tensorflow
TFMA and TF KERAS 2.0 model on pretrained model
Hello all,
I am referring here to stackoverflow that I have published couple of days ago: [https://stackoverflow.com/questions/56248024/tensorflow-model-analysis-tfma-for-keras-model] I didn't receive any response. There might not be too many people that uses TFX with TF KERAS 2.0 at the moment. So, I am trying my luck here. In general, I want to analyze a pre-trained (VGG16) model. The model was
- imported with TF KERAS 2.0 API
- saved
- loaded and converted to estimator using the keras to estimator API
However, the execution requires all VGG features in TF features format. What is the right way to extract these features? Can anyone refer me to an example where TFX is being used with pre-trained model.
The code is available at the stackoverflow. if easer, I can copy it to here - let me know Many thanks, eilalan
TF2.0 is not yet supported. We are actively working on it, so check back in a few weeks.
I understand. Therefore, I used
estimator_model = tf.keras.estimator.model_to_estimator(new_model,model_dir=TF_MODEL_DIR)]]
The error is raising from the VGG model that is being transferred
`# Load model
new_model = keras.models.load_model(model_name)
new_model.summary()
# keras model to estimator
estimator_model = tf.keras.estimator.model_to_estimator(new_model,model_dir=TF_MODEL_DIR)]]
# The receiver function for the estimator
def eval_input_receiver_1_fn():
serialized_tf_example = tf.compat.v1.placeholder(dtype=tf.string, shape=[None], name='input_example_placeholder')
receiver_tensors = {'examples': serialized_tf_example}
validation_features_columns = [tf.feature_column.numeric_column("image",shape=(192,192)),
tf.feature_column.categorical_column_with_vocabulary_list("label",["normal_healthy","sick"])]
feature_spec = tf.feature_column.make_parse_example_spec(validation_features_columns)
features = tf.io.parse_example(serialized_tf_example, feature_spec)
return tfma.export.EvalInputReceiver(
features=features,
receiver_tensors=receiver_tensors,
labels=features['label'])
import os
import shutil
from pathlib import Path
def up_one_dir(path):
"""Move all file in path up one
"""
parent_dir = str(Path(path).parents[0])
for f in os.listdir(path):
shutil.copy(os.path.join(path,f),parent_dir)
#shutil.rmtree(path)
up_one_dir(KERAS_FOLDER)
tfma.export.export_eval_savedmodel(estimator=estimator_model,
export_dir_base=EXPORT_DIR,
eval_input_receiver_fn=eval_input_receiver_1_fn)
The following error is fired regarding the pre-trained model features:
KeyErrorTraceback (most recent call last)
<ipython-input-137-b275096a314a> in <module>()
1 tfma.export.export_eval_savedmodel(estimator=estimator_model,
2 export_dir_base=EXPORT_DIR,
----> 3 eval_input_receiver_fn=eval_input_receiver_1_fn)
/usr/local/envs/py2env/lib/python2.7/site-packages/tensorflow_model_analysis/util.pyc in wrapped_fn(*args, **kwargs)
171 (fn.__name__, kwargs.keys()))
172
--> 173 return fn(**kwargs_to_pass)
174
175 return wrapped_fn
/usr/local/envs/py2env/lib/python2.7/site-packages/tensorflow_model_analysis/eval_saved_model/export.pyc in export_eval_savedmodel(estimator, export_dir_base, eval_input_receiver_fn, serving_input_receiver_fn, assets_extra, checkpoint_path)
472 },
473 assets_extra=assets_extra,
--> 474 checkpoint_path=checkpoint_path)
475
476
/usr/local/envs/py2env/lib/python2.7/site-packages/tensorflow/python/util/deprecation.pyc in new_func(*args, **kwargs)
322 'in a future version' if date is None else ('after %s' % date),
323 instructions)
--> 324 return func(*args, **kwargs)
325 return tf_decorator.make_decorator(
326 func, new_func, 'deprecated',
/usr/local/envs/py2env/lib/python2.7/site-packages/tensorflow_estimator/contrib/estimator/python/estimator/export.pyc in export_all_saved_models(estimator, export_dir_base, input_receiver_fn_map, assets_extra, as_text, checkpoint_path)
206 assets_extra=assets_extra,
207 as_text=as_text,
--> 208 checkpoint_path=checkpoint_path)
/usr/local/envs/py2env/lib/python2.7/site-packages/tensorflow_estimator/python/estimator/estimator.pyc in experimental_export_all_saved_models(self, export_dir_base, input_receiver_fn_map, assets_extra, as_text, checkpoint_path)
820 self._add_meta_graph_for_mode(
821 builder, input_receiver_fn_map, checkpoint_path,
--> 822 save_variables, mode=model_fn_lib.ModeKeys.EVAL)
823 save_variables = False
824 if input_receiver_fn_map.get(model_fn_lib.ModeKeys.PREDICT):
/usr/local/envs/py2env/lib/python2.7/site-packages/tensorflow_estimator/python/estimator/estimator.pyc in _add_meta_graph_for_mode(self, builder, input_receiver_fn_map, checkpoint_path, save_variables, mode, export_tags, check_variables)
895 labels=getattr(input_receiver, 'labels', None),
896 mode=mode,
--> 897 config=self.config)
898
899 export_outputs = model_fn_lib.export_outputs_for_mode(
/usr/local/envs/py2env/lib/python2.7/site-packages/tensorflow_estimator/python/estimator/estimator.pyc in _call_model_fn(self, features, labels, mode, config)
1110
1111 logging.info('Calling model_fn.')
