How to finetune SlowFast pretrained model on custom AVA dataset in PytorchVideo?

[Balakishan Molankula]

I want to perform custom training(finetune) on my dataset(AVA format) using pretrained slowfast model from Pytorchvideo model zoo.

---

slowfast model link: https://dl.fbaipublicfiles.com/pytorchvideo/model_zoo/ava/SLOWFAST_8x8_R50_DETECTION.pyth

---

I am unable to find anything related to this in documentation. Could someone help me on this?

[Hammer-888]

i also need this guide to help me

[C-popKingJames]

我也想知道如何用pytorchvideo训练自己的数据集

[kalyanvasudev]

Hi We don't have plans to support this Pytorchvideo anytime soon. I'd encourage users to the AVA benchmarks in pyslowfast repo to do this - https://github.com/facebookresearch/slowfast