NINE EIGHTY NINE",
                    "token": "E N T E R <space> N I N E <space> E I G H T Y <space> N I N E",
                    "tokenid": "7 16 22 7 20 2 16 11 16 7 2 7 11 9 10 22 27 2 16 11 16 7"
                }
            ],
            "utt2spk": "fclc"
        },
        "fclc-an150-b": {
            "input": [
                {
                    "feat": "/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/dump/train_dev/deltafalse/feats.3.ark:13",
                    "name": "input1",
                    "shape": [
                        428,
                        83
                    ]
                }
            ],
            "output": [
                {
                    "name": "target1",
                    "shape": [
                        34,
                        30
                    ],
                    "text": "ERASE C K C W FOURTEEN EIGHTY FIVE",
                    "token": "E R A S E <space> C <space> K <space> C <space> W <space> F O U R T E E N <space> E I G H T Y <space> F I V E",
                    "tokenid": "7 20 3 21 7 2 5 2 13 2 5 2 25 2 8 17 23 20 22 7 7 16 2 7 11 9 10 22 27 2 8 11 24 7"
                }
            ],
            "utt2spk": "fclc"
        },
        "fclc-cen1-b": {
            "input": [
                {
                    "feat": "/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/dump/train_dev/deltafalse/feats.3.ark:36234",
                    "name": "input1",
                    "shape": [
                        428,
                        83
                    ]
                }
            ],
            "output": [
                {
                    "name": "target1",
                    "shape": [
                        15,
                        30
                    ],
                    "text": "C O N N E L L Y",
                    "token": "C <space> O <space> N <space> N <space> E <space> L <space> L <space> Y",
                    "tokenid": "5 2 17 2 16 2 16 2 7 2 14 2 14 2 27"
                }
            ],
            "utt2spk": "fclc"
        },
        "fclc-cen2-b": {
            "input": [
                {
                    "feat": "/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/dump/train_dev/deltafalse/feats.3.ark:72455",
                    "name": "input1",
                    "shape": [
                        248,
                        83
                    ]
                }
            ],
            "output": [
                {
                    "name": "target1",
                    "shape": [
                        9,
                        30
                    ],
                    "text": "C I N D Y",
                    "token": "C <space> I <space> N <space> D <space> Y",
                    "tokenid": "5 2 11 2 16 2 6 2 27"
                }
            ],
            "utt2spk": "fclc"
        },
        "fclc-cen3-b": {
            "input": [
                {
                    "feat": "/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/dump/train_dev/deltafalse/feats.3.ark:93736",
                    "name": "input1",
                    "shape": [
                        188,
                        83
                    ]
                }
            ],
            "output": [
                {
                    "name": "target1",
                    "shape": [
                        20,
                        30
                    ],
                    "text": "FIVE FIVE THREE FIVE",
                    "token": "F I V E <space> F I V E <space> T H R E E <space> F I V E",
                    "tokenid": "8 11 24 7 2 8 11 24 7 2 22 10 20 7 7 2 8 11 24 7"
                }
            ],
            "utt2spk": "fclc"
        },
        "fclc-cen4-b": {
            "input": [
                {
                    "feat": "/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/dump/train_dev/deltafalse/feats.3.ark:110037",
                    "name": "input1",
                    "shape": [
                        368,
                        83
                    ]
                }
            ],
            "output": [
                {
                    "name": "target1",
                    "shape": [
                        17,
                        30
                    ],
                    "text": "A Y L E S B O R O",
                    "token": "A <space> Y <space> L <space> E <space> S <space> B <space> O <space> R <space> O",
                    "tokenid": "3 2 27 2 14 2 7 2 21 2 4 2 17 2 20 2 17"
                }
            ],
            "utt2spk": "fclc"
        },
        "fclc-cen5-b": {
            "input": [
                {
                    "feat": "/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/dump/train_dev/deltafalse/feats.3.ark:141278",
                    "name": "input1",
                    "shape": [
                        428,
                        83
                    ]
                }
            ],
            "output": [
                {
                    "name": "target1",
                    "shape": [
                        19,
                        30
                    ],
                    "text": "P I T T S B U R G H",
                    "token": "P <space> I <space> T <space> T <space> S <space> B <space> U <space> R <space> G <space> H",
                    "tokenid": "18 2 11 2 22 2 22 2 21 2 4 2 23 2 20 2 9 2 10"
                }
            ],
            "utt2spk": "fclc"
        },
        "fclc-cen6-b": {
            "input": [
                {
                    "feat": "/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/dump/train_dev/deltafalse/feats.3.ark:177499",
                    "name": "input1",
                    "shape": [
                        208,
                        83
                    ]
                }
            ],
            "output": [
                {
                    "name": "target1",
                    "shape": [
                        22,
                        30
                    ],
                    "text": "ONE FIVE TWO ONE SEVEN",
                    "token": "O N E <space> F I V E <space> T W O <space> O N E <space> S E V E N",
                    "tokenid": "17 16 7 2 8 11 24 7 2 22 25 17 2 17 16 7 2 21 7 24 7 16"
                }
            ],
            "utt2spk": "fclc"
        },
        "fclc-cen7-b": {
            "input": [
                {
                    "feat": "/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/dump/train_dev/deltafalse/feats.3.ark:195460",
                    "name": "input1",
                    "shape": [
                        298,
                        83
                    ]
                }
            ],
            "output": [
                {
                    "name": "target1",
                    "shape": [
                        35,
                        30
                    ],
                    "text": "SIX EIGHT SEVEN FIVE EIGHT OH EIGHT",
                    "token": "S I X <space> E I G H T <space> S E V E N <space> F I V E <space> E I G H T <space> O H <space> E I G H T",
                    "tokenid": "21 11 26 2 7 11 9 10 22 2 21 7 24 7 16 2 8 11 24 7 2 7 11 9 10 22 2 17 10 2 7 11 9 10 22"
                }
            ],
            "utt2spk": "fclc"
        },
        "fclc-cen8-b": {
            "input": [
                {
                    "feat": "/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/dump/train_dev/deltafalse/feats.3.ark:220891",
                    "name": "input1",
                    "shape": [
                        268,
                        83
                    ]
                }
            ],
            "output": [
                {
                    "name": "target1",
                    "shape": [
                        34,
                        30
                    ],
                    "text": "JUNE ELEVENTH NINETEEN SIXTY SEVEN",
                    "token": "J U N E <space> E L E V E N T H <space> N I N E T E E N <space> S I X T Y <space> S E V E N",
                    "tokenid": "12 23 16 7 2 7 14 7 24 7 16 22 10 2 16 11 16 7 22 7 7 16 2 21 11 26 22 27 2 21 7 24 7 16"
                }
            ],
            "utt2spk": "fclc"
        },
        "fejs-an36-b": {
            "input": [
                {
                    "feat": "/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/dump/train_dev/deltafalse/feats.3.ark:243832",
                    "name": "input1",
                    "shape": [
                        378,
                        83
                    ]
                }
            ],
            "output": [
                {
                    "name": "target1",
                    "shape": [
                        26,
                        30
                    ],
                    "text": "RUBOUT C Y R B SEVEN EIGHT",
                    "token": "R U B O U T <space> C <space> Y <space> R <space> B <space> S E V E N <space> E I G H T",
                    "tokenid": "20 23 4 17 23 22 2 5 2 27 2 20 2 4 2 21 7 24 7 16 2 7 11 9 10 22"
                }
            ],
            "utt2spk": "fejs"
        },
        "fejs-an37-b": {
            "input": [
                {
                    "feat": "/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/dump/train_dev/deltafalse/feats.3.ark:275903",
                    "name": "input1",
                    "shape": [
                        468,
                        83
                    ]
                }
            ],
            "output": [
                {
                    "name": "target1",
                    "shape": [
                        27,
                        30
                    ],
                    "text": "ERASE J G Q Y EIGHT SIX ONE",
                    "token": "E R A S E <space> J <space> G <space> Q <space> Y <space> E I G H T <space> S I X <space> O N E",
                    "tokenid": "7 20 3 21 7 2 12 2 9 2 19 2 27 2 7 11 9 10 22 2 21 11 26 2 17 16 7"
                }
            ],
            "utt2spk": "fejs"
        },
        "fejs-an38-b": {
            "input": [
                {
                    "feat": "/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/dump/train_dev/deltafalse/feats.3.ark:315444",
                    "name": "input1",
                    "shape": [
                        78,
                        83
                    ]
                }
            ],
            "output": [
                {
                    "name": "target1",
                    "shape": [
                        2,
                        30
                    ],
                    "text": "GO",
                    "token": "G O",
                    "tokenid": "9 17"
                }
            ],
            "utt2spk": "fejs"
        },
        "fejs-an39-b": {
            "input": [
                {
                    "feat": "/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/dump/train_dev/deltafalse/feats.3.ark:322615",
                    "name": "input1",
                    "shape": [
                        268,
                        83
                    ]
                }
            ],
            "output": [
                {
                    "name": "target1",
                    "shape": [
                        12,
                        30
                    ],
                    "text": "G C D O FOUR",
                    "token": "G <space> C <space> D <space> O <space> F O U R",
                    "tokenid": "9 2 5 2 6 2 17 2 8 17 23 20"
                }
            ],
            "utt2spk": "fejs"
        },
        "fejs-an40-b": {
            "input": [
                {
                    "feat": "/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/dump/train_dev/deltafalse/feats.4.ark:12",
                    "name": "input1",
                    "shape": [
                        308,
                        83
                    ]
                }
            ],
            "output": [
                {
                    "name": "target1",
                    "shape": [
                        21,
                        30
                    ],
                    "text": "Z I R K SIX FOUR FOUR",
                    "token": "Z <space> I <space> R <space> K <space> S I X <space> F O U R <space> F O U R",
                    "tokenid": "28 2 11 2 20 2 13 2 21 11 26 2 8 17 23 20 2 8 17 23 20"
                }
            ],
            "utt2spk": "fejs"
        },
        "fejs-cen1-b": {
            "input": [
                {
                    "feat": "/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/dump/train_dev/deltafalse/feats.4.ark:26273",
                    "name": "input1",
                    "shape": [
                        438,
                        83
                    ]
                }
            ],
            "output": [
                {
                    "name": "target1",
                    "shape": [
                        17,
                        30
                    ],
                    "text": "S C A V I N C K Y",
                    "token": "S <space> C <space> A <space> V <space> I <space> N <space> C <space> K <space> Y",
                    "tokenid": "21 2 5 2 3 2 24 2 11 2 16 2 5 2 13 2 27"
                }
            ],
            "utt2spk": "fejs"
        },
        "fejs-cen2-b": {
            "input": [
                {
                    "feat": "/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/dump/train_dev/deltafalse/feats.4.ark:63324",
                    "name": "input1",
                    "shape": [
                        228,
                        83
                    ]
                }
            ],
            "output": [
                {
                    "name": "target1",
                    "shape": [
                        7,
                        30
                    ],
                    "text": "J E A N",
                    "token": "J <space> E <space> A <space> N",
                    "tokenid": "12 2 7 2 3 2 16"
                }
            ],
            "utt2spk": "fejs"
        },
        "fejs-cen3-b": {
            "input": [
                {
                    "feat": "/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/dump/train_dev/deltafalse/feats.4.ark:82945",
                    "name": "input1",
                    "shape": [
                        278,
                        83
                    ]
                }
            ],
            "output": [
                {
                    "name": "target1",
                    "shape": [
                        17,
                        30
                    ],
                    "text": "ONE ZERO SIX FIVE",
                    "token": "O N E <space> Z E R O <space> S I X <space> F I V E",
                    "tokenid": "17 16 7 2 28 7 20 17 2 21 11 26 2 8 11 24 7"
                }
            ],
            "utt2spk": "fejs"
        },
        "fejs-cen4-b": {
            "input": [
                {
                    "feat": "/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/dump/train_dev/deltafalse/feats.4.ark:106716",
                    "name": "input1",
                    "shape": [
                        528,
                        83
                    ]
                }
            ],
            "output": [
                {
                    "name": "target1",
                    "shape": [
                        23,
                        30
                    ],
                    "text": "F I N D L E Y D R I V E",
                    "token": "F <space> I <space> N <space> D <space> L <space> E <space> Y <space> D <space> R <space> I <space> V <space> E",
                    "tokenid": "8 2 11 2 16 2 6 2 14 2 7 2 27 2 6 2 20 2 11 2 24 2 7"
                }
            ],
            "utt2spk": "fejs"
        },
        "fejs-cen5-b": {
            "input": [
                {
                    "feat": "/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/dump/train_dev/deltafalse/feats.4.ark:151237",
                    "name": "input1",
