Resolved: Specifying batch_size in many ways

Question

Hello,
Can I specify batch_size as one of the arguments in model.fit() method?
Would that work same way as predetermining batch_size using tf.batch method?

Thanks.

Answer 1

Hi,

Specifying the "batch_size" parameter in the fit method works for inputs such as numpy array. In the MNIST example we are working with dataset object, that's why we are separately shuffling and batching it.
When working with NumPy arrays, the batch size in the fit method automatically shuffles the dataset and batch it.

Best,
365 Team

Resolved: Specifying batch_size in many ways

Submit an answer

related questions