MLX Community · First MLX build 22 Dec 2024
deepseek-vl2-tiny
MLX
TextVision deepseek
A 3.4B parameter vision and text model, with 5 MLX builds, from 3-bit to bf16.
- Parameters
- 3.4B From Hugging Face
- Quantisation
- 4-bit deepseek-vl2-tiny-4bit
- Memory needed
- 5 GB Estimate, unified memory
- Download
- 2.1 GB Weights, estimated
- Downloads
- 170 This build, last 30 days
- Task
- Vision and text Text, Vision
- MLX builds
- 5 On mlx-community
- Devices
- Mac Apple silicon
Memory
Will it fit?Needs about 5 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 · mlx-vlm$ pip install -U mlx-vlm
$ python -m mlx_vlm.generate --model mlx-community/deepseek-vl2-tiny-4bit --image photo.jpg --prompt "Describe this image"MLX builds
5 on mlx-community| deepseek-vl2-tiny-3bit | 3-bit | 1.7 GB | 4 GB | 70 |
|---|---|---|---|---|
| deepseek-vl2-tiny-4bit | 4-bit | 2.1 GB | 5 GB | 170 |
| deepseek-vl2-tiny-6bit | 6-bit | 2.9 GB | 6 GB | 93 |
| deepseek-vl2-tiny-8bit | 8-bit | 3.7 GB | 7 GB | 93 |
| deepseek-vl2-tiny-bf16 | bf16 | 6.7 GB | 10 GB | 81 |
Sizes are estimated from Hugging Face’s parameter counts and the quantisation. Downloads are for the last 30 days.