MLC Chat

missing a few big models still like llama 3 (as of right now at least), otherwise pretty good

Dommage que cette appli ne soit pas adaptée à l’iPhone SE3 (pas assez de VRam)et dommage que le développeur ne l’indique pas explicitement en description. J’espère que cela fonctionnera sur le prochain iPhone SE.

I love how fast the LLMs in this app are. Is this because of the chosen quantization? I usually use q8 on my desktop computer, and it take minutes to generate a text. Anyway. I wish it was possible to run Dolphin Mixtral (because I heavily dislike the alignment of regular Mistral), and LLaVA (because local image recognition directly from a smartphone would be amazing). Keep up the good work!

Download button dissapeared

Not enough VRAM on iPhone 13 mini ? So I can’t do anything in the app :(

The download model button was disabled in the new version

Sorry, the system cannot provide 3814.7MB VRAM as requested to the app, so we cannot initialize this model on this device.

The app is great but can you add shortcuts support?

Tried many times, the conversations cannot be saved. It’s really a nuisance! What’s the point if the conversation isn’t saved for later use? Disappointed!

The app works well, but the instructions on how to upload other models is unclear, leaving me to only use Mistral and nothing else.