We load a pre-trained model.
from keras.applications.resnet50 import ResNet50
from keras.preprocessing import image
from keras.applications.resnet50 import preprocess_input, decode_predictions
import numpy as np
model = ResNet50(weights='imagenet')
Using TensorFlow backend.
Now we download the cat image. Source: https://stackoverflow.com/questions/384759/how-to-convert-a-pil-image-into-a-numpy-array.
url = "https://d1wn0q81ehzw6k.cloudfront.net/additional/thul/media/0eaa14d11e8930f5?w=400&h=400"
from PIL import Image
import requests
from io import BytesIO
response = requests.get(url)
raw_img = Image.open(BytesIO(response.content))
And transform it to a 224, 224 pixel image.
from skimage.transform import resize
img = np.array(raw_img.getdata()).reshape(raw_img.size[0], raw_img.size[1], 3).astype(np.float)
img = resize(img, (224, 224, 3), mode='constant')
We can plot this:
import matplotlib.pyplot as plt
%matplotlib inline
plt.imshow(img / 255.)
plt.colorbar()
<matplotlib.colorbar.Colorbar at 0x11e33fe10>
And transform it into a format that the machine learning algorithm expects.
x = image.img_to_array(raw_img.copy()) # expects a PIL image
x = resize(x.astype(np.float), (224, 224, 3), mode='constant')
x = np.expand_dims(x, axis=0) # adds a new dimension at the start
x = preprocess_input(x) # preprocesses it
preds = model.predict(x)
# decode the results into a list of tuples (class, description, probability)
# (one such list for each sample in the batch)
print('Predicted:', decode_predictions(preds, top=3)[0])
Predicted: [('n02123045', 'tabby', 0.72113997), ('n02123159', 'tiger_cat', 0.15921114), ('n02124075', 'Egyptian_cat', 0.096362673)]
Okay it works.
import tensorflow as tf
from cleverhans.utils_keras import KerasModelWrapper
from cleverhans.attacks import FastGradientMethod
import keras.backend
sess = tf.Session()
keras.backend.set_session(sess)
x_ph = tf.placeholder(tf.float32, shape=(None, 224, 224, 3))
# Initialize the Fast Gradient Sign Method (FGSM) attack object and graph
wrap = KerasModelWrapper(model)
fgsm = FastGradientMethod(wrap, sess=sess)
fgsm_params = {'eps': 0.3,
'clip_min': 0.,
'clip_max': 1.}
adv_x = fgsm.generate(x_ph, **fgsm_params)
# Consider the attack to be constant
adv_x = tf.stop_gradient(adv_x)
preds_adv = model(adv_x)
from cleverhans.utils_tf import model_eval
x.shape
(1, 224, 224, 3)
preds_adv
<tf.Tensor 'resnet50_1/fc1000/Softmax:0' shape=(?, 1000) dtype=float32>
keras.layers.core.K.set_learning_phase(0)
tf.global_variables_initializer().run(session=sess)
sess.run(preds_adv, {x_ph:x})
--------------------------------------------------------------------------- InvalidArgumentError Traceback (most recent call last) ~/anaconda/lib/python3.6/site-packages/tensorflow/python/client/session.py in _do_call(self, fn, *args) 1326 try: -> 1327 return fn(*args) 1328 except errors.OpError as e: ~/anaconda/lib/python3.6/site-packages/tensorflow/python/client/session.py in _run_fn(session, feed_dict, fetch_list, target_list, options, run_metadata) 1305 feed_dict, fetch_list, target_list, -> 1306 status, run_metadata) 1307 ~/anaconda/lib/python3.6/contextlib.py in __exit__(self, type, value, traceback) 88 try: ---> 89 next(self.gen) 90 except StopIteration: ~/anaconda/lib/python3.6/site-packages/tensorflow/python/framework/errors_impl.py in raise_exception_on_not_ok_status() 465 compat.as_text(pywrap_tensorflow.TF_Message(status)), --> 466 pywrap_tensorflow.TF_GetCode(status)) 467 finally: InvalidArgumentError: You must