Skip to content

PyTorch implementation of “Brain-inspired learning to Deeper Inductive Reasoning for Video Captioning“ (International Journal of Machine Learning and Cybernetics, 2023.05)

Notifications You must be signed in to change notification settings

Canyizl/DIR-for-Video-Captioning

Repository files navigation

Brain-inspired learning to Deeper Inductive Reasoning for Video Captioning (https://doi.org/10.1007/s13042-023-01876-9)

This repo is accepted by "International Journal of Machine Learning and Cybernetics" (IF=5.6 2023.06)

Some files you may need(ckpt,json,datasets...): https://pan.baidu.com/s/1ZraMbY8eEFKNHBgkz_RiUA(rxlq)

The "caption_eval" is also recommended to download in any other repo, cuz it noticed me some files is larger than 50MB and I am not sure it is uploaded successfully.

Running:

The "visible.py" can visualize some video samples, so you can run first.

Please notice that we choose the videos in "MSR-VTT". So you should download the "./dataset/MSR-VTT/*", "dict2.pt" and "warm_up.pt".

Just run the visible.py and see what happened !

The code is some kind of mass cuz I tried many times in this experiment.

You can train your own model in MSVD and MSR-VTT, and we do not offer any pretrained models becuase of the mass version control...

Citation

If you find our code or our paper useful for your research, please cite our work:

Coming soon.

Contact

If you have any question or comment, please contact 1962410334@hhu.edu.cn.

It also should be noticed that I will studied in ZJU soon. So you can contact my personal email 975136225@qq.com.

About

PyTorch implementation of “Brain-inspired learning to Deeper Inductive Reasoning for Video Captioning“ (International Journal of Machine Learning and Cybernetics, 2023.05)

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published