Hugging Face · First MLX build 26 Nov 2024
SmolVLM-Instruct
MLX
TextVision apache-2.0
A 2.2B parameter vision and text model from Hugging Face, with 4 MLX builds, from 4-bit to bf16.
- Parameters
- 2.2B From Hugging Face
- Quantisation
- 4-bit SmolVLM-Instruct-4bit
- Memory needed
- 4 GB Estimate, unified memory
- Download
- 1.5 GB Weights, estimated
- Downloads
- 703 This build, last 30 days
- Task
- Vision and text Text, Vision
- MLX builds
- 4 On mlx-community
- Devices
- Mac Apple silicon
Memory
Will it fit?Needs about 4 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/SmolVLM-Instruct-4bit --image photo.jpg --prompt "Describe this image"MLX builds
4 on mlx-community| SmolVLM-Instruct-4bit | 4-bit | 1.5 GB | 4 GB | 703 |
|---|---|---|---|---|
| SmolVLM-Instruct-6bit | 6-bit | 2.0 GB | 5 GB | 47 |
| SmolVLM-Instruct-8bit | 8-bit | 2.5 GB | 5 GB | 62 |
| SmolVLM-Instruct-bf16 | bf16 | 4.5 GB | 8 GB | 74 |
Sizes are estimated from Hugging Face’s parameter counts and the quantisation. Downloads are for the last 30 days.