                    "shape": [
                        408,
                        83
                    ]
                }
            ],
            "output": [
                {
                    "name": "target1",
                    "shape": [
                        19,
                        30
                    ],
                    "text": "P I T T S B U R G H",
                    "token": "P <space> I <space> T <space> T <space> S <space> B <space> U <space> R <space> G <space> H",
                    "tokenid": "18 2 11 2 22 2 22 2 21 2 4 2 23 2 20 2 9 2 10"
                }
            ],
            "utt2spk": "fejs"
        },
        "fejs-cen6-b": {
            "input": [
                {
                    "feat": "/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/dump/train_dev/deltafalse/feats.4.ark:185798",
                    "name": "input1",
                    "shape": [
                        198,
                        83
                    ]
                }
            ],
            "output": [
                {
                    "name": "target1",
                    "shape": [
                        20,
                        30
                    ],
                    "text": "ONE FIVE TWO TWO ONE",
                    "token": "O N E <space> F I V E <space> T W O <space> T W O <space> O N E",
                    "tokenid": "17 16 7 2 8 11 24 7 2 22 25 17 2 22 25 17 2 17 16 7"
                }
            ],
            "utt2spk": "fejs"
        },
        "fejs-cen7-b": {
            "input": [
                {
                    "feat": "/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/dump/train_dev/deltafalse/feats.4.ark:202929",
                    "name": "input1",
                    "shape": [
                        298,
                        83
                    ]
                }
            ],
            "output": [
                {
                    "name": "target1",
                    "shape": [
                        34,
                        30
                    ],
                    "text": "TWO SIX EIGHT THREE EIGHT ZERO TWO",
                    "token": "T W O <space> S I X <space> E I G H T <space> T H R E E <space> E I G H T <space> Z E R O <space> T W O",
                    "tokenid": "22 25 17 2 21 11 26 2 7 11 9 10 22 2 22 10 20 7 7 2 7 11 9 10 22 2 28 7 20 17 2 22 25 17"
                }
            ],
            "utt2spk": "fejs"
        },
        "fejs-cen8-b": {
            "input": [
                {
                    "feat": "/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/dump/train_dev/deltafalse/feats.4.ark:228360",
                    "name": "input1",
                    "shape": [
                        218,
                        83
                    ]
                }
            ],
            "output": [
                {
                    "name": "target1",
                    "shape": [
                        21,
                        30
                    ],
                    "text": "TWO FIFTEEN SIXTY TWO",
                    "token": "T W O <space> F I F T E E N <space> S I X T Y <space> T W O",
                    "tokenid": "22 25 17 2 8 11 8 22 7 7 16 2 21 11 26 22 27 2 22 25 17"
                }
            ],
            "utt2spk": "fejs"
        },
        "ffmm-an291-b": {
            "input": [
                {
                    "feat": "/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/dump/train_dev/deltafalse/feats.4.ark:247152",
                    "name": "input1",
                    "shape": [
                        148,
                        83
                    ]
                }
            ],
            "output": [
                {
                    "name": "target1",
                    "shape": [
                        10,
                        30
                    ],
                    "text": "ENTER FIVE",
                    "token": "E N T E R <space> F I V E",
                    "tokenid": "7 16 22 7 20 2 8 11 24 7"
                }
            ],
            "utt2spk": "ffmm"
        },
        "ffmm-an292-b": {
            "input": [
                {
                    "feat": "/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/dump/train_dev/deltafalse/feats.4.ark:260134",
                    "name": "input1",
                    "shape": [
                        168,
                        83
                    ]
                }
            ],
            "output": [
                {
                    "name": "target1",
                    "shape": [
                        19,
                        30
                    ],
                    "text": "ENTER SEVENTY EIGHT",
                    "token": "E N T E R <space> S E V E N T Y <space> E I G H T",
                    "tokenid": "7 16 22 7 20 2 21 7 24 7 16 22 27 2 7 11 9 10 22"
                }
            ],
            "utt2spk": "ffmm"
        },
        "ffmm-an293-b": {
            "input": [
                {
                    "feat": "/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/dump/train_dev/deltafalse/feats.4.ark:274776",
                    "name": "input1",
                    "shape": [
                        338,
                        83
                    ]
                }
            ],
            "output": [
                {
                    "name": "target1",
                    "shape": [
                        27,
                        30
                    ],
                    "text": "K L M U FIVE TWO SEVEN FOUR",
                    "token": "K <space> L <space> M <space> U <space> F I V E <space> T W O <space> S E V E N <space> F O U R",
                    "tokenid": "13 2 14 2 15 2 23 2 8 11 24 7 2 22 25 17 2 21 7 24 7 16 2 8 17 23 20"
                }
            ],
            "utt2spk": "ffmm"
        },
        "ffmm-an294-b": {
            "input": [
                {
                    "feat": "/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/dump/train_dev/deltafalse/feats.4.ark:303528",
                    "name": "input1",
                    "shape": [
                        268,
                        83
                    ]
                }
            ],
            "output": [
                {
                    "name": "target1",
                    "shape": [
                        21,
                        30
                    ],
                    "text": "H B G L SEVENTY SEVEN",
                    "token": "H <space> B <space> G <space> L <space> S E V E N T Y <space> S E V E N",
                    "tokenid": "10 2 4 2 9 2 14 2 21 7 24 7 16 22 27 2 21 7 24 7 16"
                }
            ],
            "utt2spk": "ffmm"
        },
        "ffmm-an295-b": {
            "input": [
                {
                    "feat": "/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/dump/train_dev/deltafalse/feats.5.ark:13",
                    "name": "input1",
                    "shape": [
                        288,
                        83
                    ]
                }
            ],
            "output": [
                {
                    "name": "target1",
                    "shape": [
                        27,
                        30
                    ],
                    "text": "V A D R THIRTY ONE OH SEVEN",
                    "token": "V <space> A <space> D <space> R <space> T H I R T Y <space> O N E <space> O H <space> S E V E N",
                    "tokenid": "24 2 3 2 6 2 20 2 22 10 11 20 22 27 2 17 16 7 2 17 10 2 21 7 24 7 16"
                }
            ],
            "utt2spk": "ffmm"
        },
        "ffmm-cen1-b": {
            "input": [
                {
                    "feat": "/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/dump/train_dev/deltafalse/feats.5.ark:24614",
                    "name": "input1",
                    "shape": [
                        328,
                        83
                    ]
                }
            ],
            "output": [
                {
                    "name": "target1",
                    "shape": [
                        13,
                        30
                    ],
                    "text": "M O D U G N O",
                    "token": "M <space> O <space> D <space> U <space> G <space> N <space> O",
                    "tokenid": "15 2 17 2 6 2 23 2 9 2 16 2 17"
                }
            ],
            "utt2spk": "ffmm"
        },
        "ffmm-cen2-b": {
            "input": [
                {
                    "feat": "/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/dump/train_dev/deltafalse/feats.5.ark:52535",
                    "name": "input1",
                    "shape": [
                        498,
                        83
                    ]
                }
            ],
            "output": [
                {
                    "name": "target1",
                    "shape": [
                        21,
                        30
                    ],
                    "text": "F R A N C E S M A R Y",
                    "token": "F <space> R <space> A <space> N <space> C <space> E <space> S <space> M <space> A <space> R <space> Y",
                    "tokenid": "8 2 20 2 3 2 16 2 5 2 7 2 21 2 15 2 3 2 20 2 27"
                }
            ],
            "utt2spk": "ffmm"
        },
        "ffmm-cen3-b": {
            "input": [
                {
                    "feat": "/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/dump/train_dev/deltafalse/feats.5.ark:94566",
                    "name": "input1",
                    "shape": [
                        198,
                        83
                    ]
                }
            ],
            "output": [
                {
                    "name": "target1",
                    "shape": [
                        20,
                        30
                    ],
                    "text": "FIFTY SIX THIRTY TWO",
                    "token": "F I F T Y <space> S I X <space> T H I R T Y <space> T W O",
                    "tokenid": "8 11 8 22 27 2 21 11 26 2 22 10 11 20 22 27 2 22 25 17"
                }
            ],
            "utt2spk": "ffmm"
        },
        "ffmm-cen4-b": {
            "input": [
                {
                    "feat": "/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/dump/train_dev/deltafalse/feats.5.ark:111697",
                    "name": "input1",
                    "shape": [
                        428,
                        83
                    ]
                }
            ],
            "output": [
                {
                    "name": "target1",
                    "shape": [
                        15,
                        30
                    ],
                    "text": "P H I L L I P S",
                    "token": "P <space> H <space> I <space> L <space> L <space> I <space> P <space> S",
                    "tokenid": "18 2 10 2 11 2 14 2 14 2 11 2 18 2 21"
                }
            ],
            "utt2spk": "ffmm"
        },
        "ffmm-cen5-b": {
            "input": [
                {
                    "feat": "/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/dump/train_dev/deltafalse/feats.5.ark:147918",
                    "name": "input1",
                    "shape": [
                        488,
                        83
                    ]
                }
            ],
            "output": [
                {
                    "name": "target1",
                    "shape": [
                        19,
                        30
                    ],
                    "text": "P I T T S B U R G H",
                    "token": "P <space> I <space> T <space> T <space> S <space> B <space> U <space> R <space> G <space> H",
                    "tokenid": "18 2 11 2 22 2 22 2 21 2 4 2 23 2 20 2 9 2 10"
                }
            ],
            "utt2spk": "ffmm"
        },
        "ffmm-cen6-b": {
            "input": [
                {
                    "feat": "/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/dump/train_dev/deltafalse/feats.5.ark:189119",
                    "name": "input1",
                    "shape": [
                        228,
                        83
                    ]
                }
            ],
            "output": [
                {
                    "name": "target1",
                    "shape": [
                        22,
                        30
                    ],
                    "text": "ONE FIVE TWO ONE SEVEN",
                    "token": "O N E <space> F I V E <space> T W O <space> O N E <space> S E V E N",
                    "tokenid": "17 16 7 2 8 11 24 7 2 22 25 17 2 17 16 7 2 21 7 24 7 16"
                }
            ],
            "utt2spk": "ffmm"
        },
        "ffmm-cen7-b": {
            "input": [
                {
                    "feat": "/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/dump/train_dev/deltafalse/feats.5.ark:208740",
                    "name": "input1",
                    "shape": [
                        488,
                        83
                    ]
                }
            ],
            "output": [
                {
                    "name": "target1",
                    "shape": [
                        46,
                        30
                    ],
                    "text": "FOUR ONE TWO FOUR TWO ONE EIGHT EIGHT NINE SIX",
                    "token": "F O U R <space> O N E <space> T W O <space> F O U R <space> T W O <space> O N E <space> E I G H T <space> E I G H T <space> N I N E <space> S I X",
                    "tokenid": "8 17 23 20 2 17 16 7 2 22 25 17 2 8 17 23 20 2 22 25 17 2 17 16 7 2 7 11 9 10 22 2 7 11 9 10 22 2 16 11 16 7 2 21 11 26"
                }
            ],
            "utt2spk": "ffmm"
        },
        "ffmm-cen8-b": {
            "input": [
                {
                    "feat": "/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/dump/train_dev/deltafalse/feats.5.ark:249941",
                    "name": "input1",
                    "shape": [
                        348,
                        83
                    ]
                }
            ],
            "output": [
                {
                    "name": "target1",
                    "shape": [
                        39,
                        30
                    ],
                    "text": "SEPTEMBER FIFTEENTH NINETEEN SIXTY FIVE",
                    "token": "S E P T E M B E R <space> F I F T E E N T H <space> N I N E T E E N <space> S I X T Y <space> F I V E",
                    "tokenid": "21 7 18 22 7 15 4 7 20 2 8 11 8 22 7 7 16 22 10 2 16 11 16 7 22 7 7 16 2 21 11 26 22 27 2 8 11 24 7"
                }
            ],
            "utt2spk": "ffmm"
        },
        "fjam-an76-b": {
            "input": [
                {
                    "feat": "/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/dump/train_dev/deltafalse/feats.5.ark:279522",
                    "name": "input1",
                    "shape": [
                        88,
                        83
                    ]
                }
            ],
            "output": [
                {
                    "name": "target1",
                    "shape": [
                        3,
                        30
                    ],
                    "text": "YES",
                    "token": "Y E S",
                    "tokenid": "27 7 21"
                }
            ],
            "utt2spk": "fjam"
        },
        "fjam-an77-b": {
            "input": [
                {
                    "feat": "/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/dump/train_dev/deltafalse/feats.5.ark:287523",
                    "name": "input1",
                    "shape": [
                        308,
                        83
                    ]
                }
            ],
            "output": [
                {
                    "name": "target1",
                    "shape": [
