Views
No views yet
gpt_oss (Mixture of Experts — 117B total, 5.1B active)1pip install mlx-lm
2
3mlx_lm.generate \
4 --model tocchitocchi/GPT-OSS-Swallow-120B-RL-v0.1-6bit-mlx \
5 --prompt "日本の首都はどこですか？" \
6 --max-tokens 2001from mlx_lm import load, generate
2
3model, tokenizer = load("tocchitocchi/GPT-OSS-Swallow-120B-RL-v0.1-6bit-mlx")
4
5messages = [{"role": "user", "content": "Pythonでフィボナッチ数列を出力するコードを書いてください"}]
6prompt = tokenizer.apply_chat_template(messages, add_generation_prompt=True)
7
8response = generate(model, tokenizer, prompt=prompt, verbose=True, max_tokens=500)