Basically i have two models, why training speed in Model 1 (30 mins per epoch) is twice faster than Model2 (1 hour per epoch). I am using keras (tf 2.9) on GPU NVIDIA A100, compute capacity 8. Why is Model 1 much faster than model 2?? Thanks!
I cannot embed media, and picture, so this is the link. link to structure
Hi @daisydaisy, Are you providing the same dataset to both models like the same batch_size, same number of training samples and providing the same type of pre processed data?
The training time also depends upon the number of training parameters the model is having. If possible could you please let us know the training parameters difference between model1 and model 2. Thank You.
Thank you so much Kiran, the params are all the same, except Dense layer (before output) in Model 1 is 64, but model 2 is 128. All other params are consistent