Prince Canuma
Prince Canuma
I opened a new issue for the pauses so can keep track and return to it.
Hey @JoeJoe1313 You are correct! ✅ The multimodal should be converted using mlx-vlm and not mlx-lm Please feel free to re-upload using mlx-vlm, if you can't just ping me and...
This is happening because new models are migrating to jinja files instead of json for chat template. I fixed it here #376. You can install from source for now it...
This is awesome! I have some ideas, let's collaborate on a PR
To solve #136 We can load KV cache ([attention sink](https://arxiv.org/abs/2309.17453)) and compute values for new image and prompt tokens. Note: 1. Save each image hash in the cached prompts/features for...
@Licyldj this should be fixed in v0.2.2
Hey @fakerybakery Thanks for reaching out and suggesting it! Please feel free to suggest new models anytime.❤️ It's definitely on the roadmap 🚀 Will start on the tts module initially...
Thanks! I will look into it. There is another issue reporting the same problem.
I wonder that as well. But being closer to the project helps. If it grows in popularity then we can consider a separate repo. cc: @lucasnewman