                        32,
                        30
                    ],
                    "text": "ERASE V X S X TWO FOUR NINE NINE",
                    "token": "E R A S E <space> V <space> X <space> S <space> X <space> T W O <space> F O U R <space> N I N E <space> N I N E",
                    "tokenid": "7 20 3 21 7 2 24 2 26 2 21 2 26 2 22 25 17 2 8 17 23 20 2 16 11 16 7 2 16 11 16 7"
                }
            ],
            "utt2spk": "fjam"
        },
        "fjam-an78-b": {
            "input": [
                {
                    "feat": "/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/dump/train_dev/deltafalse/feats.5.ark:313784",
                    "name": "input1",
                    "shape": [
                        198,
                        83
                    ]
                }
            ],
            "output": [
                {
                    "name": "target1",
                    "shape": [
                        12,
                        30
                    ],
                    "text": "R Q Z H FOUR",
                    "token": "R <space> Q <space> Z <space> H <space> F O U R",
                    "tokenid": "20 2 19 2 28 2 10 2 8 17 23 20"
                }
            ],
            "utt2spk": "fjam"
        },
        "fjam-an79-b": {
            "input": [
                {
                    "feat": "/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/dump/train_dev/deltafalse/feats.6.ark:12",
                    "name": "input1",
                    "shape": [
                        308,
                        83
                    ]
                }
            ],
            "output": [
                {
                    "name": "target1",
                    "shape": [
                        30,
                        30
                    ],
                    "text": "RUBOUT J U I P THREE TWO EIGHT",
                    "token": "R U B O U T <space> J <space> U <space> I <space> P <space> T H R E E <space> T W O <space> E I G H T",
                    "tokenid": "20 23 4 17 23 22 2 12 2 23 2 11 2 18 2 22 10 20 7 7 2 22 25 17 2 7 11 9 10 22"
                }
            ],
            "utt2spk": "fjam"
        },
        "fjam-an80-b": {
            "input": [
                {
                    "feat": "/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/dump/train_dev/deltafalse/feats.6.ark:26273",
                    "name": "input1",
                    "shape": [
                        88,
                        83
                    ]
                }
            ],
            "output": [
                {
                    "name": "target1",
                    "shape": [
                        4,
                        30
                    ],
                    "text": "STOP",
                    "token": "S T O P",
                    "tokenid": "21 22 17 18"
                }
            ],
            "utt2spk": "fjam"
        },
        "fjam-cen1-b": {
            "input": [
                {
                    "feat": "/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/dump/train_dev/deltafalse/feats.6.ark:34274",
                    "name": "input1",
                    "shape": [
                        238,
                        83
                    ]
                }
            ],
            "output": [
                {
                    "name": "target1",
                    "shape": [
                        13,
                        30
                    ],
                    "text": "M A R N E L L",
                    "token": "M <space> A <space> R <space> N <space> E <space> L <space> L",
                    "tokenid": "15 2 3 2 20 2 16 2 7 2 14 2 14"
                }
            ],
            "utt2spk": "fjam"
        },
        "fjam-cen2-b": {
            "input": [
                {
                    "feat": "/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/dump/train_dev/deltafalse/feats.6.ark:54725",
                    "name": "input1",
                    "shape": [
                        148,
                        83
                    ]
                }
            ],
            "output": [
                {
                    "name": "target1",
                    "shape": [
                        9,
                        30
                    ],
                    "text": "J U L I E",
                    "token": "J <space> U <space> L <space> I <space> E",
                    "tokenid": "12 2 23 2 14 2 11 2 7"
                }
            ],
            "utt2spk": "fjam"
        },
        "fjam-cen3-b": {
            "input": [
                {
                    "feat": "/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/dump/train_dev/deltafalse/feats.6.ark:67706",
                    "name": "input1",
                    "shape": [
                        178,
                        83
                    ]
                }
            ],
            "output": [
                {
                    "name": "target1",
                    "shape": [
                        16,
                        30
                    ],
                    "text": "SEVEN FIFTY FIVE",
                    "token": "S E V E N <space> F I F T Y <space> F I V E",
                    "tokenid": "21 7 24 7 16 2 8 11 8 22 27 2 8 11 24 7"
                }
            ],
            "utt2spk": "fjam"
        },
        "fjam-cen4-b": {
            "input": [
                {
                    "feat": "/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/dump/train_dev/deltafalse/feats.6.ark:83177",
                    "name": "input1",
                    "shape": [
                        328,
                        83
                    ]
                }
            ],
            "output": [
                {
                    "name": "target1",
                    "shape": [
                        19,
                        30
                    ],
                    "text": "M E M O R Y L A N E",
                    "token": "M <space> E <space> M <space> O <space> R <space> Y <space> L <space> A <space> N <space> E",
                    "tokenid": "15 2 7 2 15 2 17 2 20 2 27 2 14 2 3 2 16 2 7"
                }
            ],
            "utt2spk": "fjam"
        },
        "fjam-cen5-b": {
            "input": [
                {
                    "feat": "/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/dump/train_dev/deltafalse/feats.6.ark:111098",
                    "name": "input1",
                    "shape": [
                        368,
                        83
                    ]
                }
            ],
            "output": [
                {
                    "name": "target1",
                    "shape": [
                        19,
                        30
                    ],
                    "text": "M C K E E S P O R T",
                    "token": "M <space> C <space> K <space> E <space> E <space> S <space> P <space> O <space> R <space> T",
                    "tokenid": "15 2 5 2 13 2 7 2 7 2 21 2 18 2 17 2 20 2 22"
                }
            ],
            "utt2spk": "fjam"
        },
        "fjam-cen6-b": {
            "input": [
                {
                    "feat": "/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/dump/train_dev/deltafalse/feats.6.ark:142339",
                    "name": "input1",
                    "shape": [
                        288,
                        83
                    ]
                }
            ],
            "output": [
                {
                    "name": "target1",
                    "shape": [
                        24,
                        30
                    ],
                    "text": "ONE FIVE ONE THREE THREE",
                    "token": "O N E <space> F I V E <space> O N E <space> T H R E E <space> T H R E E",
                    "tokenid": "17 16 7 2 8 11 24 7 2 17 16 7 2 22 10 20 7 7 2 22 10 20 7 7"
                }
            ],
            "utt2spk": "fjam"
        },
        "fjam-cen7-b": {
            "input": [
                {
                    "feat": "/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/dump/train_dev/deltafalse/feats.6.ark:166940",
                    "name": "input1",
                    "shape": [
                        238,
                        83
                    ]
                }
            ],
            "output": [
                {
                    "name": "target1",
                    "shape": [
                        33,
                        30
                    ],
                    "text": "SIX SIX FOUR SEVEN TWO NINE THREE",
                    "token": "S I X <space> S I X <space> F O U R <space> S E V E N <space> T W O <space> N I N E <space> T H R E E",
                    "tokenid": "21 11 26 2 21 11 26 2 8 17 23 20 2 21 7 24 7 16 2 22 25 17 2 16 11 16 7 2 22 10 20 7 7"
                }
            ],
            "utt2spk": "fjam"
        },
        "fjam-cen8-b": {
            "input": [
                {
                    "feat": "/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/dump/train_dev/deltafalse/feats.6.ark:187391",
                    "name": "input1",
                    "shape": [
                        208,
                        83
                    ]
                }
            ],
            "output": [
                {
                    "name": "target1",
                    "shape": [
                        31,
                        30
                    ],
                    "text": "MAY TWENTY FIRST NINETEEN SIXTY",
                    "token": "M A Y <space> T W E N T Y <space> F I R S T <space> N I N E T E E N <space> S I X T Y",
                    "tokenid": "15 3 27 2 22 25 7 16 22 27 2 8 11 20 21 22 2 16 11 16 7 22 7 7 16 2 21 11 26 22 27"
                }
            ],
            "utt2spk": "fjam"
        },
        "fjdn-an121-b": {
            "input": [
                {
                    "feat": "/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/dump/train_dev/deltafalse/feats.6.ark:205353",
                    "name": "input1",
                    "shape": [
                        68,
                        83
                    ]
                }
            ],
            "output": [
                {
                    "name": "target1",
                    "shape": [
                        4,
                        30
                    ],
                    "text": "HELP",
                    "token": "H E L P",
                    "tokenid": "10 7 14 18"
                }
            ],
            "utt2spk": "fjdn"
        },
        "fjdn-an122-b": {
            "input": [
                {
                    "feat": "/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/dump/train_dev/deltafalse/feats.6.ark:211695",
                    "name": "input1",
                    "shape": [
                        158,
                        83
                    ]
                }
            ],
            "output": [
                {
                    "name": "target1",
                    "shape": [
                        24,
                        30
                    ],
                    "text": "ENTER EIGHT TWO FOUR TWO",
                    "token": "E N T E R <space> E I G H T <space> T W O <space> F O U R <space> T W O",
                    "tokenid": "7 16 22 7 20 2 7 11 9 10 22 2 22 25 17 2 8 17 23 20 2 22 25 17"
                }
            ],
            "utt2spk": "fjdn"
        },
        "fjdn-an123-b": {
            "input": [
                {
                    "feat": "/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/dump/train_dev/deltafalse/feats.7.ark:13",
                    "name": "input1",
                    "shape": [
                        128,
                        83
                    ]
                }
            ],
            "output": [
                {
                    "name": "target1",
                    "shape": [
                        16,
                        30
                    ],
                    "text": "ENTER FIFTY FOUR",
                    "token": "E N T E R <space> F I F T Y <space> F O U R",
                    "tokenid": "7 16 22 7 20 2 8 11 8 22 27 2 8 17 23 20"
                }
            ],
            "utt2spk": "fjdn"
        },
        "fjdn-an124-b": {
            "input": [
                {
                    "feat": "/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/dump/train_dev/deltafalse/feats.7.ark:11335",
                    "name": "input1",
                    "shape": [
                        268,
                        83
                    ]
                }
            ],
            "output": [
                {
                    "name": "target1",
                    "shape": [
                        27,
                        30
                    ],
                    "text": "RUBOUT O Q I M I THIRTY TWO",
                    "token": "R U B O U T <space> O <space> Q <space> I <space> M <space> I <space> T H I R T Y <space> T W O",
                    "tokenid": "20 23 4 17 23 22 2 17 2 19 2 11 2 15 2 11 2 22 10 11 20 22 27 2 22 25 17"
                }
            ],
            "utt2spk": "fjdn"
        },
        "fjdn-an125-b": {
            "input": [
                {
                    "feat": "/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/dump/train_dev/deltafalse/feats.7.ark:34277",
                    "name": "input1",
                    "shape": [
                        108,
                        83
                    ]
                }
            ],
            "output": [
                {
                    "name": "target1",
                    "shape": [
                        13,
                        30
                    ],
                    "text": "ENTER FIFTEEN",
                    "token": "E N T E R <space> F I F T E E N",
                    "tokenid": "7 16 22 7 20 2 8 11 8 22 7 7 16"
                }
            ],
            "utt2spk": "fjdn"
        },
        "fjdn-cen1-b": {
            "input": [
                {
                    "feat": "/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/dump/train_dev/deltafalse/feats.7.ark:43938",
                    "name": "input1",
                    "shape": [
                        148,
                        83
                    ]
                }
            ],
            "output": [
                {
                    "name": "target1",
                    "shape": [
                        11,
                        30
                    ],
                    "text": "N E L S O N",
                    "token": "N <space> E <space> L <space> S <space> O <space> N",
                    "tokenid": "16 2 7 2 14 2 21 2 17 2 16"
                }
            ],
            "utt2spk": "fjdn"
        },
        "fjdn-cen2-b": {
            "input": [
                {
                    "feat": "/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/dump/train_dev/deltafalse/feats.7.ark:56919",
                    "name": "input1",
                    "shape": [
                        178,
                        83
                    ]
                }
            ],
            "output": [
                {
                    "name": "target1",
                    "shape": [
                        15,
                        30
                    ],
                    "text": "J E N N I F E R",
                    "token": "J <space> E <space> N <space> N <space> I <space> F <space> E <space> R",
                    "tokenid": "12 2 7 2 16 2 16 2 11 2 8 2 7 2 20"
                }
            ],
            "utt2spk": "fjdn"
        },
        "fjdn-cen3-b": {
            "input": [
                {
                    "feat": "/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/dump/train_dev/deltafalse/feats.7.ark:72390",
                    "name": "input1",
                    "shape": [
                        98,
