Spaces:
Sleeping
Sleeping
|
Download models/README.md from ofraij123/BridgeLinkASL: direct link, hf CLI and curl.
- Browser
- Download file 982 Bytes
-
https://huggingface.co/spaces/ofraij123/BridgeLinkASL/resolve/main/models/README.md
- Command line
-
hf download hf://spaces/ofraij123/BridgeLinkASL/models/README.md
-
curl -L -o README.md https://huggingface.co/spaces/ofraij123/BridgeLinkASL/resolve/main/models/README.md
982 Bytes
A newer version of the Gradio SDK is available: 6.32.0
Model Weights
Place the trained model files here after running the Colab training notebook.
Required files
| File | Source | Description |
|---|---|---|
sign_transformer_best.pt |
Google Drive → BridgeLink-ASL → models/ | Trained Transformer weights + label map + config (~5 MB) |
labels.json |
Google Drive → BridgeLink-ASL → models/ | Label index → gloss name mapping |
How to get them
- Open Google Drive →
BridgeLink-ASL/models/ - Download
sign_transformer_best.ptandlabels.json - Place both files in this
models/directory
The app loads sign_transformer_best.pt at startup. If the file is missing,
the app will start but show a "Model: NOT LOADED" banner and predictions
won't work.
Alternative: Hugging Face Hub
Instead of committing the .pt file to git, you can upload it to a HF model
repo and set HF_MODEL_REPO=your-username/repo-name as an environment
variable. The app will auto-download from the Hub at startup.