Replies: 5 comments
|
They've completely lost the initiative in AI. I feel like they don't actually care about it at all—it's just a gimmick so users don't completely abandon them. |
|
just switch entirely to MLX-VLM the developers have a passion for the work. It can handle many text only models as well. |
|
Sh*t! I just realised that Awni (who own this repo and co-created MLX) left Apple for Anthropic. His GitHub account is mostly inactive since the last MLX release, five months ago. 🥹 |
|
Looks like @zcbenz joined as a new maintainer (hi Cheng, congrats). |
|
Sorry for the long wait, an inference engine definitely needs much more than 1~2 maintainers to work on especially with exponentially increase of new model architectures and decoding optimizations, and we had made Awni overworked on this. We are currently building a new team for mlx-lm but it is going to take a long time for us to find the right people, hire them, and get them onboarded. In the meantime we will do our best to improve things in mlx-lm while working on mlx, but still development is going to be slow for a long while, and I don't want to see anyone in the team getting burnt out again. |
Uh oh!
There was an error while loading. Please reload this page.
i might have missed an announcement or a thread, but I don't get it why no new releases are being cut? is the project dead, despite the commits?
All reactions