search

layer6ai-labs / xpool

https://layer6ai-labs.github.io/xpool/
110 stars 9 forks source link

Question about the validation #4

Closed bighuang624 closed 2 years ago

bighuang624 commented 2 years ago

Hi, much thanks for your nice paper and the well-organized open-resourced codebase. I am a beginner of the video-text retrieval area, and I noticed that you seemed to have used the test split of the datasets for validation, according to the Line.40 of train.py. I want to know if I misunderstood this, or is this the common practice in this area? Looking forward to any response.

NoelVouitsis commented 2 years ago

Hi there, thank you for taking interest in our work! This is a convention following previous work such as CLIP4Clip [1]. Thanks!

[1] Huaishao Luo, Lei Ji, Ming Zhong, Yang Chen, Wen Lei, Nan Duan, and Tianrui Li. Clip4clip: An empirical study of clip for end to end video clip retrieval. arXiv preprint arXiv:2104.08860, 2021.