SimoLoca

@klauscc

Hi @klauscc , thanks for the suggestion! At the moment I tried using a loop until a valid segments has not been found and it seems to work. I have...

Hi, thanks for the reply. Regarding question 3, can you please explain me why should I need to change that part of the code?

Hi @klauscc, thanks for the fast reply. I tried as you have mentioned, precisely after L23 I've inserted: ```python for name, param in looper.train_engine.model.backbone.named_parameters(): print("name: ", name, "grad: ", param.grad)...

I check that following the readme for training the model, without loading checkpoint. And also I did not change the config file. Did I get it wrong the way to...

Hi @klauscc, I've resolved the issue. There are no problem with the code, there were some errors on my config file, so forgive me if i disturbed you too much....

Ok, thanks you. So I need to make some changes in [SwinTransformer3D](https://github.com/klauscc/TALLFormer/blob/5519140e39095cd87d9b50420bde912975cae9fb/vedatad/models/backbones/vswin.py#L587) or in [ChunkVideoSwin](https://github.com/klauscc/TALLFormer/blob/5519140e39095cd87d9b50420bde912975cae9fb/vedatad/models/backbones/chunk_model.py#L21)?

If I'm not mistaken, it's because when training the model it will employ the video's memory bank to alleviate workload on the GPU, while when testing it will load the...