                        83
                    ]
                }
            ],
            "output": [
                {
                    "name": "target1",
                    "shape": [
                        12,
                        30
                    ],
                    "text": "THIRTY THREE",
                    "token": "T H I R T Y <space> T H R E E",
                    "tokenid": "22 10 11 20 22 27 2 22 10 20 7 7"
                }
            ],
            "utt2spk": "fjdn"
        },
        "fjdn-cen4-b": {
            "input": [
                {
                    "feat": "/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/dump/train_dev/deltafalse/feats.7.ark:81221",
                    "name": "input1",
                    "shape": [
                        148,
                        83
                    ]
                }
            ],
            "output": [
                {
                    "name": "target1",
                    "shape": [
                        9,
                        30
                    ],
                    "text": "K U N T Z",
                    "token": "K <space> U <space> N <space> T <space> Z",
                    "tokenid": "13 2 23 2 16 2 22 2 28"
                }
            ],
            "utt2spk": "fjdn"
        },
        "fjdn-cen5-b": {
            "input": [
                {
                    "feat": "/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/dump/train_dev/deltafalse/feats.7.ark:94202",
                    "name": "input1",
                    "shape": [
                        228,
                        83
                    ]
                }
            ],
            "output": [
                {
                    "name": "target1",
                    "shape": [
                        15,
                        30
                    ],
                    "text": "B E R K E L E Y",
                    "token": "B <space> E <space> R <space> K <space> E <space> L <space> E <space> Y",
                    "tokenid": "4 2 7 2 20 2 13 2 7 2 14 2 7 2 27"
                }
            ],
            "utt2spk": "fjdn"
        },
        "fjdn-cen6-b": {
            "input": [
                {
                    "feat": "/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/dump/train_dev/deltafalse/feats.7.ark:113823",
                    "name": "input1",
                    "shape": [
                        148,
                        83
                    ]
                }
            ],
            "output": [
                {
                    "name": "target1",
                    "shape": [
                        21,
                        30
                    ],
                    "text": "OH SEVEN NINE TWO TWO",
                    "token": "O H <space> S E V E N <space> N I N E <space> T W O <space> T W O",
                    "tokenid": "17 10 2 21 7 24 7 16 2 16 11 16 7 2 22 25 17 2 22 25 17"
                }
            ],
            "utt2spk": "fjdn"
        },
        "fjdn-cen7-b": {
            "input": [
                {
                    "feat": "/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/dump/train_dev/deltafalse/feats.7.ark:126804",
                    "name": "input1",
                    "shape": [
                        258,
                        83
                    ]
                }
            ],
            "output": [
                {
                    "name": "target1",
                    "shape": [
                        43,
                        30
                    ],
                    "text": "TWO OH ONE FOUR SIX FOUR SIX OH EIGHT THREE",
                    "token": "T W O <space> O H <space> O N E <space> F O U R <space> S I X <space> F O U R <space> S I X <space> O H <space> E I G H T <space> T H R E E",
                    "tokenid": "22 25 17 2 17 10 2 17 16 7 2 8 17 23 20 2 21 11 26 2 8 17 23 20 2 21 11 26 2 17 10 2 7 11 9 10 22 2 22 10 20 7 7"
                }
            ],
            "utt2spk": "fjdn"
        },
        "fjdn-cen8-b": {
            "input": [
                {
                    "feat": "/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/dump/train_dev/deltafalse/feats.7.ark:148915",
                    "name": "input1",
                    "shape": [
                        168,
                        83
                    ]
                }
            ],
            "output": [
                {
                    "name": "target1",
                    "shape": [
                        32,
                        30
                    ],
                    "text": "MARCH TWENTIETH NINETEEN SEVENTY",
                    "token": "M A R C H <space> T W E N T I E T H <space> N I N E T E E N <space> S E V E N T Y",
                    "tokenid": "15 3 20 5 10 2 22 25 7 16 22 11 7 22 10 2 16 11 16 7 22 7 7 16 2 21 7 24 7 16 22 27"
                }
            ],
            "utt2spk": "fjdn"
        },
        "fjmd-an10-b": {
            "input": [
                {
                    "feat": "/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/dump/train_dev/deltafalse/feats.7.ark:163556",
                    "name": "input1",
                    "shape": [
                        158,
                        83
                    ]
                }
            ],
            "output": [
                {
                    "name": "target1",
                    "shape": [
                        17,
                        30
                    ],
                    "text": "G TWO EIGHTY NINE",
                    "token": "G <space> T W O <space> E I G H T Y <space> N I N E",
                    "tokenid": "9 2 22 25 17 2 7 11 9 10 22 27 2 16 11 16 7"
                }
            ],
            "utt2spk": "fjmd"
        },
        "fjmd-an6-b": {
            "input": [
                {
                    "feat": "/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/dump/train_dev/deltafalse/feats.8.ark:11",
                    "name": "input1",
                    "shape": [
                        78,
                        83
                    ]
                }
            ],
            "output": [
                {
                    "name": "target1",
                    "shape": [
                        2,
                        30
                    ],
                    "text": "NO",
                    "token": "N O",
                    "tokenid": "16 17"
                }
            ],
            "utt2spk": "fjmd"
        },
        "fjmd-an7-b": {
            "input": [
                {
                    "feat": "/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/dump/train_dev/deltafalse/feats.8.ark:7181",
                    "name": "input1",
                    "shape": [
                        218,
                        83
                    ]
                }
            ],
            "output": [
                {
                    "name": "target1",
                    "shape": [
                        26,
                        30
                    ],
                    "text": "O U FOUR SEVEN TWENTY NINE",
                    "token": "O <space> U <space> F O U R <space> S E V E N <space> T W E N T Y <space> N I N E",
                    "tokenid": "17 2 23 2 8 17 23 20 2 21 7 24 7 16 2 22 25 7 16 22 27 2 16 11 16 7"
                }
            ],
            "utt2spk": "fjmd"
        },
        "fjmd-an8-b": {
            "input": [
                {
                    "feat": "/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/dump/train_dev/deltafalse/feats.8.ark:25971",
                    "name": "input1",
                    "shape": [
                        178,
                        83
                    ]
                }
            ],
            "output": [
                {
                    "name": "target1",
                    "shape": [
                        19,
                        30
                    ],
                    "text": "B SEVEN TWENTY FIVE",
                    "token": "B <space> S E V E N <space> T W E N T Y <space> F I V E",
                    "tokenid": "4 2 21 7 24 7 16 2 22 25 7 16 22 27 2 8 11 24 7"
                }
            ],
            "utt2spk": "fjmd"
        },
        "fjmd-an9-b": {
            "input": [
                {
                    "feat": "/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/dump/train_dev/deltafalse/feats.8.ark:41441",
                    "name": "input1",
                    "shape": [
                        178,
                        83
                    ]
                }
            ],
            "output": [
                {
                    "name": "target1",
                    "shape": [
                        19,
                        30
                    ],
                    "text": "J FOUR THIRTY THREE",
                    "token": "J <space> F O U R <space> T H I R T Y <space> T H R E E",
                    "tokenid": "12 2 8 17 23 20 2 22 10 11 20 22 27 2 22 10 20 7 7"
                }
            ],
            "utt2spk": "fjmd"
        },
        "fjmd-cen1-b": {
            "input": [
                {
                    "feat": "/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/dump/train_dev/deltafalse/feats.8.ark:56912",
                    "name": "input1",
                    "shape": [
                        158,
                        83
                    ]
                }
            ],
            "output": [
                {
                    "name": "target1",
                    "shape": [
                        9,
                        30
                    ],
                    "text": "D R A V K",
                    "token": "D <space> R <space> A <space> V <space> K",
                    "tokenid": "6 2 20 2 3 2 24 2 13"
                }
            ],
            "utt2spk": "fjmd"
        },
        "fjmd-cen2-b": {
            "input": [
                {
                    "feat": "/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/dump/train_dev/deltafalse/feats.8.ark:70723",
                    "name": "input1",
                    "shape": [
                        278,
                        83
                    ]
                }
            ],
            "output": [
                {
                    "name": "target1",
                    "shape": [
                        15,
                        30
                    ],
                    "text": "J E A N E T T E",
                    "token": "J <space> E <space> A <space> N <space> E <space> T <space> T <space> E",
                    "tokenid": "12 2 7 2 3 2 16 2 7 2 22 2 22 2 7"
                }
            ],
            "utt2spk": "fjmd"
        },
        "fjmd-cen3-b": {
            "input": [
                {
                    "feat": "/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/dump/train_dev/deltafalse/feats.8.ark:94494",
                    "name": "input1",
                    "shape": [
                        98,
                        83
                    ]
                }
            ],
            "output": [
                {
                    "name": "target1",
                    "shape": [
                        7,
                        30
                    ],
                    "text": "ONE TEN",
                    "token": "O N E <space> T E N",
                    "tokenid": "17 16 7 2 22 7 16"
                }
            ],
            "utt2spk": "fjmd"
        },
        "fjmd-cen4-b": {
            "input": [
                {
                    "feat": "/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/dump/train_dev/deltafalse/feats.8.ark:103325",
                    "name": "input1",
                    "shape": [
                        348,
                        83
                    ]
                }
            ],
            "output": [
                {
                    "name": "target1",
                    "shape": [
                        21,
                        30
                    ],
                    "text": "S P R I N G H O U S E",
                    "token": "S <space> P <space> R <space> I <space> N <space> G <space> H <space> O <space> U <space> S <space> E",
                    "tokenid": "21 2 18 2 20 2 11 2 16 2 9 2 10 2 17 2 23 2 21 2 7"
                }
            ],
            "utt2spk": "fjmd"
        },
        "fjmd-cen5-b": {
            "input": [
                {
                    "feat": "/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/dump/train_dev/deltafalse/feats.8.ark:132906",
                    "name": "input1",
                    "shape": [
                        358,
                        83
                    ]
                }
            ],
            "output": [
                {
                    "name": "target1",
                    "shape": [
                        23,
                        30
                    ],
                    "text": "W R I G H T S V I L L E",
                    "token": "W <space> R <space> I <space> G <space> H <space> T <space> S <space> V <space> I <space> L <space> L <space> E",
                    "tokenid": "25 2 20 2 11 2 9 2 10 2 22 2 21 2 24 2 11 2 14 2 14 2 7"
                }
            ],
            "utt2spk": "fjmd"
        },
        "fjmd-cen6-b": {
            "input": [
                {
                    "feat": "/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/dump/train_dev/deltafalse/feats.8.ark:163317",
                    "name": "input1",
                    "shape": [
                        198,
                        83
                    ]
                }
            ],
            "output": [
                {
                    "name": "target1",
                    "shape": [
                        25,
                        30
                    ],
                    "text": "ONE SEVEN THREE SIX EIGHT",
                    "token": "O N E <space> S E V E N <space> T H R E E <space> S I X <space> E I G H T",
                    "tokenid": "17 16 7 2 21 7 24 7 16 2 22 10 20 7 7 2 21 11 26 2 7 11 9 10 22"
                }
            ],
            "utt2spk": "fjmd"
        },
        "fjmd-cen7-b": {
            "input": [
                {
                    "feat": "/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/dump/train_dev/deltafalse/feats.8.ark:180448",
                    "name": "input1",
                    "shape": [
                        218,
                        83
                    ]
                }
            ],
            "output": [
                {
                    "name": "target1",
                    "shape": [
                        28,
                        30
                    ],
                    "text": "TWO FIVE TWO TWO OH TWO NINE",
                    "token": "T W O <space> F I V E <space> T W O <space> T W O <space> O H <space> T W O <space> N I N E",
                    "tokenid": "22 25 17 2 8 11 24 7 2 22 25 17 2 22 25 17 2 17 10 2 22 25 17 2 16 11 16 7"
                }
            ],
            "utt2spk": "fjmd"
        },
        "fjmd-cen8-b": {
            "input": [
                {
                    "feat": "/home/skarita/Downloads/interspeech2019-tutorial/notebooks/interspeech2019_asr/espnet/egs/an4/asr1/dump/train_dev/deltafalse/feats.8.ark:199239",
                    "name": "input1",
                    "shape": [
                        218,
                        83
                    ]
                }
            ],
            "output": [
                {
                    "name": "target1",
                    "shape": [
                        23,
                        30
                    ],
                    "text": "ELEVEN NINE SIXTY EIGHT",
                    "token": "E L E V E N <space> N I N E <space> S I X T Y <space> E I G H T",
                    "tokenid": "7 14 7 24 7 16 2 16 11 16 7 2 21 11 26 22 27 2 7 11 9 10 22"
                }
            ],
            "utt2spk": "fjmd"
        }
    }
}

