search

ClementPinard / FlowNetPytorch

Pytorch implementation of FlowNet by Dosovitskiy et al.
MIT License
843 stars 205 forks source link

start_epoch called before parsing the argument #118

Open rashadkp opened 6 months ago

rashadkp commented 6 months ago

https://github.com/ClementPinard/FlowNetPytorch/blob/60c72a9c054e6e525eca5cc129ccafc11758d88f/main.py#L91C1-L91C7

in the line 91 ,it should be : n_iter = int(args.start_epoch)

ClementPinard commented 6 months ago

Actually, it should be 0, n_iter is only here for tensorflow purpose. I am surprised this does not make the code crash, since the variable start_epoch is not defined :thinking:

We could probably do better by inferring the n_iter from args.start_epoch and dataset size (like len(train_set) * args.start_epoch )

ClementPinard commented 6 months ago

Should be solved with last commits. Please check it out ! :sunglasses: