0
$\begingroup$

Categories to learn and predict:

df.race.unique()
array(['0', '1', '3', '2', '4'], dtype=object)

Data:

train_generator = image_gen.flow_from_dataframe(
    df_train,
    x_col="img_name",
    y_col="race",
    directory=str(data_folder),
    class_mode="sparse",
    target_size=(IMAGE_SIZE, IMAGE_SIZE),
    batch_size=BATCH_SIZE,
    shuffle=True,
)

val_generator = image_gen.flow_from_dataframe(
    df_val,
    x_col="img_name",
    y_col="race",
    directory=str(data_folder),
    class_mode="sparse",
    target_size=(IMAGE_SIZE, IMAGE_SIZE),
    batch_size=BATCH_SIZE,
    shuffle=False,
)

Model load and fit:

vggface_model = load_model("resnet50face.h5")
base_model = tf.keras.Model([vggface_model.input], vggface_model.get_layer("flatten_1").output)
base_model.trainable = False

last_layer = base_model.get_layer('avg_pool').output
hidden_layer = Flatten(name='flatten')(last_layer)
out_layer = Dense(5, activation='softmax', name='gender_classifier')(hidden_layer)
custom_base_model = tf.keras.Model(base_model.input, out_layer)

custom_base_model.compile(
              optimizer=tf.keras.optimizers.Adam(learning_rate=0.0001),
              loss="categorical_crossentropy",
              metrics=['accuracy'])

history = custom_base_model.fit(
    x=train_generator, 
    validation_data=val_generator, 
    steps_per_epoch=20, 
    epochs=40)

Error message:

ValueError: in user code:

    /usr/local/lib/python3.7/dist-packages/tensorflow/python/keras/engine/training.py:855 train_function  *
        return step_function(self, iterator)
    /usr/local/lib/python3.7/dist-packages/tensorflow/python/keras/engine/training.py:845 step_function  **
        outputs = model.distribute_strategy.run(run_step, args=(data,))
    /usr/local/lib/python3.7/dist-packages/tensorflow/python/distribute/distribute_lib.py:1285 run
        return self._extended.call_for_each_replica(fn, args=args, kwargs=kwargs)
    /usr/local/lib/python3.7/dist-packages/tensorflow/python/distribute/distribute_lib.py:2833 call_for_each_replica
        return self._call_for_each_replica(fn, args, kwargs)
    /usr/local/lib/python3.7/dist-packages/tensorflow/python/distribute/distribute_lib.py:3608 _call_for_each_replica
        return fn(*args, **kwargs)
    /usr/local/lib/python3.7/dist-packages/tensorflow/python/keras/engine/training.py:838 run_step  **
        outputs = model.train_step(data)
    /usr/local/lib/python3.7/dist-packages/tensorflow/python/keras/engine/training.py:797 train_step
        y, y_pred, sample_weight, regularization_losses=self.losses)
    /usr/local/lib/python3.7/dist-packages/tensorflow/python/keras/engine/compile_utils.py:204 __call__
        loss_value = loss_obj(y_t, y_p, sample_weight=sw)
    /usr/local/lib/python3.7/dist-packages/tensorflow/python/keras/losses.py:155 __call__
        losses = call_fn(y_true, y_pred)
    /usr/local/lib/python3.7/dist-packages/tensorflow/python/keras/losses.py:259 call  **
        return ag_fn(y_true, y_pred, **self._fn_kwargs)
    /usr/local/lib/python3.7/dist-packages/tensorflow/python/util/dispatch.py:206 wrapper
        return target(*args, **kwargs)
    /usr/local/lib/python3.7/dist-packages/tensorflow/python/keras/losses.py:1644 categorical_crossentropy
        y_true, y_pred, from_logits=from_logits)
    /usr/local/lib/python3.7/dist-packages/tensorflow/python/util/dispatch.py:206 wrapper
        return target(*args, **kwargs)
    /usr/local/lib/python3.7/dist-packages/tensorflow/python/keras/backend.py:4862 categorical_crossentropy
        target.shape.assert_is_compatible_with(output.shape)
    /usr/local/lib/python3.7/dist-packages/tensorflow/python/framework/tensor_shape.py:1161 assert_is_compatible_with
        raise ValueError("Shapes %s and %s are incompatible" % (self, other))

    ValueError: Shapes (None, 1) and (None, 5) are incompatible
$\endgroup$
2
  • $\begingroup$ Have you tried converting your labels to a one-hot encoding? I think at the moment your generator returns the label index, which is a single value, whereas you have created a model with neurons in the last layer. $\endgroup$
    – Oxbowerce
    Jun 17, 2021 at 13:19
  • $\begingroup$ It looks strange but it start working with class_mode="categorical" in flow_from_dataframe(. Strange becase I check twice categorical is "2D numpy array .." from this doc here. Maybe becase I used str obj labels in data frame (not int)? $\endgroup$
    – Vyacheslav
    Jun 17, 2021 at 13:28

1 Answer 1

0
$\begingroup$

Problem was solved when I changed:

class_mode="sparse"

to:

class_mode="categorical"

for both data sources in image_gen.flow_from_dataframe(..)

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.