3.3 NN Training (Stage 3 - 4)

You can configure NN training with conf/train_xxx.yaml

image.png

In [21]:
!tree espnet/egs/voxforge/asr1/conf
espnet/egs/voxforge/asr1/conf
├── decode.yaml -> tuning/decode_pytorch_transformer.yaml
├── fbank.conf
├── gpu.conf
├── pitch.conf
├── queue.conf
├── slurm.conf
├── train.yaml -> tuning/train_pytorch_transformer.yaml
└── tuning
    ├── decode_pytorch_transformer.yaml
    ├── decode_rnn.yaml
    ├── decode_transducer.yaml
    ├── lm.yaml
    ├── train_pytorch_transformer_d6-2048.yaml
    ├── train_pytorch_transformer.yaml
    ├── train_rnn.yaml
    └── train_transducer.yaml

1 directory, 15 files

ESPnet training features

A lot of options at espnet/espnet/bin/asr_train.py

  • NN backends: pytorch, chainer
  • Predefined models
    • RNN (LSTM, GRU, VGG, etc) with Attention + CTC (dot, location, multi-head, etc)
    • Transformer + CTC
    • RNN Transducer with Attention
  • Speech enhancement (joint training):
    • beamformer
    • dereverb (WPE, DNN-WPE)
    • speech separation
  • Data augmentation: SpecAugment, speed perturbation, etc
  • Minibatch strategy
    • sorting, category, counting (tensor elements, sequence frames, sequences)
  • Multi GPU training
  • Half/mixed precision training
  • Regularizations: dropout, label smoothing, weight noise, weight decay, etc

Training config: RNN with Attention + CTC

In [5]:
!cat espnet/egs/an4/asr1/conf/train_mtlalpha0.5.yaml
# minibatch related
batch-size: 30
maxlen-in: 800  # if input length  > maxlen_in, batchsize is automatically reduced
maxlen-out: 150 # if output length > maxlen_out, batchsize is automatically reduced
# optimization related
sortagrad: 0 # Feed samples from shortest to longest ; -1: enabled for all epochs, 0: disabled, other: enabled for 'other' epochs
opt: adadelta
epochs: 20
patience: 3

# scheduled sampling option
sampling-probability: 0.0

# encoder related
etype: blstmp     # encoder architecture type
elayers: 4
eunits: 320
eprojs: 320
subsample: "1_2_2_1_1" # skip every n frame from input to nth layers
# decoder related
dlayers: 1
dunits: 300
# attention related
atype: location
adim: 320
aconv-chans: 10
aconv-filts: 100

# hybrid CTC/attention
mtlalpha: 0.5

Training config: Transformer + CTC

In [38]:
!cat espnet/egs/librispeech/asr1/conf/tuning/train_pytorch_transformer_large_ngpu4.yaml
# This configuration requires 4 gpus with 12GB memory
accum-grad: 4
adim: 512
aheads: 8
backend: pytorch
batch-bins: 15000000
dlayers: 6
dropout-rate: 0.1
dunits: 2048
elayers: 12
epochs: 120
eunits: 2048
grad-clip: 5
lsm-weight: 0.1
model-module: espnet.nets.pytorch_backend.e2e_asr_transformer:E2E
mtlalpha: 0.3
opt: noam
patience: 0
sortagrad: 0
transformer-attn-dropout-rate: 0.0
transformer-init: pytorch
transformer-input-layer: conv2d
transformer-length-normalized-loss: false
transformer-lr: 10.0
transformer-warmup-steps: 25000

Training config: RNN Transducer (with Attention)

In [8]:
!cat espnet/egs/voxforge/asr1/conf/tuning/train_transducer.yaml
# minibatch related
batch-size: 20
maxlen-in: 800
maxlen-out: 150

# optimization related
criterion: loss
early-stop-criterion: "validation/main/loss"
sortagrad: 0
opt: adadelta
epochs: 20
patience: 3

# network architecture
## encoder related
etype: vggblstm
elayers: 4
eunits: 256
eprojs: 256
subsample: "1_1_1_1_1"
dropout-rate: 0.5
## decoder related
dtype: lstm
dlayers: 1
dec-embed-dim: 256
dunits: 256
dropout-rate-decoder: 0.2
dropout-rate-embed-decoder: 0.2
## attention related
atype: location
adim: 256
aheads: 1
awin: 5
aconv-chans: 10
aconv-filts: 100
## joint network related
joint-dim: 256

# transducer related
## mtlalpha should be set to 1.0 (CTC) to use transducer
mtlalpha: 1.0
## switch to 'rnnt-att' to use transducer with attention
rnnt-mode: 'rnnt'
model-module: "espnet.nets.pytorch_backend.e2e_asr_transducer:E2E"

Run LM and ASR NN training

RNNLM and RNN with attention and CTC in AN4

In [0]:
# WARNING: This code takes several minutes!
!cd espnet/egs/an4/asr1; ./run.sh  --ngpu 1 --stage 3 --stop-stage 4 --train-config ./conf/train_mtlalpha0.5.yaml
dictionary: data/lang_1char/train_nodev_units.txt
stage 3: LM Preparation
WARNING:root:OOV rate = 0.00 %
stage 4: Network Training

TIPS 1/3: change_yaml.py

Tweak training config. for example: --train-config $(change_yaml.py train.yaml -a lr=10.0)

  • creates the config with new name: train_lr10.0.yaml
  • exp results are stored in dir: exp/train_lr10.0/results
  • useful hyperparameter search by array jobs
In [0]:
# WARNING: This code takes several minutes!
!cd espnet/egs/an4/asr1; source path.sh; \
  ./run.sh  --ngpu 1 --stage 4 --stop-stage 4 \
  --train-config $(change_yaml.py ./conf/train_mtlalpha0.5.yaml -a eunits=100 -a epochs=5)
dictionary: data/lang_1char/train_nodev_units.txt
stage 4: Network Training

TIPS 2/3: tensorboard

To find the best config, view tensorboard

In [7]:
!pip install -q tf-nightly-2.0-preview
# Load the TensorBoard notebook extension
%load_ext tensorboard 
%tensorboard --logdir espnet/egs/an4/asr1/tensorboard
---------------------------------------------------------------------------
ModuleNotFoundError                       Traceback (most recent call last)
<ipython-input-7-cd6d82ef7510> in <module>
      1 get_ipython().system('pip install -q tf-nightly-2.0-preview')
      2 # Load the TensorBoard notebook extension
----> 3 get_ipython().run_line_magic('load_ext', 'tensorboard ')
      4 get_ipython().run_line_magic('tensorboard', '--logdir espnet/egs/an4/asr1/tensorboard')

~/Documents/repos/espnet/tools/venv/lib/python3.7/site-packages/IPython/core/interactiveshell.py in run_line_magic(self, magic_name, line, _stack_depth)
   2312                 kwargs['local_ns'] = sys._getframe(stack_depth).f_locals
   2313             with self.builtin_trap:
-> 2314                 result = fn(*args, **kwargs)
   2315             return result
   2316 

</home/skarita/Documents/repos/espnet/tools/venv/lib/python3.7/site-packages/decorator.py:decorator-gen-64> in load_ext(self, module_str)

~/Documents/repos/espnet/tools/venv/lib/python3.7/site-packages/IPython/core/magic.py in <lambda>(f, *a, **k)
    185     # but it's overkill for just that one bit of state.
    186     def magic_deco(arg):
--> 187         call = lambda f, *a, **k: f(*a, **k)
    188 
    189         if callable(arg):

~/Documents/repos/espnet/tools/venv/lib/python3.7/site-packages/IPython/core/magics/extension.py in load_ext(self, module_str)
     31         if not module_str:
     32             raise UsageError('Missing module name.')
---> 33         res = self.shell.extension_manager.load_extension(module_str)
     34 
     35         if res == 'already loaded':

~/Documents/repos/espnet/tools/venv/lib/python3.7/site-packages/IPython/core/extensions.py in load_extension(self, module_str)
     78             if module_str not in sys.modules:
     79                 with prepended_to_syspath(self.ipython_extension_dir):
---> 80                     mod = import_module(module_str)
     81                     if mod.__file__.startswith(self.ipython_extension_dir):
     82                         print(("Loading extensions from {dir} is deprecated. "

~/Documents/repos/espnet/tools/venv/lib/python3.7/importlib/__init__.py in import_module(name, package)
    125                 break
    126             level += 1
--> 127     return _bootstrap._gcd_import(name[level:], package, level)
    128 
    129 

~/Documents/repos/espnet/tools/venv/lib/python3.7/importlib/_bootstrap.py in _gcd_import(name, package, level)

~/Documents/repos/espnet/tools/venv/lib/python3.7/importlib/_bootstrap.py in _find_and_load(name, import_)

~/Documents/repos/espnet/tools/venv/lib/python3.7/importlib/_bootstrap.py in _find_and_load_unlocked(name, import_)

ModuleNotFoundError: No module named 'tensorboard '

image.png

TIPS 3/3 log files

To monitor training, exp/train*/results/ contains useful files:

  • loss.png train/valid loss values
  • acc.png train/valid accuracy
  • cer.png train/valid character error rate
  • att_ws/*.png attention plots
In [6]:
import glob
from IPython.display import Image, display_png
expdir = "espnet/egs/an4/asr1/exp/train_nodev_pytorch_train_mtlalpha0.5/results/"
for name in ["loss.png", "acc.png", "cer.png"]:
    print(name)
    display_png(Image(expdir + name, width=500))
loss.png
acc.png
cer.png

Attention plots (RNN)

A diagonal attention is good measure to check training in ASR

</table> RNN attentions at the first (left) and last (right) epoch

Attention plots (Transformer)

Transformer has many attentions. We recommend to monitor the last decoder layer's attention to encoded features.

trnf-attn

Transformer multi (=4) head attention plot in the last decoder layer at the last epoch

Attention plots (Transformer)

Transformer has many attentions. We recommend to monitor the last decoder layer's attention to encoded features.

trnf-attn

Transformer multi (=4) head attention plot in the first decoder layer at the last epoch

3.4 Decoding and evaluation (Stage 5)

The last stage of ASR recipe

image.png

decoding YAML config

very similar to training YAML

  • decoding score:

$\mathrm{argmax}_y (1 - \lambda) \log P_{dec}(y|x) + \lambda \log P_{ctc}(y|x) + \gamma \log P_{lm}(y) + b |y|$

In [0]:
!cat espnet/egs/an4/asr1/conf/decode_ctcweight0.5.yaml
# decoding parameter
beam-size: 20
penalty: 0.0
maxlenratio: 0.0
minlenratio: 0.0
ctc-weight: 0.5
lm-weight: 1.0