-> 1112 model_fn_results = self._model_fn(features=features, **kwargs)
1113 logging.info('Done calling model_fn.')
1114
/usr/local/envs/py2env/lib/python2.7/site-packages/tensorflow_estimator/python/estimator/keras.pyc in model_fn(features, labels, mode)
276
277 model = _clone_and_build_model(mode, keras_model, custom_objects, features,
--> 278 labels)
279 model_output_names = []
280 # We need to make sure that the output names of the last layer in the model
/usr/local/envs/py2env/lib/python2.7/site-packages/tensorflow_estimator/python/estimator/keras.pyc in _clone_and_build_model(mode, keras_model, custom_objects, features, labels)
184 K.set_learning_phase(mode == model_fn_lib.ModeKeys.TRAIN)
185 input_tensors, target_tensors = _convert_estimator_io_to_keras(
--> 186 keras_model, features, labels)
187
188 compile_clone = (mode != model_fn_lib.ModeKeys.PREDICT)
/usr/local/envs/py2env/lib/python2.7/site-packages/tensorflow_estimator/python/estimator/keras.pyc in _convert_estimator_io_to_keras(keras_model, features, labels)
157
158 input_tensors = _to_ordered_tensor_list(
--> 159 features, input_names, 'features', 'inputs')
160 target_tensors = _to_ordered_tensor_list(
161 labels, output_names, 'labels', 'outputs')
/usr/local/envs/py2env/lib/python2.7/site-packages/tensorflow_estimator/python/estimator/keras.pyc in _to_ordered_tensor_list(obj, key_order, obj_name, order_name)
139 order_name=order_name, order_keys=set(key_order),
140 obj_name=obj_name, obj_keys=set(obj.keys()),
--> 141 different_keys=different_keys))
142
143 return [_convert_tensor(obj[key]) for key in key_order]
KeyError: "The dictionary passed into features does not have the expected inputs keys defined in the keras model.\n\tExpected keys: set([u'vgg16_input'])\n\tfeatures keys: set(['image', 'label'])\n\tDifference: set(['image', 'label', u'vgg16_input'])"
Can I make that work?
I should have been clearer. Support for TF 2.0 estimators (keras model_to_estimator or otherwise) is not yet supported.
That said, the error you are getting is because keras does not allow you to pass more features than the model defines. You need to either add vgg16_input to the model or remove it from the features.
Thank you so much for your response. will try to add to my features list. I want to make sure that I can have viz support for transfer models. Event if it is not perfect at the moment. will keep you posted if I have any additional issue. Thanks again, eilalan
It doesn't look like that I will able to load the TF 2.0 model with TF 1.13.1 that supports TFMA - an improper config error is being fired. I will wait for your updated version. Meanwhile, please let me know if you are familiar with other tools to analyze models that is saved as h5. Thanks, eilalan
I'm having a similar issue when attempting to use TFMA to slice and display statistics on raw features before they have gone through the TFTransform component. Are there any updates on this issue?
As of now, keras model_to_estimator requires that only the features used by the model be passed, so you would not be able to slice on additional features without additional work run another feature extractor in the pipeline.
+cc @tanzhenyu