问题
I am facing below error when trying to train a multi-class classification model ( 4 classes) for Image dataset. Even though my output tensor is of shape 4 I am facing below issue. Please let me know how to fix this issue.
Epoch 1/10
---------------------------------------------------------------------------
InvalidArgumentError Traceback (most recent call last)
<ipython-input-30-01c6f78f4d4f> in <module>
4 epochs=epochs,
5 validation_data=val_data_gen,
----> 6 validation_steps=total_val // batch_size
7 )
/usr/local/lib/python3.6/dist-packages/tensorflow_core/python/keras/engine/training.py in fit_generator(self, generator, steps_per_epoch, epochs, verbose, callbacks, validation_data, validation_steps, validation_freq, class_weight, max_queue_size, workers, use_multiprocessing, shuffle, initial_epoch)
1294 shuffle=shuffle,
1295 initial_epoch=initial_epoch,
-> 1296 steps_name='steps_per_epoch')
1297
1298 def evaluate_generator(self,
/usr/local/lib/python3.6/dist-packages/tensorflow_core/python/keras/engine/training_generator.py in model_iteration(model, data, steps_per_epoch, epochs, verbose, callbacks, validation_data, validation_steps, validation_freq, class_weight, max_queue_size, workers, use_multiprocessing, shuffle, initial_epoch, mode, batch_size, steps_name, **kwargs)
263
264 is_deferred = not model._is_compiled
--> 265 batch_outs = batch_function(*batch_data)
266 if not isinstance(batch_outs, list):
267 batch_outs = [batch_outs]
/usr/local/lib/python3.6/dist-packages/tensorflow_core/python/keras/engine/training.py in train_on_batch(self, x, y, sample_weight, class_weight, reset_metrics)
1015 self._update_sample_weight_modes(sample_weights=sample_weights)
1016 self._make_train_function()
-> 1017 outputs = self.train_function(ins) # pylint: disable=not-callable
1018
1019 if reset_metrics:
/usr/local/lib/python3.6/dist-packages/tensorflow_core/python/keras/backend.py in __call__(self, inputs)
3474
3475 fetched = self._callable_fn(*array_vals,
-> 3476 run_metadata=self.run_metadata)
3477 self._call_fetch_callbacks(fetched[-len(self._fetches):])
3478 output_structure = nest.pack_sequence_as(
/usr/local/lib/python3.6/dist-packages/tensorflow_core/python/client/session.py in __call__(self, *args, **kwargs)
1470 ret = tf_session.TF_SessionRunCallable(self._session._session,
1471 self._handle, args,
-> 1472 run_metadata_ptr)
1473 if run_metadata:
1474 proto_data = tf_session.TF_GetBuffer(run_metadata_ptr)
InvalidArgumentError: 2 root error(s) found.
(0) Invalid argument: Incompatible shapes: [4,3] vs. [4,4]
[[{{node loss_2/predictions_loss/logistic_loss/mul}}]]
[[loss_2/mul/_19047]]
(1) Invalid argument: Incompatible shapes: [4,3] vs. [4,4]
[[{{node loss_2/predictions_loss/logistic_loss/mul}}]]
0 successful operations.
0 derived errors ignored.
My batch size is 4 and below is last few layers of my model.
conv5_block16_2_conv (Conv2D) (None, 16, 16, 32) 36864 conv5_block16_1_relu[0][0]
__________________________________________________________________________________________________
conv5_block16_concat (Concatena (None, 16, 16, 1024) 0 conv5_block15_concat[0][0]
conv5_block16_2_conv[0][0]
__________________________________________________________________________________________________
bn (BatchNormalization) (None, 16, 16, 1024) 4096 conv5_block16_concat[0][0]
__________________________________________________________________________________________________
relu (Activation) (None, 16, 16, 1024) 0 bn[0][0]
__________________________________________________________________________________________________
avg_pool (GlobalAveragePooling2 (None, 1024) 0 relu[0][0]
__________________________________________________________________________________________________
predictions (Dense) (None, 4) 4100 avg_pool[0][0]
==================================================================================================
Loss function
model.compile(optimizer='adam',
loss=tf.keras.losses.BinaryCrossentropy(from_logits=True))
回答1:
I think, there is nothing wrong with the shapes, but with the loss function, you are trying to use. Ideally for multiclass classification, the final layer has to have softmax activation (for your logits to sum up to 1) and use CategoricalCrossentropy as your loss function if your labels are one-hot and SparseCategoricalCrossentropy if your labels are integers. Tensorflow documentation attached below. https://www.tensorflow.org/api_docs/python/tf/keras/losses/CategoricalCrossentropy
Changes that are to be done to your code
# adding softmax activation to final dense layer
predictions = Dense(4, activation='softmax')(avg_pool)
# assuming you have one-hot labels
model.compile(optimizer='adam',loss=tf.keras.losses.CategoricalCrossentropy(from_logits=True))
来源：https://stackoverflow.com/questions/61054359/invalidargumenterror-2-root-errors-found-0-invalid-argument-incompatible