MLX Community · First MLX build 5 Jul 2024
nanoLLaVA-1.5
MLX
apache-2.0
A 1.1B parameter model, with 5 MLX builds, from 3-bit to bf16.
- Parameters
- 1.1B From Hugging Face
- Quantisation
- 8-bit nanoLLaVA-1.5-8bit
- Memory needed
- 4 GB Estimate, unified memory
- Download
- 1.1 GB Weights, estimated
- Downloads
- 84 This build, last 30 days
- Task
- —
- MLX builds
- 5 On mlx-community
- Devices
- Mac Apple silicon
Memory
Will it fit?Needs about 4 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/nanoLLaVA-1.5-8bitMLX builds
5 on mlx-community| nanoLLaVA-1.5-3bit | 3-bit | 463 MB | 3 GB | 31 |
|---|---|---|---|---|
| nanoLLaVA-1.5-4bit | 4-bit | 594 MB | 3 GB | 50 |
| nanoLLaVA-1.5-6bit | 6-bit | 857 MB | 3 GB | 33 |
| nanoLLaVA-1.5-8bit | 8-bit | 1.1 GB | 4 GB | 84 |
| nanoLLaVA-1.5-bf16 | bf16 | 2.1 GB | 5 GB | 33 |
Sizes are estimated from Hugging Face’s parameter counts and the quantisation. Downloads are for the last 30 days.