search

snap-research / Panda-70M

[CVPR 2024] Panda-70M: Captioning 70M Videos with Multiple Cross-Modality Teachers
https://snap-research.github.io/Panda-70M/
438 stars 15 forks source link

Could you please open you retrieval data of the 100K videos? #20

Open LukeALee opened 3 months ago

LukeALee commented 3 months ago

As mentioned in your paper, no single caption model excels in all types of videos. This presents a significant challenge for researchers, making it crucial to have a robust baseline dataset for training retrieval models. I'm curious if you would consider making this data available to the research community？Thanks a lot !

tsaishien-chen commented 3 months ago

Hi @LukeALee, thanks for your interest in our fine-grained retrieval dataset! To release this data, we need additional approval from our legal team. I'll let you know asap when we are approved to release them.