Run ASR decoding

In [0]:
# WARNING: This code takes several minutes!
!cd espnet/egs/an4/asr1; source path.sh; \
  ./run.sh --stage 5 --decode-config $(change_yaml.py conf/decode_ctcweight0.5.yaml -a batchsize=0) --train-config conf/train_mtlalpha0.5.yaml
dictionary: data/lang_1char/train_nodev_units.txt
stage 5: Decoding
2019-09-11 11:20:45,892 (splitjson:40) INFO: /usr/bin/python3 /content/espnet/egs/an4/asr1/../../../utils/splitjson.py --parts 8 dump/train_dev/deltafalse/data.json
2019-09-11 11:20:45,894 (splitjson:52) INFO: number of utterances = 100
2019-09-11 11:20:45,895 (splitjson:40) INFO: /usr/bin/python3 /content/espnet/egs/an4/asr1/../../../utils/splitjson.py --parts 8 dump/test/deltafalse/data.json
2019-09-11 11:20:45,897 (splitjson:52) INFO: number of utterances = 130
2019-09-11 11:30:47,555 (concatjson:36) INFO: /usr/bin/python3 /content/espnet/egs/an4/asr1/../../../utils/concatjson.py exp/train_nodev_pytorch_train_mtlalpha0.5/decode_train_dev_decode_ctcweight0.5_batchsize0_lm_word100/data.1.json exp/train_nodev_pytorch_train_mtlalpha0.5/decode_train_dev_decode_ctcweight0.5_batchsize0_lm_word100/data.2.json exp/train_nodev_pytorch_train_mtlalpha0.5/decode_train_dev_decode_ctcweight0.5_batchsize0_lm_word100/data.3.json exp/train_nodev_pytorch_train_mtlalpha0.5/decode_train_dev_decode_ctcweight0.5_batchsize0_lm_word100/data.4.json exp/train_nodev_pytorch_train_mtlalpha0.5/decode_train_dev_decode_ctcweight0.5_batchsize0_lm_word100/data.5.json exp/train_nodev_pytorch_train_mtlalpha0.5/decode_train_dev_decode_ctcweight0.5_batchsize0_lm_word100/data.6.json exp/train_nodev_pytorch_train_mtlalpha0.5/decode_train_dev_decode_ctcweight0.5_batchsize0_lm_word100/data.7.json exp/train_nodev_pytorch_train_mtlalpha0.5/decode_train_dev_decode_ctcweight0.5_batchsize0_lm_word100/data.8.json
2019-09-11 11:30:47,557 (concatjson:46) INFO: new json has 100 utterances
2019-09-11 11:30:48,707 (json2trn:43) INFO: /usr/bin/python3 /content/espnet/egs/an4/asr1/../../../utils/json2trn.py exp/train_nodev_pytorch_train_mtlalpha0.5/decode_train_dev_decode_ctcweight0.5_batchsize0_lm_word100/data.json data/lang_1char/train_nodev_units.txt --num-spkrs 1 --refs exp/train_nodev_pytorch_train_mtlalpha0.5/decode_train_dev_decode_ctcweight0.5_batchsize0_lm_word100/ref.trn --hyps exp/train_nodev_pytorch_train_mtlalpha0.5/decode_train_dev_decode_ctcweight0.5_batchsize0_lm_word100/hyp.trn
2019-09-11 11:30:48,707 (json2trn:45) INFO: reading exp/train_nodev_pytorch_train_mtlalpha0.5/decode_train_dev_decode_ctcweight0.5_batchsize0_lm_word100/data.json
2019-09-11 11:30:48,711 (json2trn:49) INFO: reading data/lang_1char/train_nodev_units.txt
write a CER (or TER) result in exp/train_nodev_pytorch_train_mtlalpha0.5/decode_train_dev_decode_ctcweight0.5_batchsize0_lm_word100/result.txt
|   SPKR      |   # Snt      # Wrd    |   Corr         Sub        Del         Ins         Err       S.Err    |
|   Sum/Avg   |    100        1915    |   78.0         7.3       14.8         0.8        22.9        71.0    |
2019-09-11 11:32:20,598 (concatjson:36) INFO: /usr/bin/python3 /content/espnet/egs/an4/asr1/../../../utils/concatjson.py exp/train_nodev_pytorch_train_mtlalpha0.5/decode_test_decode_ctcweight0.5_batchsize0_lm_word100/data.1.json exp/train_nodev_pytorch_train_mtlalpha0.5/decode_test_decode_ctcweight0.5_batchsize0_lm_word100/data.2.json exp/train_nodev_pytorch_train_mtlalpha0.5/decode_test_decode_ctcweight0.5_batchsize0_lm_word100/data.3.json exp/train_nodev_pytorch_train_mtlalpha0.5/decode_test_decode_ctcweight0.5_batchsize0_lm_word100/data.4.json exp/train_nodev_pytorch_train_mtlalpha0.5/decode_test_decode_ctcweight0.5_batchsize0_lm_word100/data.5.json exp/train_nodev_pytorch_train_mtlalpha0.5/decode_test_decode_ctcweight0.5_batchsize0_lm_word100/data.6.json exp/train_nodev_pytorch_train_mtlalpha0.5/decode_test_decode_ctcweight0.5_batchsize0_lm_word100/data.7.json exp/train_nodev_pytorch_train_mtlalpha0.5/decode_test_decode_ctcweight0.5_batchsize0_lm_word100/data.8.json
2019-09-11 11:32:20,599 (concatjson:46) INFO: new json has 130 utterances
2019-09-11 11:32:20,997 (json2trn:43) INFO: /usr/bin/python3 /content/espnet/egs/an4/asr1/../../../utils/json2trn.py exp/train_nodev_pytorch_train_mtlalpha0.5/decode_test_decode_ctcweight0.5_batchsize0_lm_word100/data.json data/lang_1char/train_nodev_units.txt --num-spkrs 1 --refs exp/train_nodev_pytorch_train_mtlalpha0.5/decode_test_decode_ctcweight0.5_batchsize0_lm_word100/ref.trn --hyps exp/train_nodev_pytorch_train_mtlalpha0.5/decode_test_decode_ctcweight0.5_batchsize0_lm_word100/hyp.trn
2019-09-11 11:32:20,998 (json2trn:45) INFO: reading exp/train_nodev_pytorch_train_mtlalpha0.5/decode_test_decode_ctcweight0.5_batchsize0_lm_word100/data.json
2019-09-11 11:32:20,999 (json2trn:49) INFO: reading data/lang_1char/train_nodev_units.txt
write a CER (or TER) result in exp/train_nodev_pytorch_train_mtlalpha0.5/decode_test_decode_ctcweight0.5_batchsize0_lm_word100/result.txt
|   SPKR     |   # Snt      # Wrd   |   Corr        Sub         Del        Ins        Err       S.Err   |
|   Sum/Avg  |    130        2565   |   86.8        6.0         7.2        0.5       13.8        60.0   |

3.5 Check evaluation results

ESPnet uses sclite in SPTK to evaluate ASR errors

  • token error rate: exp/(train dir)/(decode dir)/result.txt
  • word error rate: exp/(train dir)/(decode dir)/result.wrd.txt
In [0]:
!ls espnet/egs/an4/asr1/exp/train_nodev_pytorch_train_mtlalpha0.5/*/result.txt | xargs -n1 grep -e Avg -e SPK -m 2
|   SPKR     |   # Snt      # Wrd   |   Corr        Sub         Del        Ins        Err       S.Err   |
|   Sum/Avg  |    130        2565   |   86.8        6.0         7.2        0.5       13.8        60.0   |
|   SPKR      |   # Snt      # Wrd    |   Corr         Sub        Del         Ins         Err       S.Err    |
|   Sum/Avg   |    100        1915    |   78.0         7.3       14.8         0.8        22.9        71.0    |
In [0]:
!head -n 36 espnet/egs/an4/asr1/exp/train_nodev_pytorch_train_mtlalpha0.5/decode_test_decode_ctcweight0.5_batchsize0_lm_word100/result.txt

                     SYSTEM SUMMARY PERCENTAGES by SPEAKER                      

,-------------------------------------------------------------------------------------------------------.
|exp/train_nodev_pytorch_train_mtlalpha0.5/decode_test_decode_ctcweight0.5_batchsize0_lm_word100/hyp.trn|
|-------------------------------------------------------------------------------------------------------|
|   SPKR     |   # Snt      # Wrd   |   Corr        Sub         Del        Ins        Err       S.Err   |
|------------+----------------------+-------------------------------------------------------------------|
|   fcaw     |     13         237   |   88.6        9.3         2.1        0.0       11.4        61.5   |
|------------+----------------------+-------------------------------------------------------------------|
|   fjlp     |     13         242   |   91.3        4.5         4.1        0.0        8.7        53.8   |
|------------+----------------------+-------------------------------------------------------------------|
|   fvap     |     13         274   |   83.2        6.6        10.2        0.0       16.8        84.6   |
|------------+----------------------+-------------------------------------------------------------------|
|   marh     |     13         276   |   84.1        8.0         8.0        1.4       17.4        61.5   |
|------------+----------------------+-------------------------------------------------------------------|
|   mdms2    |     13         274   |   92.0        5.1         2.9        1.1        9.1        53.8   |
|------------+----------------------+-------------------------------------------------------------------|
|   menk     |     13         258   |   89.9        5.8         4.3        0.4       10.5        53.8   |
|------------+----------------------+-------------------------------------------------------------------|
|   miry     |     13         274   |   93.1        4.7         2.2        1.8        8.8        53.8   |
|------------+----------------------+-------------------------------------------------------------------|
|   mjgm     |     13         226   |   80.5        4.4        15.0        0.0       19.5        46.2   |
|------------+----------------------+-------------------------------------------------------------------|
|   mjwl     |     13         241   |   76.8        5.8        17.4        0.4       23.7        69.2   |
|------------+----------------------+-------------------------------------------------------------------|
|   mmxg     |     13         263   |   87.1        6.1         6.8        0.0       12.9        61.5   |
|=======================================================================================================|
|   Sum/Avg  |    130        2565   |   86.8        6.0         7.2        0.5       13.8        60.0   |
|=======================================================================================================|
|    Mean    |   13.0       256.5   |   86.7        6.0         7.3        0.5       13.9        60.0   |
|    S.D.    |    0.0        18.6   |    5.4        1.6         5.4        0.7        5.2        10.8   |
|   Median   |   13.0       260.5   |   87.8        5.8         5.6        0.2       12.2        57.7   |
`-------------------------------------------------------------------------------------------------------'
In [0]:
!tail -n 7 espnet/egs/an4/asr1/exp/train_nodev_pytorch_train_mtlalpha0.5/decode_test_decode_ctcweight0.5_batchsize0_lm_word100/result.txt
id: (mmxg-mmxg-cen8-b)
Scores: (#C #S #D #I) 26 2 8 0
REF:  O C t O B e r <space> t w e n t y <space> f o u r <space> n i n E T E e N <space> s e v e n T Y 
HYP:  E N t * * e r <space> t w e n t y <space> f o u r <space> n i n * * * e * <space> s e v e n * * 
Eval: S S   D D                                                       D D D   D                   D D 


ASR result as data.json

Find detail results in exp/xxx/decode_yyy/data.json

In [13]:
!head -n30 espnet/egs/an4/asr1/exp/train_nodev_pytorch_train_mtlalpha0.5/decode_test_decode_ctcweight0.5_lm_word100/data.json
{
    "utts": {
        "fcaw-an406-b": {
            "output": [
                {
                    "name": "target1[1]",
                    "rec_text": "<blank><blank>RUBOU<blank>T<blank><blank><blank> G N E F THREE NINE<eos>",
                    "rec_token": "<blank> <blank> R U B O U <blank> T <blank> <blank> <blank> <space> G <space> N <space> E <space> F <space> T H R E E <space> N I N E <eos>",
                    "rec_tokenid": "0 0 20 23 4 17 23 0 22 0 0 0 2 9 2 16 2 7 2 8 2 22 10 20 7 7 2 16 11 16 7 29",
                    "score": -38.3045539855957,
                    "shape": [
                        25,
                        30
                    ],
                    "text": "RUBOUT G M E F THREE NINE",
                    "token": "R U B O U T <space> G <space> M <space> E <space> F <space> T H R E E <space> N I N E",
                    "tokenid": "20 23 4 17 23 22 2 9 2 15 2 7 2 8 2 22 10 20 7 7 2 16 11 16 7"
                }
            ],
            "utt2spk": "fcaw"
        },
        "fcaw-an407-b": {
            "output": [
                {
                    "name": "target1[1]",
                    "rec_text": "ERASE C Q Q F SEVEN<eos>",
                    "rec_token": "E R A S E <space> C <space> Q <space> Q <space> F <space> S E V E N <eos>",
                    "rec_tokenid": "7 20 3 21 7 2 5 2 19 2 19 2 8 2 21 7 24 7 16 29",
                    "score": -34.00189971923828,
                    "shape": [

