search

OpenGVLab / UniFormerV2

[ICCV2023] UniFormerV2: Spatiotemporal Learning by Arming Image ViTs with Video UniFormer
https://arxiv.org/abs/2211.09552
Apache License 2.0
291 stars 18 forks source link

wether the pre-training dataset K710 is available ? #36

Closed peiliu0408 closed 1 year ago

peiliu0408 commented 1 year ago

just as paper presented, the pre-training on K710, i.e., the combination of K400, 600, 700, results shorter training cost on downstreaming task, I wonder whether this annotation file is available?

Andy1621 commented 1 year ago

Hi! Please Check DATASET.md. The annotation file can found here.