feed a value for placeholder tensor 'bn_conv1/keras_learning_phase' with dtype bool [[Node: bn_conv1/keras_learning_phase = Placeholder[dtype=DT_BOOL, shape=<unknown>, _device="/job:localhost/replica:0/task:0/cpu:0"]()]] During handling of the above exception, another exception occurred: InvalidArgumentError Traceback (most recent call last) <ipython-input-31-c577f35e605e> in <module>() 1 keras.layers.core.K.set_learning_phase(0) 2 tf.global_variables_initializer().run(session=sess) ----> 3 sess.run(preds_adv, {x_ph:x}) ~/anaconda/lib/python3.6/site-packages/tensorflow/python/client/session.py in run(self, fetches, feed_dict, options, run_metadata) 893 try: 894 result = self._run(None, fetches, feed_dict, options_ptr, --> 895 run_metadata_ptr) 896 if run_metadata: 897 proto_data = tf_session.TF_GetBuffer(run_metadata_ptr) ~/anaconda/lib/python3.6/site-packages/tensorflow/python/client/session.py in _run(self, handle, fetches, feed_dict, options, run_metadata) 1122 if final_fetches or final_targets or (handle and feed_dict_tensor): 1123 results = self._do_run(handle, final_targets, final_fetches, -> 1124 feed_dict_tensor, options, run_metadata) 1125 else: 1126 results = [] ~/anaconda/lib/python3.6/site-packages/tensorflow/python/client/session.py in _do_run(self, handle, target_list, fetch_list, feed_dict, options, run_metadata) 1319 if handle is None: 1320 return self._do_call(_run_fn, self._session, feeds, fetches, targets, -> 1321 options, run_metadata) 1322 else: 1323 return self._do_call(_prun_fn, self._session, handle, feeds, fetches) ~/anaconda/lib/python3.6/site-packages/tensorflow/python/client/session.py in _do_call(self, fn, *args) 1338 except KeyError: 1339 pass -> 1340 raise type(e)(node_def, op, message) 1341 1342 def _extend_graph(self): InvalidArgumentError: You must feed a value for placeholder tensor 'bn_conv1/keras_learning_phase' with dtype bool [[Node: bn_conv1/keras_learning_phase = Placeholder[dtype=DT_BOOL, shape=<unknown>, _device="/job:localhost/replica:0/task:0/cpu:0"]()]] Caused by op 'bn_conv1/keras_learning_phase', defined at: File "/Users/kappamaki/anaconda/lib/python3.6/runpy.py", line 193, in _run_module_as_main "__main__", mod_spec) File "/Users/kappamaki/anaconda/lib/python3.6/runpy.py", line 85, in _run_code exec(code, run_globals) File "/Users/kappamaki/anaconda/lib/python3.6/site-packages/ipykernel_launcher.py", line 16, in <module> app.launch_new_instance() File "/Users/kappamaki/anaconda/lib/python3.6/site-packages/traitlets/config/application.py", line 658, in launch_instance app.start() File "/Users/kappamaki/anaconda/lib/python3.6/site-packages/ipykernel/kernelapp.py", line 477, in start ioloop.IOLoop.instance().start() File "/Users/kappamaki/anaconda/lib/python3.6/site-packages/zmq/eventloop/ioloop.py", line 177, in start super(ZMQIOLoop, self).start() File "/Users/kappamaki/anaconda/lib/python3.6/site-packages/tornado/ioloop.py", line 888, in start handler_func(fd_obj, events) File "/Users/kappamaki/anaconda/lib/python3.6/site-packages/tornado/stack_context.py", line 277, in null_wrapper return fn(*args, **kwargs) File "/Users/kappamaki/anaconda/lib/python3.6/site-packages/zmq/eventloop/zmqstream.py", line 440, in _handle_events self._handle_recv() File "/Users/kappamaki/anaconda/lib/python3.6/site-packages/zmq/eventloop/zmqstream.py", line 472, in _handle_recv self._run_callback(callback, msg) File "/Users/kappamaki/anaconda/lib/python3.6/site-packages/zmq/eventloop/zmqstream.py", line 