DeepSeek · First MLX build 20 Jan 2025
DeepSeek-R1-Distill-Qwen-7B
MLX
Text mit
A 7.6B parameter text generation model from DeepSeek, with 9 MLX builds, from 3-bit to bf16.
- Parameters
- 7.6B From Hugging Face
- Quantisation
- 4-bit DeepSeek-R1-Distill-Qwen-7B-4bit
- Memory needed
- 7 GB Estimate, unified memory
- Download
- 4.3 GB Weights, estimated
- Downloads
- 2.7K This build, last 30 days
- Task
- Text generation Text
- MLX builds
- 9 On mlx-community
- Devices
- Mac Apple silicon
Memory
Will it fit?Needs about 7 GB at 4-bit
8 GB
16 GB
24 GB
32 GB
64 GB
128 GB
Fits with headroomTight, close other appsWon’t fit
Estimated from the size of the weights, plus a fifth for context and working memory and 2 GB for macOS and your apps. Long contexts need more.
Run it
Terminal · Hugging Face CLI$ pip install -U huggingface_hub
$ hf download mlx-community/DeepSeek-R1-Distill-Qwen-7B-4bitMLX builds
9 on mlx-community| DeepSeek-R1-Distill-Qwen-7B-3bit | 3-bit | 3.3 GB | 6 GB | 63 |
|---|---|---|---|---|
| DeepSeek-R1-Distill-Qwen-7B-4bit | 4-bit | 4.3 GB | 7 GB | 2.7K |
| DeepSeek-R1-Distill-Qwen-7B-4bit-mlx | 4-bit | 4.3 GB | 7 GB | 88 |
| DeepSeek-R1-Distill-Qwen-7B-6bit | 6-bit | 6.2 GB | 9 GB | 86 |
| DeepSeek-R1-Distill-Qwen-7B-8bit | 8-bit | 8.1 GB | 12 GB | 733 |
| DeepSeek-R1-Distill-Qwen-7B-chat-mlx-16bit | 16-bit | 15.2 GB | 20 GB | 792 |
| DeepSeek-R1-Distill-Qwen-7B | Unstated | 15.2 GB | 20 GB | 636 |
| DeepSeek-R1-Distill-Qwen-7B-bf16 | bf16 | 15.2 GB | 20 GB | 67 |
| DeepSeek-R1-Distill-Qwen-7B-MLX | Unstated | — | — | 0 |
Sizes are estimated from Hugging Face’s parameter counts and the quantisation. Downloads are for the last 30 days.