---
title: DeepSeek-R1-Distill-Qwen-1.5B-6.25bit-G128
canonical_url: "https://www.modelscope.cn/models/mlx-community/DeepSeek-R1-Distill-Qwen-1.5B-6.25bit-G128"
md_url: "https://www.modelscope.cn/models/mlx-community/DeepSeek-R1-Distill-Qwen-1.5B-6.25bit-G128.md"
repository: mlx-community/DeepSeek-R1-Distill-Qwen-1.5B-6.25bit-G128
last_updated: 2025-02-17
license: mit
pipeline_tag: text-generation
tasks:
  - text-generation
model_type:
  - qwen2
architectures:
  - Qwen2ForCausalLM
base_model:
  - deepseek-ai/DeepSeek-R1-Distill-Qwen-1.5B
base_model_relation: quantized
parameters: 361.1M
tensor_type:
  - U32
  - BF16
library_name:
  - mlx
  - safetensors
  - pytorch
frameworks:
  - pytorch
inference_backends:
  - "deploy_task text/emb"
  - "lmdeploy 0.9.1"
  - "lmdeploy_turbomind 0.9.1"
  - "sglang 0.5.2"
  - "vllm 0.9.2"
downloads: 19
stars: 0
tags:
  - mlx
---

# DeepSeek-R1-Distill-Qwen-1.5B-6.25bit-G128

> DeepSeek-R1-Distill-Qwen-1.5B-6.25bit-G128 - mlx-community 在 ModelScope 开源的模型。bobig/DeepSeek-R1-Distill-Qwen-1.5B-6bit

mlx-community/DeepSeek-R1-Distill-Qwen-1.5B-6.25bit-G128 是 ModelScope 魔搭社区上的 361.1M 参数text-generation模型，采用 mit 许可，基于 deepseek-ai/DeepSeek-R1-Distill-Qwen-1.5B 构建，可用 deploy_task text/emb、lmdeploy 0.9.1、lmdeploy_turbomind 0.9.1 部署。

- **Repository**: mlx-community/DeepSeek-R1-Distill-Qwen-1.5B-6.25bit-G128
- **License**: mit
- **Tasks**: text-generation
- **Parameters**: 361.1M
- **Base model**: deepseek-ai/DeepSeek-R1-Distill-Qwen-1.5B
- **Inference backends**: deploy_task text/emb, lmdeploy 0.9.1, lmdeploy_turbomind 0.9.1, sglang 0.5.2, vllm 0.9.2
- **Tags**: mlx
- **Downloads**: 19
- **Stars**: 0
- **Last updated**: 2025-02-17

Source: https://www.modelscope.cn/models/mlx-community/DeepSeek-R1-Distill-Qwen-1.5B-6.25bit-G128

---

# bobig/DeepSeek-R1-Distill-Qwen-1.5B-6bit

This works well as a draft model for speculative decoding in LMstudio 3.10 with:

mlx-community/FuseO1-DeepSeekR1-Qwen2.5-Coder-32B-4.25bit-G128


The Model [bobig/DeepSeek-R1-Distill-Qwen-1.5B-6bit](https://huggingface.co/bobig/DeepSeek-R1-Distill-Qwen-1.5B-6bit) was
converted to MLX format from [deepseek-ai/DeepSeek-R1-Distill-Qwen-1.5B](https://huggingface.co/deepseek-ai/DeepSeek-R1-Distill-Qwen-1.5B)
using mlx-lm version **0.21.4**.  Group Size 128, bfloat16
## Use with mlx

```bash
pip install mlx-lm
```

```python
from mlx_lm import load, generate

model, tokenizer = load("bobig/DeepSeek-R1-Distill-Qwen-1.5B-6bit")

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)
```