showlab / UniVTG

[ICCV2023] UniVTG: Towards Unified Video-Language Temporal Grounding

Home Page:https://arxiv.org/abs/2307.16715

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

What is the version of slowfast(R50)?

hellojet opened this issue · comments

I tried to extract video features through slowfast 8x8 R50 on the tvsum data set. There are some differences from the features downloaded in the feature file you gave.

given: XzYM3PfTM4w.npz
image

extract by slowfast 8x8 R50
image

This ended up not working well when I went to test

I am using this version now:

image

I know the reason. I accidentally switched the positions of clip and slowfast during concat, which resulted in poor results.

Hi @hellojet , have you found the reason? can you please use this codebase for feature extraction. https://github.com/linjieli222/HERO_Video_Feature_Extractor for easy reproduction.