Needs updating to support Qwen3_moe models

#59
by insmek - opened

Trying to convert "Qwen/Qwen3-235B-A22B-Instruct-2507" and unfortunately it keeps erroring out. It seems like newer versions of mlx-lm>=0.24.0 support conversion of Qwen3_moe so I suppose that's the one to aim for.

Thanks for the great tool! It's a huge boon for Apple Silicon users.

Sign up or log in to comment