Python 3.11 Recommended for nano-vllm | Generated by AI
For this project (nano-vllm), the pyproject.toml defines:
requires-python = ">=3.10,<3.13"
So valid choices are Python 3.10, 3.11, or 3.12.
Recommendation: Python 3.11
Here’s why:
| Factor | 3.11 | 3.12 |
|---|---|---|
| triton support ✅ | Excellent, widely tested | Good, but some edge cases historically |
| torch compatibility ✅ | Fully supported | Supported |
| CUDA graph stability ✅ | Most battle-tested | Fine, but newer |
| Package ecosystem ✅ | Maximum compatibility for ML | Some packages were late to support |
Since this project depends on triton>=3.0.0 and torch>=2.4.0 and uses CUDA graphs + tensor parallelism (GPU-centric ML), Python 3.11 is the safest, most battle-tested choice.
How to set it in Zed
- Open Command Palette (
Cmd+Shift+P) - Run “Zed: Select Python Environment”
- Choose your Python 3.11 interpreter
Or create a .python-version file at the project root:
echo "3.11" > /mnt/data/nano-vllm/.python-version
Then select that venv/interpreter in Zed.
