Add new text-to-image LocalApp JoyFusion(macOS)

huggingface / huggingface.js

Utilities to use the Hugging Face Hub API

https://hf.co/docs/huggingface.js

MIT License

1.37k stars 213 forks source link

Add new text-to-image LocalApp JoyFusion(macOS) #740

Closed jiangdi0924 closed 1 month ago

jiangdi0924 commented 4 months ago

JoyFusion is a native application for macOS , built upon Stable Diffusion and CoreML technologies. It can locally run CoreML-based text-to-image models.

Currently, only CoreML model ZIP file imports are supported, but more file types will be supported in the future. For now, you can find CoreML files in the following repos or community: https://huggingface.co/Norton0924/Joyfusion/tree/main https://huggingface.co/coreml-community

👇Here is App Icon SVG joyfusion (1)

👇Demo Video

https://github.com/huggingface/huggingface.js/assets/21988177/2db2c186-a0dc-4707-8ce5-de1e74071edb

julien-c commented 1 month ago

Hi @jiangdi0924, any way to support all coreml models (no zip file needed), or alternatively, do you have a tag on model repos you support? i.e. repos could specify a joyfusion tag in their model card to be deemed as compatible with JoyFusion?

Thanks!

jiangdi0924 commented 1 month ago

Hi @jiangdi0924, any way to support all coreml models (no zip file needed), or alternatively, do you have a tag on model repos you support? i.e. repos could specify a joyfusion tag in their model card to be deemed as compatible with JoyFusion?

Thanks!

Hello Julien-C,

Thank you for your feedback. Indeed, at the moment, we only support the zip file format for coreML models. Currently, there are a significant number of coreML model zips available on Hugging Face. We plan to support CoreML Repos with non-zip file types in the next version.

This PR has been pending for a while, and if possible, we would like to merge it and release a version. After the merge, we will continue to optimize the file types for importing models and the user experience.

Additionally, regarding your suggestion for the JoyFusion tag, there are indeed some limitations at present. Once again, thank you for your feedback.

pcuenca commented 1 month ago

works well

Tested here, works fine for me too.

enzostvs commented 1 month ago

Hi @jiangdi0924 Could you run the linter on your PR to format it properly please ? You can install the prettier extension on VSCode, should be automatic after a save

Thank you 🙏

jiangdi0924 commented 1 month ago

Hi @jiangdi0924 Could you run the linter on your PR to format it properly please ? You can install the prettier extension on VSCode, should be automatic after a save

Thank you 🙏 👋 @enzostvs , sorry about the formatting issues in the previous code. I’ve tried using Prettier to it and made another commit (Tag: 066a1722f55f2b71868abe6e9bcb124b67e1cbd9). If there are any problems, please let me know anytime.

enzostvs commented 1 month ago

Test passed, PR has been merged! Thank you @jiangdi0924 for your contribution 🚀

julien-c commented 1 month ago

made a small tweak (for now) in #847 requiring an explicit joyfusion tag on the model

enzostvs commented 1 month ago

Hey @jiangdi0924 JoyFusion has been added to HF local apps! You can see it on my screen:

You can now communicate on it if you need to !

Thanks again for your contribution!

jiangdi0924 commented 1 month ago

Hey @jiangdi0924 JoyFusion has been added to HF local apps! You can see it on my screen:

You can now communicate on it if you need to !

Thanks again for your contribution!

Hi @enzostvs , Thanks for the support! It’s awesome to see JoyFusion on huggingface’s Local Apps. Looking forward to our next collaboration! 🫡