Mask Rcnn total steps

[p.vahidinia]

Hi.
Is total steps in mask rcnn spec file same as number of epochs?

[Morganh]

See NVIDIA TAO Documentation
total_steps : The total number of training iterations

[p.vahidinia]

yes you are right. but in this link:

Poor metric results after retraining maskrcnn using TLT notebook

@ghazni After checking, your previous learning rate(0.005) is fine. But need to enlarge the total_step. total_steps = total_images * total_epochs / batch_size / nGPUs Your batchsize=2, the same as blog’s. Your training gpus is 2, while blog is using 8gpus. Your total_steps is 100k, while blog’s total_step is 100k. So, please increase total number of steps to 400k.

you mentioned that:
total_steps = total_images * total_epochs / batch_size / nGPUs

That is what I thought total steps and total epochs were different

[Morganh]

Yes, total_steps = total_images * total_epochs / batch_size / nGPUs