search

xinghaochen / SqueezeTime

Official PyTorch implementation of "No Time to Waste: Squeeze Time into Channel for Mobile Video Understanding"
Apache License 2.0
30 stars 1 forks source link

Regarding the annotation of Ava datasets #6

Open steven-spec opened 1 week ago

steven-spec commented 1 week ago

Thank you for your outstanding work i want to train the mode with ava dataset ,but I don't know if the annotation is the same as slowfast's one second frame annotation, or if all 30 frames are annotated in one second Look forward to your reply

zyjwuyan commented 6 days ago

We prepared the ava2.1 dataset following the default instruction of 'https://github.com/open-mmlab/mmaction2/blob/main/tools/data/ava/README.md', please follow this instruction to prepare your data.