README.md

metadata

base_model: macadeliccc/Samantha-Qwen-2-7B
datasets:
  - macadeliccc/opus_samantha
  - HuggingfaceH4/ultrachat_200k
  - teknium/OpenHermes-2.5
  - Sao10K/Claude-3-Opus-Instruct-15K
language:
  - en
  - zh
license: apache-2.0
tags:
  - mlx

mlx-community/Samantha-Qwen-2-7B-4bit

The Model mlx-community/Samantha-Qwen-2-7B-4bit was converted to MLX format from macadeliccc/Samantha-Qwen-2-7B using mlx-lm version 0.19.0.

Use with mlx

pip install mlx-lm

from mlx_lm import load, generate

model, tokenizer = load("mlx-community/Samantha-Qwen-2-7B-4bit")

prompt="hello"

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

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