MarcusNerva / HMN

[CVPR2022] Official code for Hierarchical Modular Network for Video Captioning. Our proposed HMN is implemented with PyTorch.
MIT License
52 stars 9 forks source link

Some issue to modify the code #6

Closed Chenfeng1271 closed 2 years ago

Chenfeng1271 commented 2 years ago

Hi, thank you for your work. When I implement this code, I meet several problems, mainly from parameter setting. For example, --language_dir_name, data/MSVD/language/vid2fillmask_MSVD.pkl in data file, missing weight_decay in cfg.train.

MarcusNerva commented 2 years ago

Hi, thank you for your work. When I implement this code, I meet several problems, mainly from parameter setting. For example, --language_dir_name, data/MSVD/language/vid2fillmask_MSVD.pkl in data file, missing weight_decay in cfg.train.

For the first problem, you may solve it by renaming the file. For the second one, I have updated my repo to solve it, please check again.