414, in _run_callback callback(*args, **kwargs) File "/Users/kappamaki/anaconda/lib/python3.6/site-packages/tornado/stack_context.py", line 277, in null_wrapper return fn(*args, **kwargs) File "/Users/kappamaki/anaconda/lib/python3.6/site-packages/ipykernel/kernelbase.py", line 283, in dispatcher return self.dispatch_shell(stream, msg) File "/Users/kappamaki/anaconda/lib/python3.6/site-packages/ipykernel/kernelbase.py", line 235, in dispatch_shell handler(stream, idents, msg) File "/Users/kappamaki/anaconda/lib/python3.6/site-packages/ipykernel/kernelbase.py", line 399, in execute_request user_expressions, allow_stdin) File "/Users/kappamaki/anaconda/lib/python3.6/site-packages/ipykernel/ipkernel.py", line 196, in do_execute res = shell.run_cell(code, store_history=store_history, silent=silent) File "/Users/kappamaki/anaconda/lib/python3.6/site-packages/ipykernel/zmqshell.py", line 533, in run_cell return super(ZMQInteractiveShell, self).run_cell(*args, **kwargs) File "/Users/kappamaki/anaconda/lib/python3.6/site-packages/IPython/core/interactiveshell.py", line 2728, in run_cell interactivity=interactivity, compiler=compiler, result=result) File "/Users/kappamaki/anaconda/lib/python3.6/site-packages/IPython/core/interactiveshell.py", line 2850, in run_ast_nodes if self.run_code(code, result): File "/Users/kappamaki/anaconda/lib/python3.6/site-packages/IPython/core/interactiveshell.py", line 2910, in run_code exec(code_obj, self.user_global_ns, self.user_ns) File "<ipython-input-1-3cd75d6c9839>", line 6, in <module> model = ResNet50(weights='imagenet') File "/Users/kappamaki/anaconda/lib/python3.6/site-packages/keras/applications/resnet50.py", line 208, in ResNet50 x = BatchNormalization(axis=bn_axis, name='bn_conv1')(x) File "/Users/kappamaki/anaconda/lib/python3.6/site-packages/keras/engine/topology.py", line 602, in __call__ output = self.call(inputs, **kwargs) File "/Users/kappamaki/anaconda/lib/python3.6/site-packages/keras/layers/normalization.py", line 190, in call training=training) File "/Users/kappamaki/anaconda/lib/python3.6/site-packages/keras/backend/tensorflow_backend.py", line 2610, in in_train_phase training = learning_phase() File "/Users/kappamaki/anaconda/lib/python3.6/site-packages/keras/backend/tensorflow_backend.py", line 115, in learning_phase name='keras_learning_phase') File "/Users/kappamaki/anaconda/lib/python3.6/site-packages/tensorflow/python/ops/array_ops.py", line 1548, in placeholder return gen_array_ops._placeholder(dtype=dtype, shape=shape, name=name) File "/Users/kappamaki/anaconda/lib/python3.6/site-packages/tensorflow/python/ops/gen_array_ops.py", line 2094, in _placeholder name=name) File "/Users/kappamaki/anaconda/lib/python3.6/site-packages/tensorflow/python/framework/op_def_library.py", line 767, in apply_op op_def=op_def) File "/Users/kappamaki/anaconda/lib/python3.6/site-packages/tensorflow/python/framework/ops.py", line 2630, in create_op original_op=self._default_original_op, op_def=op_def) File "/Users/kappamaki/anaconda/lib/python3.6/site-packages/tensorflow/python/framework/ops.py", line 1204, in __init__ self._traceback = self._graph._extract_stack() # pylint: disable=protected-access InvalidArgumentError (see above for traceback): You must feed a value for placeholder tensor 'bn_conv1/keras_learning_phase' with dtype bool [[Node: bn_conv1/keras_learning_phase = Placeholder[dtype=DT_BOOL, shape=<unknown>, _device="/job:localhost/replica:0/task:0/cpu:0"]()]]
model_eval(sess, x, y )
preds_adv.