Swathikiran Sudhakaran

I am not sure what exactly the problem is. I would assume it to be the issue with the extracted frames. Can you also try to have some runs by...

Hi, I am not sure what is wrong. I had two more runs and got >37%. For visualization, we followed the approach from this repo https://github.com/alexandrosstergiou/Saliency-Tubes-Visual-Explanations-for-Spatio-Temporal-Convolutions You can remove the...

I used ffmpeg 4.0 for extracting the frames

You can find the log file here: https://drive.google.com/file/d/1oyDajU3EFHYdR_PpjzqlM6J7BDCVS-OJ/view?usp=sharing

This is with the setting from the paper. Evaluating in every epoch should not make any difference.

That was my answer. This log file corresponds to the setting from paper (from the published code).

I checked the scheduler and the learning rate should follow the trend in the log file I shared. Which pytorch version are you using?

The error shows that not all videos contain a minimum of 20 frames. Either update the sequence length to an appropriate value or update the dataloader to take into consideration...

Thank you for the quick reply. So is it not possible to get the results reported in the paper by running the published code without any changes? What is the...

I used the same prompt as above. The modifications I made are after [here ](https://github.com/microsoft/i-Code/blob/main/i-Code-Doc/benchmarker/utils/pregenerated.py#L214)as follows: 1) prepend the input_ids (item_dict["input_ids"]) with prompt_token_ids 2) prepend the attention mask (item_dict["attention_mask"]) with...