Hello!
I'm trying to implement a LSTM model for anomaly detection in OpenVino
I got the Keras model from a blog post then converted it to a frozen Tensorflow model following another, but I'm stuck at the "optimising the model and converting it to IR" part.
I'm getting that error:
[ ERROR ] Shape [ -1 10 256 256 1] is not fully defined for output 0 of "time_distributed_1_input_4". Use --input_shape with positive integers to override model input shapes.
[ ERROR ] Cannot infer shapes or values for node "time_distributed_1_input_4".
I understand that we need to specify the batch size, but the original model seems to have a flexible one, am I forced to fix it in the code and retrain it? Or is there a workaround?
Thanks a lot.
Additional info:
Shape of the model:
Here is the part related to model creation:
seq = Sequential()
seq.add(TimeDistributed(Conv2D(128, (11, 11), strides=4, padding="same"), batch_input_shape=(None, 10, 256, 256, 1)))
seq.add(LayerNormalization())
seq.add(TimeDistributed(Conv2D(64, (5, 5), strides=2, padding="same")))
seq.add(LayerNormalization())
# # # # #
seq.add(ConvLSTM2D(64, (3, 3), padding="same", return_sequences=True))
seq.add(LayerNormalization())
seq.add(ConvLSTM2D(32, (3, 3), padding="same", return_sequences=True))
seq.add(LayerNormalization())
seq.add(ConvLSTM2D(64, (3, 3), padding="same", return_sequences=True))
seq.add(LayerNormalization())
# # # # #
seq.add(TimeDistributed(Conv2DTranspose(64, (5, 5), strides=2, padding="same")))
seq.add(LayerNormalization())
seq.add(TimeDistributed(Conv2DTranspose(128, (11, 11), strides=4, padding="same")))
seq.add(LayerNormalization())
seq.add(TimeDistributed(Conv2D(1, (11, 11), activation="sigmoid", padding="same")))
print(seq.summary())
seq.compile(loss='mse', optimizer=tf.keras.optimizers.Adam(lr=1e-4, decay=1e-5, epsilon=1e-6))
seq.fit(training_set, training_set,
batch_size=Config.BATCH_SIZE, epochs=Config.EPOCHS, shuffle=False)
Here are the errors I got in the workbench when removing a dimension, or replacing it with 190 and 200:
Link Copied
