Alibaba Qwen · First MLX build 6 Jun 2024
Qwen2-72B
MLX
tongyi-qianwen
A 20B parameter model from Alibaba Qwen, with 3 MLX builds, from 2-bit to 8-bit.
- Parameters
- 20B From Hugging Face
- Quantisation
- 8-bit dolphin-2.9.2-qwen2-72b-8bit
- Memory needed
- 29 GB Estimate, unified memory
- Download
- 23.9 GB Weights, estimated
- Downloads
- 48 This build, last 30 days
- Task
- —
- MLX builds
- 3 On mlx-community
- Devices
- Mac Apple silicon
Memory
Will it fit?Needs about 29 GB at 8-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/dolphin-2.9.2-qwen2-72b-8bitMLX builds
3 on mlx-community| dolphin-2.9.2-qwen2-72b-2bit | 2-bit | 6.0 GB | 9 GB | 29 |
|---|---|---|---|---|
| dolphin-2.9.2-qwen2-72b-4bit | 4-bit | 9.7 GB | 13 GB | 35 |
| dolphin-2.9.2-qwen2-72b-8bit | 8-bit | 23.9 GB | 29 GB | 48 |
Sizes are estimated from Hugging Face’s parameter counts and the quantisation. Downloads are for the last 30 days.