Update metadata and add paper link
#1
by nielsr HF Staff - opened
This PR updates the model card for EVOKE:
- Changes the pipeline tag from
text-to-videotoimage-to-videoto accurately reflect the model's primary conditioning mode. - Adds
library_name: diffusersto enable appropriate library integration and discoverability. - Adds a direct link to the paper on the Hugging Face Hub in the header.
No other modifications are made; the existing usage instructions, notes, and citation remain as provided by the authors.
Thanks @nielsr ! The paper link was a good catch — it's live on main now,
along with fixes to the arXiv badge, the GitHub URL, and the download
commands.
I'm keeping pipeline_tag: text-to-video though: EVOKE runs t2v / i2v / v2v
and t2v is a first-class mode, so no single tag really fits — both are
already in tags. And I'd rather hold off on declaring
library_name: diffusers, since the repo has no root model_index.json
and needs a diffusers fork that isn't on PyPI, so the generated snippet
wouldn't run.
Closing since main has moved past this branch. Thanks again!
SII-YuanyangYin changed pull request status to closed