Windows compile (#845)

* Custom python 3.13 entry in pyproject.toml

* amend

* Note about compile on Windows

* update
This commit is contained in:
Sebastian Raschka
2025-09-26 12:01:19 -05:00
committed by GitHub
parent 3b83705988
commit 322000d833
2 changed files with 19 additions and 1 deletions

View File

@@ -174,6 +174,24 @@ After:
- `Step tok/sec: 112046`
- `Reserved memory: 6.1875 GB`
<br>
---
**Windows note**
- Compilation can be tricky on Windows
- `torch.compile()` uses Inductor, which JIT-compiles kernels and needs a working C/C++ toolchain
- For CUDA, Inductor also depends on Triton, available via the community package `triton-windows`
- If you see `cl not found`, [install Visual Studio Build Tools with the "C++ workload"](https://learn.microsoft.com/en-us/cpp/build/vscpp-step-0-installation?view=msvc-170) and run Python from the "x64 Native Tools" prompt
- If you see `triton not found` with CUDA, install `triton-windows` (for example, `uv pip install "triton-windows<3.4"`).
- For CPU, a reader further recommended following this [PyTorch Inductor guide for Windows](https://docs.pytorch.org/tutorials/unstable/inductor_windows.html)
- Here, it is important to install the English language package when installing Visual Studio 2022 to avoid a UTF-8 error
- Also, please note that the code needs to be run via the "Visual Studio 2022 Developer Command Prompt" rather than a notebook
- If this setup proves tricky, you can skip compilation; **compilation is optional, and all code examples work fine without it**
---
&nbsp;
### 9. Vocabulary padding