Ponku

I agree with @dataplayer12. Whilst going through some tests in `prototype` the fetch all like behaviour might raise issues when dealing with constructors for future models we might plan on...

@datumbox Sorry about that, everything should be fine now. > @TeodorPoncu It seems that in a recent commit, you accidentally updated all the expected files for all models. Could you...

An approach similar to the one in the optical flow dataset would indeed lead to tuples with only one element @NicolasHug and require to user to make checks on the...

@datumbox There are multiple people [stating](https://github.com/megvii-research/CREStereo/issues/23) on the original repo that they have trouble reproducing the performance claimed by the authors with the training code provided by them. This is...

@datumbox I have already managed to work around both the NaN and some of the performance drop by simply adjusting the point at which the learning rate starts decaying, and...

Yes, this could be closed.

@datumbox no need to rush it to finish it by the time the internship is done. Happy to keep contributing even afterwards, since I want to add a proposal for...

Running the deployed weights with the following command: `torchrun --nproc_per_node=1 train.py --model maxvit_t --interpolation bicubic --batch-size 1 --test-only --weights MaxVit_T_Weights.IMAGENET1K_V1` Yields the following results: `Test: Acc@1 83.700 Acc@5 96.722`

Hey @tpet! Thank you for your input and feedback as it is very valuable! At that time the picture wasn't very clear of how users would prefer to interact with...

Hey @NicolasHug! I personally was not aware of that difference in parameter specification between PyTorch and TensorFlow. I do not recall that coming up during reviews (I've double checked with...