File size: 1,261 Bytes
e7b4310 2677e94 e7b4310 2677e94 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 |
---
license: apache-2.0
tags:
- mlx
base_model: GreenBitAI/DeepSeek-R1-Distill-Qwen-1.5B-layer-mix-bpw-4.0
---
# GreenBitAI/DeepSeek-R1-Distill-Qwen-1.5B-layer-mix-bpw-4.0-mlx
This quantized low-bit model [GreenBitAI/DeepSeek-R1-Distill-Qwen-1.5B-layer-mix-bpw-4.0-mlx](https://huggingface.co/GreenBitAI/DeepSeek-R1-Distill-Qwen-1.5B-layer-mix-bpw-4.0-mlx) was converted to MLX format from [`GreenBitAI/DeepSeek-R1-Distill-Qwen-1.5B-layer-mix-bpw-4.0`](https://huggingface.co/GreenBitAI/DeepSeek-R1-Distill-Qwen-1.5B-layer-mix-bpw-4.0) using gbx-lm version **0.3.5**.
Refer to the [original model card](https://huggingface.co/GreenBitAI/DeepSeek-R1-Distill-Qwen-1.5B-layer-mix-bpw-4.0) for more details on the model.
## Use with mlx
```bash
pip install gbx-lm
```
```python
from gbx_lm import load, generate
model, tokenizer = load("GreenBitAI/DeepSeek-R1-Distill-Qwen-1.5B-layer-mix-bpw-4.0-mlx")
prompt = "How can I make an apple cake"
if tokenizer.chat_template is not None:
messages = [{"role": "user", "content": prompt}]
prompt = tokenizer.apply_chat_template(
messages, add_generation_prompt=True
)
prompt = tokenizer.decode(prompt)
response = generate(model, tokenizer, prompt=prompt, verbose=True, max_tokens=4096)
```
|