is it possible to resume training progress from checkpoint obtained before(snapshot)

[xpeng]

Hi guys, i meet this certain situation, i confirm an amazing effect on some checkpoint but the training progress is stopped before. so i think if i can continue the training instead of discarding those intermediate result(for example have been trained 8K+ iters) and starting an new one(this will consume more unnecessary time), this will be very helpful to save our time.

or if i missed something in train.lua?

thanks in advance.

[HichameMoriceau]

Hi @xpeng,

I think you might want to add something such as net = torch.load(model_filename):cuda() where model_filename is the path to your chosen model.t7 file, if you make this "resume" feature a CLI argument this should be quite straightforward to implement. Hope this helps!

[xpeng]

@HichameMoriceau thanks for help, i will try it!