mirrors/uv
1
0
Fork 0
mirror of https://github.com/astral-sh/uv.git synced 2025-08-04 19:08:04 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions 2

Use positive (rather than negative) markers in PyTorch examples (#11112)

Browse source 
## Summary

Maybe slightly controversial because it's more verbose, but we really
want to limit these indexes to Linux and Windows, rather than ignoring
them on Darwin. E.g., we'd also want to ignore them on other platforms.

Further down, I use markers that look like this in the more complete
examples, so this feels more consistent.
This commit is contained in:
Charlie Marsh 2025-01-30 14:00:35 -05:00 • committed by GitHub
parent 5c0fdfd7ce
commit 9f8cd92bd3
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
1 changed files with 18 additions and 18 deletions
Download patch file Download diff file Expand all files Collapse all files

36
docs/guides/integration/pytorch.md
View file

@ -142,53 +142,53 @@ Next, update the `pyproject.toml` to point `torch` and `torchvision` to the desi
=== "CUDA 11.8"
PyTorch doesn't publish CUDA builds for macOS. As such, we gate on `platform_system` to instruct uv to ignore
the PyTorch index when resolving for macOS.
PyTorch doesn't publish CUDA builds for macOS. As such, we gate on `platform_system` to instruct uv to use
the PyTorch index on Linux and Windows, but fall back to PyPI on macOS:
```toml
[tool.uv.sources]
torch = [
{ index = "pytorch-cu118", marker = "platform_system != 'Darwin'" },
{ index = "pytorch-cu118", marker = "platform_system == 'Linux' or platform_system == 'Windows'" },
]
torchvision = [
{ index = "pytorch-cu118", marker = "platform_system != 'Darwin'" },
{ index = "pytorch-cu118", marker = "platform_system == 'Linux' or platform_system == 'Windows'" },
]
```
=== "CUDA 12.1"
PyTorch doesn't publish CUDA builds for macOS. As such, we gate on `platform_system` to instruct uv to ignore
the PyTorch index when resolving for macOS.
PyTorch doesn't publish CUDA builds for macOS. As such, we gate on `platform_system` to instruct uv to limit
the PyTorch index to Linux and Windows, falling back to PyPI on macOS:
```toml
[tool.uv.sources]
torch = [
{ index = "pytorch-cu121", marker = "platform_system != 'Darwin'" },
{ index = "pytorch-cu121", marker = "platform_system == 'Linux' or platform_system == 'Windows'" },
]
torchvision = [
{ index = "pytorch-cu121", marker = "platform_system != 'Darwin'" },
{ index = "pytorch-cu121", marker = "platform_system == 'Linux' or platform_system == 'Windows'" },
]
```
=== "CUDA 12.4"
PyTorch doesn't publish CUDA builds for macOS. As such, we gate on `platform_system` to instruct uv to ignore
the PyTorch index when resolving for macOS.
PyTorch doesn't publish CUDA builds for macOS. As such, we gate on `platform_system` to instruct uv to limit
the PyTorch index to Linux and Windows, falling back to PyPI on macOS:
```toml
[tool.uv.sources]
torch = [
{ index = "pytorch-cu124", marker = "platform_system != 'Darwin'" },
{ index = "pytorch-cu124", marker = "platform_system == 'Linux' or platform_system == 'Windows'" },
]
torchvision = [
{ index = "pytorch-cu124", marker = "platform_system != 'Darwin'" },
{ index = "pytorch-cu124", marker = "platform_system == 'Linux' or platform_system == 'Windows'" },
]
```
=== "ROCm6"
PyTorch doesn't publish ROCm6 builds for macOS or Windows. As such, we gate on `platform_system` to instruct uv to
ignore the PyTorch index when resolving for those platforms.
PyTorch doesn't publish ROCm6 builds for macOS or Windows. As such, we gate on `platform_system` to instruct uv
to limit the PyTorch index to Linux, falling back to PyPI on macOS and Windows:
```toml
[tool.uv.sources]
@ -202,16 +202,16 @@ Next, update the `pyproject.toml` to point `torch` and `torchvision` to the desi
=== "Intel GPUs"
PyTorch doesn't publish Intel GPU builds for macOS. As such, we gate on `platform_system` to instruct uv to ignore
the PyTorch index when resolving for macOS.
PyTorch doesn't publish Intel GPU builds for macOS. As such, we gate on `platform_system` to instruct uv to limit
the PyTorch index to Linux and Windows, falling back to PyPI on macOS:
```toml
[tool.uv.sources]
torch = [
{ index = "pytorch-xpu", marker = "platform_system != 'Darwin'" },
{ index = "pytorch-xpu", marker = "platform_system == 'Linux' or platform_system == 'Windows'" },
]
torchvision = [
{ index = "pytorch-xpu", marker = "platform_system != 'Darwin'" },
{ index = "pytorch-xpu", marker = "platform_system == 'Linux' or platform_system == 'Windows'" },
]
# Intel GPU support relies on `pytorch-triton-xpu` on Linux, which should also be installed from the PyTorch index
# (and included in `project.dependencies`).