NVlabs / prismer

The implementation of "Prismer: A Vision-Language Model with Multi-Task Experts".
https://shikun.io/projects/prismer
Other
1.3k stars 75 forks source link

Transformers integration and hosting on HuggingFace? #3

Closed altryne closed 1 year ago

altryne commented 1 year ago

Hey, could you guys make this available on hugging face? Not only due to easier pre-trained weights downloading, but also for being able to add this amazing work to coparisons like this one by @nielsr

https://huggingface.co/spaces/nielsr/comparing-captioning-models

and compare visually.

Thank you for releasing this work open source!

lorenmt commented 1 year ago

We are working on that. Stay tuned.

lorenmt commented 1 year ago

Solved in the latest commit.

altryne commented 1 year ago

@lorenmt thanks for the launch! Is it just me or is it been building for the past 2 hours? 🙏

CleanShot 2023-03-12 at 22 47 48@2x
lorenmt commented 1 year ago

Sorry. We were trying further optimize the inference speed. It should be working now.