4-bit Open on Hugging Face
Parameters
—
Weights (est.)
—
Mac memory (est.)
—
Size is estimated from the parameter count and quantisation. Memory allows a fifth extra for context plus 2 GB for the system; long contexts need more.
Run it
With mlx-vlm on an Apple silicon Mac:
pip install mlx-vlmpython -m mlx_vlm.generate --model mlx-community/gemma-3-12b-pt-4bit --image photo.jpg --prompt "Describe this image"
Details
- Base model
- google/gemma-3-12b
- Vendor
- Google · Gemma
- Task
- Vision + text
- Quantisation
- 4-bit
- Licence
- gemma
- Downloads
- 68 in the last 30 days · 1.2K all time · 1 likes
- Published
- 2025-03-12 · updated 2025-03-18 (1y ago)
- Tags
- gemma3
Other MLX builds of google/gemma-3-12b
| Build | Quantisation | Weights | Mac memory | Downloads |
|---|---|---|---|---|
| gemma-3-12b-pt-bf16 | bf16 | ≈ 26.4 GB | 32 GB+ | 31 |
| gemma-3-12b-pt-4bit (this one) | 4-bit | — | — | 68 |
| gemma-3-12b-pt-8bit | 8-bit | — | — | 30 |
| gemma-3-12b-pt-6bit | 6-bit | — | — | 29 |