Issue training model on batch (ValueError: tf.function-decorated function tried to create variables on non-first call.)

[sehgalsakshi]

Hi,
I'm getting this issue even after annotating loss function with @tf.function.
I google more about it. Seems like loss function needs to be initialized at every epoch because it breaks after the first epoch!
Could you please help me out with this asap!

[sehgalsakshi]

@tf.function
def get_compiled_model():
model = create_model(args, maxlen, vocab)
model.get_layer('word_emb').trainable = False
model.compile(optimizer=optimizer, loss=max_margin_loss, metrics=[max_margin_loss])
return model

And on adding wrapper around model, I'm getting the below error

[JoannaSimm]

I'm getting the same error. @madrugado is there a solution for this problem?