README.md

metadata

license: llama3.1
language:
  - en
inference: false
fine-tuning: false
tags:
  - nvidia
  - llama3.1
  - mlx
datasets:
  - nvidia/HelpSteer2
base_model: nvidia/Llama-3.1-Nemotron-70B-Instruct-HF
pipeline_tag: text-generation
library_name: transformers

win28703/Llama-3.1-Nemotron-70B-Instruct-HF-Q8-mlx

The Model win28703/Llama-3.1-Nemotron-70B-Instruct-HF-Q8-mlx was converted to MLX format from nvidia/Llama-3.1-Nemotron-70B-Instruct-HF using mlx-lm version 0.19.1.

Use with mlx

pip install mlx-lm

from mlx_lm import load, generate

model, tokenizer = load("win28703/Llama-3.1-Nemotron-70B-Instruct-HF-Q8-mlx")

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)