README.md

metadata

library_name: transformers
license: apache-2.0
language:
  - ja
base_model: abeja/ABEJA-Qwen2.5-32b-Japanese-v0.1
pipeline_tag: text-generation
tags:
  - mlx

mlx-community/ABEJA-Qwen2.5-32b-Japanese-v0.1-4bit

The Model mlx-community/ABEJA-Qwen2.5-32b-Japanese-v0.1-4bit was converted to MLX format from abeja/ABEJA-Qwen2.5-32b-Japanese-v0.1 using mlx-lm version 0.21.1.

Use with mlx

pip install mlx-lm

from mlx_lm import load, generate

model, tokenizer = load("mlx-community/ABEJA-Qwen2.5-32b-Japanese-v0.1-4bit")

prompt = "hello"

if tokenizer.chat_template is not None:
    messages = [{"role": "user", "content": prompt}]
    prompt = tokenizer.apply_chat_template(
        messages, add_generation_prompt=True
    )

response = generate(model, tokenizer, prompt=prompt, verbose=True)