I hope this message finds you well. I have a query regarding the process of continuing training with pretrained model weights and would appreciate your guidance on this matter.
Specifically, I am wondering how to properly proceed with loading pretrained weights for model continuation. For instance, if I wish to load _./pretrained_model/videomamba_m16_k400_mask_ft_f8res224.pth, would it be sufficient to include the following line in my script?
bash
Copy code
--finetune ./pretrained_model/videomamba_m16_k400_mask_ft_f8_res224.pth
My confusion arises as, after following this procedure, the accuracy remains at 0% for the initial epochs of training on the k400 sub-dataset. I am keen to understand where I might be encountering an issue in this process.
I hope this message finds you well. I have a query regarding the process of continuing training with pretrained model weights and would appreciate your guidance on this matter.
Specifically, I am wondering how to properly proceed with loading pretrained weights for model continuation. For instance, if I wish to load _./pretrained_model/videomamba_m16_k400_mask_ft_f8res224.pth, would it be sufficient to include the following line in my script?
bash Copy code
--finetune ./pretrained_model/videomamba_m16_k400_mask_ft_f8_res224.pth
My confusion arises as, after following this procedure, the accuracy remains at 0% for the initial epochs of training on the k400 sub-dataset. I am keen to understand where I might be encountering an issue in this process.Thank you very much for your time and assistance.