MLX Community · First MLX build 6 Oct 2024
llava-interleave-qwen-7b
MLX
TextVision tongyi-qianwen-research
A 8.1B parameter vision and text model, with 4 MLX builds, from 4-bit to 8-bit.
- Parameters
- 8.1B From Hugging Face
- Quantisation
- 4-bit llava-interleave-qwen-7b-4bit
- Memory needed
- 8 GB Estimate, unified memory
- Download
- 5.2 GB Weights, estimated
- Downloads
- 30 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 8 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/llava-interleave-qwen-7b-4bit --image photo.jpg --prompt "Describe this image"MLX builds
4 on mlx-community| llava-interleave-qwen-7b-4bit | 4-bit | 5.2 GB | 8 GB | 30 |
|---|---|---|---|---|
| llava-interleave-qwen-7b-bf16 | bf16 | 16.3 GB | 21 GB | 24 |
| llava-interleave-qwen-7b-6bit | 6-bit | — | — | 30 |
| llava-interleave-qwen-7b-8bit | 8-bit | — | — | 23 |
Sizes are estimated from Hugging Face’s parameter counts and the quantisation. Downloads are for the last 30 days.