4. Use ESPnet in Python

  1. Load the speech features
  2. Load the pretrained model
  3. Recognize the speech by the model
  4. Visualizations (attention, ctc)

4.1 Load speech features

In [23]:
%matplotlib inline
import json
import matplotlib.pyplot as plt
import kaldiio

# ESPnet summarizes dataset to JSON
root = "espnet/egs/an4/asr1"
with open(root + "/dump/test/deltafalse/data.json", "r") as f:
    test_json = json.load(f)["utts"]

key, info = list(test_json.items())[10]
fbank = kaldiio.load_mat(info["input"][0]["feat"])

# plot the speech feature
plt.matshow(fbank.T[::-1])
plt.title(key + ": " + info["output"][0]["text"])
Out[23]:
Text(0.5, 1.05, 'fcaw-cen6-b: ONE FIVE TWO THREE SIX')

4.2 Load pretrained model

pretrained model configuration (JSON) and snapshots (pickle) are available in exp/train_xxx/results

In [27]:
!ls espnet/egs/an4/asr1/exp/train_nodev_pytorch_train_mtlalpha0.5/results
acc.png		model.json	 snapshot.ep.13  snapshot.ep.19  snapshot.ep.6
att_ws		model.loss.best  snapshot.ep.14  snapshot.ep.2	 snapshot.ep.7
cer.png		snapshot.ep.1	 snapshot.ep.15  snapshot.ep.20  snapshot.ep.8
log		snapshot.ep.10	 snapshot.ep.16  snapshot.ep.3	 snapshot.ep.9
loss.png	snapshot.ep.11	 snapshot.ep.17  snapshot.ep.4
model.acc.best	snapshot.ep.12	 snapshot.ep.18  snapshot.ep.5

4.2 Load pretrained model

let's load it from python

In [28]:
import json
import torch
from espnet.nets.pytorch_backend.e2e_asr import E2E

model_dir = "espnet/egs/an4/asr1/exp/train_nodev_pytorch_train_mtlalpha0.5/results"

# load model
with open(model_dir + "/model.json", "r") as f:
    idim, odim, conf = json.load(f)
model = E2E.build(idim, odim, **conf)
model.load_state_dict(torch.load(model_dir + "/model.acc.best"))
model.cpu().eval()
vocab = conf["char_list"]
print(vocab)
model
['<blank>', '<unk>', '<space>', 'A', 'B', 'C', 'D', 'E', 'F', 'G', 'H', 'I', 'J', 'K', 'L', 'M', 'N', 'O', 'P', 'Q', 'R', 'S', 'T', 'U', 'V', 'W', 'X', 'Y', 'Z', '<eos>']
Out[28]:
E2E(
  (enc): Encoder(
    (enc): ModuleList(
      (0): RNNP(
        (birnn0): LSTM(83, 320, batch_first=True, bidirectional=True)
        (bt0): Linear(in_features=640, out_features=320, bias=True)
        (birnn1): LSTM(320, 320, batch_first=True, bidirectional=True)
        (bt1): Linear(in_features=640, out_features=320, bias=True)
        (birnn2): LSTM(320, 320, batch_first=True, bidirectional=True)
        (bt2): Linear(in_features=640, out_features=320, bias=True)
        (birnn3): LSTM(320, 320, batch_first=True, bidirectional=True)
        (bt3): Linear(in_features=640, out_features=320, bias=True)
      )
    )
  )
  (ctc): CTC(
    (ctc_lo): Linear(in_features=320, out_features=30, bias=True)
    (ctc_loss): CTCLoss()
  )
  (att): ModuleList(
    (0): AttLoc(
      (mlp_enc): Linear(in_features=320, out_features=320, bias=True)
      (mlp_dec): Linear(in_features=300, out_features=320, bias=False)
      (mlp_att): Linear(in_features=10, out_features=320, bias=False)
      (loc_conv): Conv2d(1, 10, kernel_size=(1, 201), stride=(1, 1), padding=(0, 100), bias=False)
      (gvec): Linear(in_features=320, out_features=1, bias=True)
    )
  )
  (dec): Decoder(
    (embed): Embedding(30, 300)
    (dropout_emb): Dropout(p=0.0)
    (decoder): ModuleList(
      (0): LSTMCell(620, 300)
    )
    (dropout_dec): ModuleList(
      (0): Dropout(p=0.0)
    )
    (output): Linear(in_features=300, out_features=30, bias=True)
    (att): ModuleList(
      (0): AttLoc(
        (mlp_enc): Linear(in_features=320, out_features=320, bias=True)
        (mlp_dec): Linear(in_features=300, out_features=320, bias=False)
        (mlp_att): Linear(in_features=10, out_features=320, bias=False)
        (loc_conv): Conv2d(1, 10, kernel_size=(1, 201), stride=(1, 1), padding=(0, 100), bias=False)
        (gvec): Linear(in_features=320, out_features=1, bias=True)
      )
    )
  )
)

4.3 Recognize the speech by the model

You can perform joint decoding with all the models (S2S, CTC, LM, etc) in ESPnet

In [35]:
import re
from espnet.nets.beam_search import BeamSearch

key, info = list(test_json.items())[10]
fbank = kaldiio.load_mat(info["input"][0]["feat"])

# setup beam search
bs = BeamSearch(
    scorers=model.scorers(), weights={"decoder": 0.5, "ctc": 0.5},
    sos=model.sos, eos=model.eos,
    beam_size=2, vocab_size=len(vocab))
# GPU decoding: model.cuda(), bs.cuda()
with torch.no_grad():
    encoded = model.encode(torch.as_tensor(fbank))
    result = bs(encoded)  # get N-best results

print("groundtruth:", info["output"][0]["text"])
print("N-best list:")
for n, hyp in enumerate(result, 1):
    text = "".join(vocab[y] for y in hyp.yseq).replace("<space>", " ").replace("<eos>", "")
    scores = {k: f"{float(v):0.3f}" for k, v in hyp.scores.items()}
    print(f"{n}: {text}, score: {scores}")
groundtruth: ONE FIVE TWO THREE SIX
N-best list:
1: ONE FIVE TWO ONE THREE SIX, score: {'decoder': '-11.122', 'ctc': '-0.071'}
2: ONE FIVE TWO ONE THRE SIX, score: {'decoder': '-15.021', 'ctc': '-0.071'}
3: ONE FIVE TWO ONE THREE SIXT, score: {'decoder': '-17.329', 'ctc': '-0.152'}
4: ONE FIVE TWO ONE THREE SIXTY, score: {'decoder': '-15.520', 'ctc': '-0.009'}
5: ONE FIVE TWO ONE THREE SIXTY E, score: {'decoder': '-21.535', 'ctc': '-0.001'}
6: ONE FIVE TWO ONE THREE SIXTY EIN, score: {'decoder': '-22.140', 'ctc': '-0.000'}
7: ONE FIVE TWO ONE THREE SIXTY EIG, score: {'decoder': '-21.472', 'ctc': '-0.001'}

4.4 Visualizations

Let's see NN internals with python.

  • Attention matrix between encoder and decoder
  • CTC posterior probability
In [36]:
# Attention plot
x = torch.as_tensor(fbank).unsqueeze(0)
y = result[0].yseq.unsqueeze(0)
attn = model.calculate_all_attentions(x, [len(fbank)], y)[0]

# plot
fig, ax = plt.subplots(2, figsize=(10, 7))
ax[0].set_title("Attention")
ax[0].matshow(attn, aspect="auto")
txt = [vocab[int(i)] for i in y[0]]
ax[0].set_yticks(range(len(txt)))
ax[0].set_yticklabels(txt)
ax[1].matshow(fbank.T[::-1], aspect="auto")
fig.tight_layout()
In [0]:
# CTC posterior plot
with torch.no_grad():
    logp = model.ctc.log_softmax(encoded.unsqueeze(0))[0]
    prob = logp.exp_().numpy()

fig, ax = plt.subplots(2, figsize=(10, 7))
ax[0].set_title("CTC posterior")
vs = set(int(y) for y in result[0].yseq)
vs.add(0)
for n, i in enumerate(vs):
    v = vocab[i]
    ax[0].plot(prob[:, i], label=v, linestyle="-" if len(v) == 1 else "--")
ax[0].legend(loc='center left', bbox_to_anchor=(1, 0.5))
ax[0].set_xlim(0, len(encoded)-1)
ax[1].matshow(fbank.T[::-1], aspect="auto")
fig.tight_layout()

5 Extend ESPnet for your research

  1. Python library structure
  2. Define your new model (Transformer)
  3. Use your new model via run.sh --model-module ...

5.1 Python library structure under espnet

  • nets: neural networks
  • bin: command line tools
  • transform: pre/post processing (data augumentation, frontend, etc)
  • asr, lm, mt, tts: task specific procedures (train, decode, etc)
  • utils: others

Let's extend nets with your new ASR model

In [37]:
!tree -L 1 espnet/espnet | grep -v pycache
espnet/espnet
├── asr
├── bin
├── __init__.py
├── lm
├── mt
├── nets
├── transform
├── tts
└── utils

9 directories, 1 file

5.2 Define your new ASR model

Each task defines its own interfaces for new models

In [0]:
ls espnet/espnet/nets/*interface.py
espnet/espnet/nets/asr_interface.py  espnet/espnet/nets/scorer_interface.py
espnet/espnet/nets/lm_interface.py   espnet/espnet/nets/tts_interface.py
espnet/espnet/nets/mt_interface.py
In [5]:
!cat espnet/espnet/nets/asr_interface.py
"""ASR Interface module."""
import argparse

from espnet.bin.asr_train import get_parser
from espnet.utils.dynamic_import import dynamic_import
from espnet.utils.fill_missing_args import fill_missing_args


class ASRInterface:
    """ASR Interface for ESPnet model implementation."""

    @staticmethod
    def add_arguments(parser):
        """Add arguments to parser."""
        return parser

    @classmethod
    def build(cls, idim: int, odim: int, **kwargs):
        """Initialize this class with python-level args.

        Args:
            idim (int): The number of an input feature dim.
            odim (int): The number of output vocab.

        Returns:
            ASRinterface: A new instance of ASRInterface.

        """
        def wrap(parser):
            return get_parser(parser, required=False)

        args = argparse.Namespace(**kwargs)
        args = fill_missing_args(args, wrap)
        args = fill_missing_args(args, cls.add_arguments)
        return cls(idim, odim, args)

    def forward(self, xs, ilens, ys):
        """Compute loss for training.

        :param xs:
            For pytorch, batch of padded source sequences torch.Tensor (B, Tmax, idim)
            For chainer, list of source sequences chainer.Variable
        :param ilens: batch of lengths of source sequences (B)
            For pytorch, torch.Tensor
            For chainer, list of int
        :param ys:
            For pytorch, batch of padded source sequences torch.Tensor (B, Lmax)
            For chainer, list of source sequences chainer.Variable
        :return: loss value
        :rtype: torch.Tensor for pytorch, chainer.Variable for chainer
        """
        raise NotImplementedError("forward method is not implemented")

    def recognize(self, x, recog_args, char_list=None, rnnlm=None):
        """Recognize x for evaluation.

        :param ndarray x: input acouctic feature (B, T, D) or (T, D)
        :param namespace recog_args: argment namespace contraining options
        :param list char_list: list of characters
        :param torch.nn.Module rnnlm: language model module
        :return: N-best decoding results
        :rtype: list
        """
        raise NotImplementedError("recognize method is not implemented")

    def recognize_batch(self, x, recog_args, char_list=None, rnnlm=None):
        """Beam search implementation for batch.

        :param torch.Tensor x: encoder hidden state sequences (B, Tmax, Henc)
        :param namespace recog_args: argument namespace containing options
        :param list char_list: list of characters
        :param torch.nn.Module rnnlm: language model module
        :return: N-best decoding results
        :rtype: list
        """
        raise NotImplementedError("Batch decoding is not supported yet.")

    def calculate_all_attentions(self, xs, ilens, ys):
        """Caluculate attention.

