HydrogenBombaklot

Can you please integrate Phi 3 training? It seems to be the odd one out now that Llama 3, Gemma 2, and Mistral are supported.

I would like to use and fine-tune an Llm2VEC model using MLX. LLM2VEC adapts decoder-only LLMs and transforms them into bidirectional encoder-decoder models. https://github.com/McGill-NLP/llm2vec Integrating LLM2VEC support into MLX would...

I would like to request support for LLM2VEC, which turns Decoder-only LLM's into bidirectional Encoder-Decoder models. https://github.com/McGill-NLP/llm2vec