Modelpublic

TyKaoz/bge-m3-6bit

sourceHugging Faceapache-2.0updated 4mo agoView on Hugging Face

0likes51downloads

Model Card

BGE-M3 — 6-bit MLX

6-bit MLX quantization of `BAAI/bge-m3`, for Apple Silicon (~457 MB). Multilingual text-embedding model (1024-dim dense vectors).

Usage

bash

pip install -U mlx-embeddings

python

from mlx_embeddings import load, generate

model, tokenizer = load("TyKaoz/bge-m3-6bit")
out = generate(model, tokenizer, texts=["Bonjour", "Hello", "你好"])
print(out.text_embeds.shape)  # (3, 1024)

Base	Tool	Precision	Size
`BAAI/bge-m3`	`mlx-embeddings`	6-bit · group 64	~457 MB

By [TyKaoz](https://www.tykaoz.bzh) — privacy-first native macOS LLM chat client. Apache 2.0, inherited from the base model.