        :param list xs_pad: list of padded input sequences [(T1, idim), (T2, idim), ...]
        :param ndarray ilens: batch of lengths of input sequences (B)
        :param list ys: list of character id sequence tensor [(L1), (L2), (L3), ...]
        :return: attention weights (B, Lmax, Tmax)
        :rtype: float ndarray
        """
        raise NotImplementedError("calculate_all_attentions method is not implemented")

    @property
    def attention_plot_class(self):
        """Get attention plot class."""
        from espnet.asr.asr_utils import PlotAttentionReport
        return PlotAttentionReport

    def encode(self, feat):
        """Encode feature in `beam_search` (optional).

        Args:
            x (numpy.ndarray): input feature (T, D)
        Returns:
            torch.Tensor for pytorch, chainer.Variable for chainer:
                encoded feature (T, D)

        """
        raise NotImplementedError("encode method is not implemented")

    def scorers(self):
        """Get scorers for `beam_search` (optional).

        Returns:
            dict[str, ScorerInterface]: dict of `ScorerInterface` objects

        """
        raise NotImplementedError("decoders method is not implemented")


predefined_asr = {
    "pytorch": {
        "rnn": "espnet.nets.pytorch_backend.e2e_asr:E2E",
        "transformer": "espnet.nets.pytorch_backend.e2e_asr_transformer:E2E",
    },
    "chainer": {
        "rnn": "espnet.nets.chainer_backend.e2e_asr:E2E",
        "transformer": "espnet.nets.chainer_backend.e2e_asr_transformer:E2E",
    }
}


def dynamic_import_asr(module, backend):
    """Import ASR models dynamically.

    Args:
        module (str): module_name:class_name or alias in `predefined_asr`
        backend (str): NN backend. e.g., pytorch, chainer

    Returns:
        type: ASR class

    """
    model_class = dynamic_import(module, predefined_asr.get(backend, dict()))
    assert issubclass(model_class, ASRInterface), f"{module} does not implement ASRInterface"
    return model_class

Custom ASR model

Create new class implementing the ASRInterface in ./custom.py

For training, forward and add_arguments are needed

In [2]:
!cat ./custom.py
import chainer
import torch
from espnet.nets.asr_interface import ASRInterface
from espnet.nets.pytorch_backend.transformer.encoder import Encoder
from espnet.nets.pytorch_backend.transformer.decoder import Decoder
from espnet.nets.pytorch_backend.transformer.mask import subsequent_mask
from espnet.nets.pytorch_backend.transformer.label_smoothing_loss import LabelSmoothingLoss
from espnet.nets.pytorch_backend.nets_utils import make_pad_mask
from espnet.nets.pytorch_backend.nets_utils import th_accuracy

class Reporter(chainer.Chain):
    def report(self, **kwargs):
        chainer.reporter.report(kwargs, self)

class ASRTransformer(ASRInterface, torch.nn.Module):
    @staticmethod
    def add_arguments(parser):
        parser.add_argument("--label-smoothing", default=0.0, type=float)
        return parser

    def __init__(self, idim, odim, args=None):
        torch.nn.Module.__init__(self)
        self.encoder = Encoder(idim, input_layer="linear")
        self.decoder = Decoder(odim)
        self.criterion = LabelSmoothingLoss(odim, -1, args.label_smoothing, True)
        self.sos = odim - 1
        self.eos = odim - 1
        self.ignore_id=-1
        self.subsample = [0]
        self.reporter = Reporter()

    # for training
    def forward(self, xs_pad, ilens, ys_pad):
        """Compute scalar loss for backprop"""
        src_mask = (~make_pad_mask(ilens.tolist())).to(xs_pad.device).unsqueeze(-2)
        hs_pad, hs_mask = self.encoder(xs_pad, src_mask)

        ys_in_pad, ys_out_pad = self.add_sos_eos(ys_pad)
        ys_mask = self.target_mask(ys_in_pad)
        pred_pad, pred_mask = self.decoder(ys_in_pad, ys_mask, hs_pad, hs_mask)

        loss = self.criterion(pred_pad, ys_out_pad)
        self.acc = th_accuracy(pred_pad.view(-1, pred_pad.size(-1)), ys_out_pad, ignore_label=self.ignore_id)
        self.reporter.report(loss=loss, acc=self.acc)
        return loss

    # for decoding
    def encode(self, feat):
        """Encode speech feature."""
        return self.encoder(feat.unsqueeze(0), None)[0][0]

    def scorers(self):
        """Scorer used in beam search"""
        return {"decoder": self.decoder}

    def add_sos_eos(self, ys_pad):
        from espnet.nets.pytorch_backend.nets_utils import pad_list
        eos = ys_pad.new([self.eos])
        sos = ys_pad.new([self.sos])
        ys = [y[y != self.ignore_id] for y in ys_pad]  # parse padded ys
        ys_in = [torch.cat([sos, y], dim=0) for y in ys]
        ys_out = [torch.cat([y, eos], dim=0) for y in ys]
        return pad_list(ys_in, self.eos), pad_list(ys_out, self.ignore_id)

    def target_mask(self, ys_in_pad):
        ys_mask = ys_in_pad != self.ignore_id
        m = subsequent_mask(ys_mask.size(-1), device=ys_mask.device).unsqueeze(0)
        return ys_mask.unsqueeze(-2) & m

Custom train config

Create new config ./train_custom.yaml

In [4]:
!cat train_custom.yaml
model-module: custom:ASRTransformer  # module_name:class_name
label-smoothing: 0.2                 # model-specific option

batch-size: 8
epochs: 5
opt: adadelta
report-interval-iters: 4

Excute run.sh

using previous custom.py and train_custom.yaml

ESPnet provides log files and tensorboard for your custom model

In [7]:
!cwd=$(pwd); cd espnet/egs/an4/asr1; \
  PYTHONPATH=$cwd ./run.sh --ngpu 0 --stage 4 \
  --train-config $cwd/train_custom.yaml --verbose 0
dictionary: data/lang_1char/train_nodev_units.txt
stage 4: Network Training
/home/skarita/Documents/repos/espnet/tools/venv/lib/python3.7/site-packages/torch/nn/_reduction.py:49: UserWarning: size_average and reduce args will be deprecated, please use reduction='none' instead.
  warnings.warn(warning.format(ret))
2019-09-14 15:02:36,017 (asr_train:290) WARNING: Skip DEBUG/INFO messages
2019-09-14 15:02:36,552 (asr:294) WARNING: cuda is not available
epoch       iteration   main/loss   main/loss_ctc  main/loss_att  validation/main/loss  validation/main/loss_ctc  validation/main/loss_att  main/acc    validation/main/acc  main/cer_ctc  validation/main/cer_ctc  elapsed_time  eps       
0           4                                                                                                                               0.0975691                                                               9.9127        1e-08       
     total [..................................................]  0.75%
this epoch [#.................................................]  3.77%
         4 iter, 0 epoch / 5 epochs
       inf iters/sec. Estimated time to finish: 0:00:00.
0           8                                                                                                                               0.182341                                                                19.2326       1e-08       
     total [..................................................]  1.51%
this epoch [###...............................................]  7.55%
         8 iter, 0 epoch / 5 epochs
   0.42919 iters/sec. Estimated time to finish: 0:20:16.243812.
0           12                                                                                                                              0.195431                                                                28.7167       1e-08       
     total [#.................................................]  2.26%
this epoch [#####.............................................] 11.32%
        12 iter, 0 epoch / 5 epochs
   0.42527 iters/sec. Estimated time to finish: 0:20:18.040786.
0           16                                                                                                                              0.230607                                                                37.4209       1e-08       
     total [#.................................................]  3.02%
this epoch [#######...........................................] 15.09%
        16 iter, 0 epoch / 5 epochs
   0.43623 iters/sec. Estimated time to finish: 0:19:38.282889.
^C

Summary

ESPnet supports your research

  1. Many recipes and pretrained models
  2. Extensible design: you can reuse many tools in ESPnet
  3. Open community-driven development

send us your pull requests!

https://github.com/espnet/espnet

Thanks for your attention

Appendix Custom model training (manual for loop)

Train the model with prepared datasets in the section 3.2

In [0]:
# create minibatch data loader
import json
import kaldiio
from torch.nn.utils.rnn import pad_sequence
from torch.utils.data import DataLoader
from espnet.utils.training.batchfy import make_batchset

root = "espnet/egs/an4/asr1"
with open(root + "/dump/train_nodev/deltafalse/data.json", "r") as f:
    train_json = json.load(f)["utts"]
with open(root + "/dump/train_dev/deltafalse/data.json", "r") as f:
    dev_json = json.load(f)["utts"]

batch_size = 16
trainset = make_batchset(train_json, batch_size)
devset = make_batchset(dev_json, batch_size)

def collate(minibatch):
    fbanks = []
    tokens = []
    for key, info in minibatch[0]:
        fbanks.append(torch.tensor(kaldiio.load_mat(info["input"][0]["feat"])))
        tokens.append(torch.tensor([int(s) for s in info["output"][0]["tokenid"].split()]))
    ilens = torch.tensor([x.shape[0] for x in fbanks])
    return pad_sequence(fbanks, batch_first=True), ilens, pad_sequence(tokens, batch_first=True)

train_loader = DataLoader(trainset, collate_fn=collate, shuffle=True, pin_memory=True)
dev_loader = DataLoader(devset, collate_fn=collate, pin_memory=True)
In [0]:
# training iteration
import numpy
from torch.nn.utils.clip_grad import clip_grad_norm_

model = ASRTransformer(83, 30)
model.cuda()
optim = torch.optim.Adam(model.parameters(), lr=0.001, betas=(0.9, 0.98))
n_epoch = 30
for epoch in range(n_epoch):
    acc = []
    model.train()
    for data in train_loader:
        loss = model(*[d.cuda() for d in data])
        optim.zero_grad()
        loss.backward()
        acc.append(model.acc)
        norm = clip_grad_norm_(model.parameters(), 10.0)
        optim.step()
    train_acc = numpy.mean(acc)

    acc = []
    model.eval()
    for data in dev_loader:
        model(*[d.cuda() for d in data])
        acc.append(model.acc)
    valid_acc = numpy.mean(acc)
    print(f"epoch: {epoch}, train acc: {train_acc:.3f}, dev acc: {valid_acc:.3f}")
epoch: 0, train acc: 0.550, dev acc: 0.580
epoch: 1, train acc: 0.652, dev acc: 0.674
epoch: 2, train acc: 0.701, dev acc: 0.689
epoch: 3, train acc: 0.722, dev acc: 0.708
epoch: 4, train acc: 0.743, dev acc: 0.706
epoch: 5, train acc: 0.758, dev acc: 0.721
epoch: 6, train acc: 0.774, dev acc: 0.730
epoch: 7, train acc: 0.791, dev acc: 0.743
epoch: 8, train acc: 0.800, dev acc: 0.751
epoch: 9, train acc: 0.811, dev acc: 0.762
epoch: 10, train acc: 0.818, dev acc: 0.756
epoch: 11, train acc: 0.829, dev acc: 0.762
epoch: 12, train acc: 0.837, dev acc: 0.766
epoch: 13, train acc: 0.843, dev acc: 0.771
epoch: 14, train acc: 0.849, dev acc: 0.780
epoch: 15, train acc: 0.853, dev acc: 0.776
epoch: 16, train acc: 0.858, dev acc: 0.763
epoch: 17, train acc: 0.868, dev acc: 0.763
epoch: 18, train acc: 0.869, dev acc: 0.757
epoch: 19, train acc: 0.868, dev acc: 0.776
epoch: 20, train acc: 0.874, dev acc: 0.775
epoch: 21, train acc: 0.880, dev acc: 0.775
epoch: 22, train acc: 0.881, dev acc: 0.768
epoch: 23, train acc: 0.885, dev acc: 0.764
epoch: 24, train acc: 0.886, dev acc: 0.766
epoch: 25, train acc: 0.889, dev acc: 0.774
epoch: 26, train acc: 0.892, dev acc: 0.766
epoch: 27, train acc: 0.895, dev acc: 0.769
epoch: 28, train acc: 0.897, dev acc: 0.772
epoch: 29, train acc: 0.899, dev acc: 0.778
In [0]:
torch.save(model.state_dict(), "model.pt")
In [0]:
# Appendix audio example in AN4
import os
import kaldiio
from IPython.display import Audio


try:
  d = os.getcwd()
  os.chdir(root)
  sr, wav = kaldiio.load_scp("data/test/wav.scp")[key]
finally:
  os.chdir(d)
Audio(wav, rate=sr)
Out[0]: