YAML Metadata Warning:empty or missing yaml metadata in repo card
Check out the documentation for more information.
SemWAM frozen assets
Content-addressed mirror of the frozen weights SemWAM training and evaluation load.
Use bash setup.sh from the SemWAM repository; it pins this repository to a commit
and verifies every file by sha256. Each asset keeps its upstream license (see
asset_manifest.json); DINOv3 materials are redistributed under dinov3/LICENSE.md.
| asset | files | license |
|---|---|---|
| wan22_ti2v5b_dit | 3 | Apache-2.0 |
| wan22_vae | 1 | Apache-2.0 |
| umt5_xxl_encoder | 1 | Apache-2.0 |
| umt5_xxl_tokenizer | 4 | Apache-2.0 |
| action_dit_backbone | 1 | Apache-2.0 (derived from Wan2.2-TI2V-5B) |
| dinov3_vitb16 | 1 | DINOv3 License (redistributed with third_party/dinov3/LICENSE.md) |
| svae_robotwin_v7 | 3 | see release README |
| svae_libero_v2 | 3 | see release README |
| prism_depth_dvae | 1 | see THUSI-Lab/PrismWAM |
| vda_metric_vitb | 1 | see depth-anything/Metric-Video-Depth-Anything-Base |
| defm_vit_s14 | 1 | see leggedrobotics/defm |
| vjepa2_1_vitl | 1 | see facebookresearch/vjepa2 |
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support