boheumd / d-nerv Goto Github PK

Hello, I want to express my appreciation for your work. But I wanted to inquire about the ’--model‘ parameter in the evaluate script. I couldn't find it in the code. Shouldn't it be the ’--weight‘ parameter instead?

Hey hello, I really appreciate your work and still have a few questions to ask. I want to retrain D-NeRV on the ucf101 data set I made, but the extraction method is not the same as the extraction method in your paper where your video will be extacted for 2fps per second. and I would like to ask whether there will be a big difference in the final result between the way of save the keyframes directly in the png format in torch and your way of compressing the keyframes with other methods and saving them?

From the paper, I saw that D-NeRV can effectively express diverse videos. Does this mean that it has the ability to generalize to unseen videos, or can it only express all videos that have appeared in the training set?

Hi. I wander about did the model on the ucf101 data set converged after training for 800 epochs? Or will more training bring better results?

Why when I use distributed training, I get stuck here in "model = torch.nn.parallel.DistributedDataParallel(model.to(local_rank), device_ids=[local_rank], output_device=local_rank, find_unused_parameters=False)"?

Can you public source code to process original video data of UVG? Thank you so much.