Checkpoint of the pre-trained dual-encoder.

facebookresearch / LaViLa

Code release for "Learning Video Representations from Large Language Models"

MIT License

491 stars 46 forks source link

Checkpoint of the pre-trained dual-encoder. #36

Open AlbertHuyb opened 6 months ago

AlbertHuyb commented 6 months ago

Hi @zhaoyue-zephyrus,

I'm really interested in your brilliant work. Could you please provide the dual encoder checkpoint at the very first stage before narrator training? I'd like to evaluate the performance of that checkpoint pre-trained on Ego4D annotations.

Thanks!

zhaoyue-zephyrus commented 6 months ago

Hi @AlbertHuyb

Please take a read at this doc: https://github.com/facebookresearch/LaViLa/blob/main/docs/PRETRAIN.md#pre-training-dual-encoder-baseline.

Best, Yue