fp16 Open on Hugging Face
Parameters
33B
Weights (est.)
≈ 65.5 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-lm on an Apple silicon Mac:
pip install mlx-lmmlx_lm.generate --model mlx-community/shisa-v2-qwen2.5-32b-fp16 --prompt "Hello"
Details
- Base model
- shisa-ai/shisa-v2-qwen2.5-32b (finetune)
- Vendor
- shisa-ai
- Task
- Text generation
- Quantisation
- fp16
- Licence
- apache-2.0
- Downloads
- 27 in the last 30 days · 272 all time · 0 likes
- Published
- 2025-04-14 · updated 2025-04-14 (1y ago)
- Languages
- ja, en
- Tags
- qwen2
Other MLX builds of shisa-ai/shisa-v2-qwen2.5-32b
| Build | Quantisation | Weights | Mac memory | Downloads |
|---|---|---|---|---|
| shisa-v2-qwen2.5-32b-4bit | 4-bit | ≈ 18.4 GB | 24 GB+ | 20 |
| shisa-v2-qwen2.5-32b-6bit | 6-bit | ≈ 26.6 GB | 32 GB+ | 21 |
| shisa-v2-qwen2.5-32b-8bit | 8-bit | ≈ 34.8 GB | 48 GB+ | 34 |
| shisa-v2-qwen2.5-32b-fp16 (this one) | fp16 | ≈ 65.5 GB | 96 GB+ | 27 |