I was wondering if someone here might take a look at an issue I put in with the keras team. I think my rationale is correct, but I might be missing something. Short answer, I had to modify the Lambda/matmul layer as it didn't seem to work with multi_gpu_model. It was using the full batchsize instead of the batchsize/GPUs that I was expecting. I moved to the Dot layer with axes=2 and it seems to work.
I have problem with saving and loading the model. And I think is because of the lambda layer.
How did you changed the lambda layer to dot layer with axes=2?
Hello,
I was wondering if someone here might take a look at an issue I put in with the keras team. I think my rationale is correct, but I might be missing something. Short answer, I had to modify the Lambda/matmul layer as it didn't seem to work with multi_gpu_model. It was using the full batchsize instead of the batchsize/GPUs that I was expecting. I moved to the Dot layer with axes=2 and it seems to work.
Ticket