PantiPenti's picture
Upload README.md with huggingface_hub
e3ae199 verified
metadata
base_model: nicoboss/DeepSeek-R1-Distill-Qwen-32B-Uncensored
library_name: peft
tags:
  - generated_from_trainer
  - mlx
  - mlx-my-repo
license: mit
datasets:
  - Guilherme34/uncensor
model-index:
  - name: DeepSeek-R1-Distill-Qwen-32B-Uncensored
    results: []

PantiPenti/DeepSeek-R1-Distill-Qwen-32B-Uncensored-mlx-6Bit

The Model PantiPenti/DeepSeek-R1-Distill-Qwen-32B-Uncensored-mlx-6Bit was converted to MLX format from nicoboss/DeepSeek-R1-Distill-Qwen-32B-Uncensored using mlx-lm version 0.22.3.

Use with mlx

pip install mlx-lm
from mlx_lm import load, generate

model, tokenizer = load("PantiPenti/DeepSeek-R1-Distill-Qwen-32B-Uncensored-mlx-6Bit")

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)