Hi,
using your train_flower.py script with my Dataset, an error occurred, that num_steps_per_epoch could not be converted from float to int.
I therefore explicitly casted num_batches_per_epoch into int, so that line 186 looked like this:
num_batches_per_epoch = int(dataset.num_samples / batch_size)
I believe that might have to do with what the batch size is. Nonetheless, it's a good quick fix and I've made the necessary changes. Thank you for raising this up!
Hi, using your train_flower.py script with my Dataset, an error occurred, that num_steps_per_epoch could not be converted from float to int.
I therefore explicitly casted num_batches_per_epoch into int, so that line 186 looked like this: num_batches_per_epoch = int(dataset.num_samples / batch_size)