Jinrui Zhang
Jinrui Zhang
Sorry for the late response, we have provided a config for reinforcement learning, you refer to this [cfg](https://github.com/zjr2000/GVL/blob/master/cfgs/anet_tsp_dvc_rl.yml) to perform reinforcement training
Thank you for pointing that out. I revisited my code and realized that I overlooked some elements related to rl during my previous cleanup. I have now updated the code...
> [#72 (comment)](https://github.com/state-spaces/mamba/issues/72#issuecomment-2027318310) > > > > > > same issue here when utilizing bi-mamba implementation refer to https://github.com/hustvl/Vim/blob/main/mamba-1p1p1/mamba_ssm/ops/selective_scan_interface.py. two ways can solve this: 1. fp32 training; 2. decrease lr....
1. Prepare the annotation file: ```train_caption_file```: training corpus, refer to this [file](https://github.com/zjr2000/GVL/blob/master/data/youmakeup/annotations/caption/train.json) ```val_caption_file```: validation corpus, refer to this [file](https://github.com/zjr2000/GVL/blob/master/data/youmakeup/annotations/caption/val.json) ```eval_gt_file_for_grounding```: validation file for video grounding, refer to this [file](https://github.com/zjr2000/GVL/blob/master/data/youmakeup/annotations/grounding/gt_val_for_grounding.json) ```dict_file```:...
Hello, and thank you for reaching out! To get started, please download the dataset annotations from the [official website](https://drive.google.com/drive/folders/10WfY6r6fFjlfR19zwC6G8kIFEcOQOQL2). Following that, you have the option to either download the videos...
> Is Kinetics-400 dataset okay？ It might not be okay; using the Kinetics-400 dataset could lead to missing out on some videos.
The architecture of the Video-Q Former adheres to the standard design of a vanilla Q-Former; however, the difference lies in its input features, which are the query tokens derived from...
> hi @zjr2000, @ttengwang , > > Thank you for sharing the work and code. I am unable to find the steps to perform multi-sentence grounding on own dataset. The...
You might want to check the file type of the `p2pd` executable. I encountered the same issue, and it turned out to be caused by a platform mismatch with `p2pd`....