soltkreig's picture
Add foundation and fine-tuned SLR model weights
2ea22a3 verified
|
Raw
History Blame Contribute Delete
750 Bytes
---
license: cc-by-sa-4.0
tags:
- sign-language
- video
- mae
- pytorch
---
# Multilingual Sign Language Recognition Model
Foundation and fine-tuned weights for video-based sign language recognition
([paper / code](https://github.com/ds-hub-sochi/multilingual-SLR-model)).
## Files
| Model | File |
|---|---|
| Foundation model | [`pretrain.pth`](pretrain.pth) |
| WLASL fine-tuned | [`wlasl.pt`](wlasl.pt) |
| AUTSL fine-tuned | [`autsl.pt`](autsl.pt) |
| GSL fine-tuned | [`gsl.pt`](gsl.pt) |
| SLOVO fine-tuned | [`slovo.pt`](slovo.pt) |
Smoothing arrays used by `mae_smoothing` are in [`smoothing/`](smoothing/).
## Citation
See the [GitHub repository](https://github.com/ds-hub-sochi/multilingual-SLR-model) for citation details.