Added "tie_word_embeddings": false this is needed to return LM_HEAD layers in MLX

aaabiao changed pull request status to merged

Sign up or log in to comment