Batch Size in Seq2Seq Model Causing Mismatch Dimension Error

[NickShahML]

Hey Fariz, I've been really struggling with this error, and frankly, I'm not sure how to address it.

You give a batch_size argument for your seq2seq model, and I'm wondering how this batch_size should relate to the batch_size parameter of model.fit

When I make batch_size of model.fit and seq2seq equivalent: I get the following error:

IndexError: index 4 is out of bounds for axis 1 with size 4

I have tried making the batch_size of seq2seq one less and one more than the model.fit batch_size and I'm met with this error:

Input dimension mis-match. Input 1 (indices start at 0) has shape[0] == 3, but the output's size on that axis is 4.

What exactly the difference between the batch_size of seq2seq and batch_size of model.fit()? Any help is greatly appreciated!

[farizrahman4u]

@LeavesBreathe batch_size in seq2seq is same as the batch_size in model.fit. Make sure that the number of samples in your training data is divisible by batch_size.

[NickShahML]

Okay I will reevaluate and try that. I'll close this assuming it works.