Files

David Koski bb7bacc077 fix for #2 -- CodeLlama crashes

- add replacement tokenizer class for unknown tokenizers
- fix quantization for models that don't have lm_head quantized

Requires https://github.com/ml-explore/mlx-swift/pull/28

2024-02-26 10:38:05 -08:00

Configuration.swift

initial commit

2024-02-22 10:41:02 -08:00

Gemma.swift

fix for #2 -- CodeLlama crashes

2024-02-26 10:38:05 -08:00

Llama.swift

fix for #2 -- CodeLlama crashes

2024-02-26 10:38:05 -08:00

LLM.h

initial commit

2024-02-22 10:41:02 -08:00

LLMModel.swift

fix for #2 -- CodeLlama crashes

2024-02-26 10:38:05 -08:00

Phi.swift

fix for #2 -- CodeLlama crashes

2024-02-26 10:38:05 -08:00

README.md

fix broken links, clarify documentation

2024-02-22 12:46:44 -08:00

Util.swift

fix for #2 -- CodeLlama crashes

2024-02-26 10:38:05 -08:00

README.md

Llama

This is a port of several models from:

https://github.com/ml-explore/mlx-examples/blob/main/llms/mlx_lm/models/

You can use this to load models from huggingface, e.g.:

https://huggingface.co/mlx-community/Mistral-7B-v0.1-hf-4bit-mlx

Currently supported model types are:

Llama / Mistral
Gemma
Phi

See Configuration.swift for more info.

See llm-tool