| --- |
| license: apache-2.0 |
| base_model: Wan-AI/Wan2.2-T2V-A14B-Diffusers |
| tags: |
| - controlnet |
| - text-to-video |
| - optical-illusions |
| --- |
| |
| # Hidden in Plain Sight — ControlNet |
|
|
| Trained ControlNet for the *Hidden in Plain Sight* thesis pipeline. Cold-started |
| from the HED architecture of `TheDenk/wan2.2-t2v-a14b-controlnet-hed-v1` and |
| trained on top of the high-noise expert of Wan 2.2 T2V-A14B with a spatial |
| face-weighted flow-matching loss. |
|
|
| ## Use |
|
|
| ```bash |
| huggingface-cli download <YOUR_HF_USERNAME>/hidden-in-plain-sight \ |
| controlnet.safetensors --local-dir models/controlnet |
| ``` |
|
|
| See [the project repo](https://github.com/<YOU>/Video_Anagrams) for the full pipeline. |
|
|
