Qwen2.5-VL-72B-Instruct-6bit
mlx-community/Qwen2.5-VL-72B-Instruct-6bit
Vision + text · transformers
6-bit Open on Hugging Face
Parameters
73B
Weights (est.)
≈ 59.7 GB
Mac memory (est.)
96 GB+
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/Qwen2.5-VL-72B-Instruct-6bit --image photo.jpg --prompt "Describe this image"
Details
- Base model
- Qwen/Qwen2.5-VL-72B-Instruct (finetune)
- Vendor
- Qwen
- Task
- Vision + text
- Quantisation
- 6-bit
- Licence
- qwen
- Downloads
- 108 in the last 30 days · 1.5K all time · 1 likes
- Published
- 2025-01-29 · updated 2025-02-25 (1y ago)
- Languages
- en
- Tags
- qwen2_5_vl, multimodal
Other MLX builds of Qwen/Qwen2.5-VL-72B-Instruct
| Build | Quantisation | Weights | Mac memory | Downloads |
|---|---|---|---|---|
| Qwen2.5-VL-72B-Instruct-3bit | 3-bit | ≈ 33.2 GB | 48 GB+ | 200 |
| Qwen2.5-VL-72B-Instruct-4bit | 4-bit | ≈ 42.3 GB | 64 GB+ | 1.8K |
| Qwen2.5-VL-72B-Instruct-6bit (this one) | 6-bit | ≈ 59.7 GB | 96 GB+ | 108 |
| Qwen2.5-VL-72B-Instruct-8bit | 8-bit | ≈ 78.0 GB | 96 GB+ | 205 |
| Qwen2.5-VL-72B-Instruct-bf16 | bf16 | ≈ 147 GB | 192 GB+ | 89 |