facebook · First MLX build 24 Apr 2026
sapiens2-seg-0.4b
MLX
Image sapiens2-license
A 407M parameter image to image model from facebook, with 8 MLX builds, from 4-bit to bf16.
- Parameters
- 407M From Hugging Face
- Quantisation
- bf16 sapiens2-seg-0.4b-bf16
- Memory needed
- 3 GB Estimate, unified memory
- Download
- 813 MB Weights, estimated
- Downloads
- 58 This build, last 30 days
- Task
- Image to image Image
- MLX builds
- 8 On mlx-community
- Devices
- Mac Apple silicon
Memory
Will it fit?Needs about 3 GB at bf16
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/sapiens2-seg-0.4b-bf16MLX builds
8 on mlx-community| sapiens2-seg-0.4b-4bit | 4-bit | 271 MB | 3 GB | 34 |
|---|---|---|---|---|
| sapiens2-seg-0.4b-nvfp4 | 4-bit | 271 MB | 3 GB | 28 |
| sapiens2-seg-0.4b-mxfp4 | MXFP4 | 271 MB | 3 GB | 26 |
| sapiens2-seg-0.4b-5bit | 5-bit | 320 MB | 3 GB | 23 |
| sapiens2-seg-0.4b-6bit | 6-bit | 370 MB | 3 GB | 25 |
| sapiens2-seg-0.4b-mxfp8 | MXFP8 | 468 MB | 3 GB | 29 |
| sapiens2-seg-0.4b-8bit | 8-bit | 468 MB | 3 GB | 25 |
| sapiens2-seg-0.4b-bf16 | bf16 | 813 MB | 3 GB | 58 |
Sizes are estimated from Hugging Face’s parameter counts and the quantisation. Downloads are for the last 30 days.