AR-Net
AR-Net copied to clipboard
mengyuest
What a wonderful work! Hello , I have read your paper and benefited a lot. But now I have some problems. I can't find the Fudan-Columbia Video Dataset（FCVID）and its homepage...
Hello, Thank you for your excellent work. I am preparing ActivityNet, FCVID, and MiniKinetics datasets following the [instructions](https://github.com/mengyuest/AR-Net#dataset-preparation) provided in README.md. I understand that [./ops/video_jpg.py](https://github.com/mengyuest/AR-Net/blob/master/ops/video_jpg.py) extracts frames using ffmpeg which...
It's definitly a through work for action recogintion compression problem. But I still wondering why you put a different weight for acc_loss and eff_loss？ Each loss supervise each module without...
Hi Yue, thanks very much for this very interesting project and your code. I have one question on the implementation of the policy network. On [this line](https://github.com/mengyuest/AR-Net/blob/2a9d72b50fed390e58cba6b3af368e3301ec078f/ops/models_ada.py#L328), we will use...
Hello teacher, I have read your AR-Net article and benefited a lot. But now I have some problems. I can't find the ActivityNet, FCVID and mini-Kinetics datasets. Is it convenient...
Hi, this is a nice paper. How to generate different AR-Net models with different flops(i.e., the red curve in Fig 3)? Do you achieve this by selecting different loss weights...
It is a wonderful work! But I find that in your implementation, you feed every frame to all the backbone (resnet50; resnet43; resnet18), which is different from the proposed method...
