MLX Community · First MLX build 27 Nov 2024
Idefics3-8B-Llama3
MLX
TextVision apache-2.0
A 8.5B parameter vision and text model, with 5 MLX builds, from 3-bit to bf16.
- Parameters
- 8.5B From Hugging Face
- Quantisation
- 4-bit Idefics3-8B-Llama3-4bit
- Memory needed
- 8 GB Estimate, unified memory
- Download
- 5.4 GB Weights, estimated
- Downloads
- 42 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 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/Idefics3-8B-Llama3-4bit --image photo.jpg --prompt "Describe this image"MLX builds
5 on mlx-community| Idefics3-8B-Llama3-3bit | 3-bit | 4.3 GB | 7 GB | 28 |
|---|---|---|---|---|
| Idefics3-8B-Llama3-4bit | 4-bit | 5.4 GB | 8 GB | 42 |
| Idefics3-8B-Llama3-6bit | 6-bit | 7.4 GB | 11 GB | 40 |
| Idefics3-8B-Llama3-8bit | 8-bit | 9.4 GB | 13 GB | 35 |
| Idefics3-8B-Llama3-bf16 | bf16 | 16.9 GB | 21 GB | 29 |
Sizes are estimated from Hugging Face’s parameter counts and the quantisation. Downloads are for the last 30 days.