HydrogenBombaklot comments

Repositories
Issues
Comments

Results 4 comments of


                                            HydrogenBombaklot

Implement gemma2 training

Is this being merged?

[Feature Request] MLX_lm: add support for GLM4 family of language models

GLM 4 9B tops the [HHEM Hallucination Leaderboard](https://huggingface.co/spaces/vectara/leaderboard) as the LLM model with the lowest hallucination rate of all models tested. I definitely want to see it supported too.

I tried madlad400, but there is a problem with the output if it is float16

Do you mind uploading your madlad-400 mlx to HF?

I tried madlad400, but there is a problem with the output if it is float16

No, I mean I want to use it and easier if it's on HF.