MaoYuxin

I try to implement it by https://github.com/ClementPinard/Pytorch-Correlation-extension, and get a local correlation matrix. Next I send it to correlation pooling and lookup module. But doing so directly causes the network...

Yes, follow the Readme, I generate the flow by liteflownet, and then perform training process by `python -m torch.distributed.launch --nproc_per_node=4 train_vimeo90k.py --world_size 4 --model_name 'IFRNet' --epochs 300 --batch_size 6 --lr_start...

I apologize for my mistake, but the accuracy in the paper should be 35.59, which is a huge difference from the results I trained, and I think this is the...

In addition, in the log I provided, the optimal precision should be 35.46, which I incorrectly described as 34.45. I am sorry for my mistake. But 35.46 is still 0.13dB...

Thank you very much for your reply, I will try the way you said. Good luck with your future research, thank you again.

感谢回复！

Thank you so much for such a quick reply, I will try your solution, thank you!

Hello, after I replaced it with bf16, the NAN problem still occurred. ``` torchrun --standalone --nproc_per_node=8 train.py \ --model DiT-XL/2 \ --batch_size 36 \ --ckpt_every 10000\ --num_workers 12 \ --enable_modulate_kernel...

Hello, I am running the CIFAR10 demo. According to the script in the readme, I expanded the batch size and the number of graphics cards, and left the rest unchanged.

Thank you so much!