mirror of
https://github.com/astral-sh/uv.git
synced 2025-07-07 21:35:00 +00:00
doc: Sync PyTorch integration index for CUDA and ROCm versions from PyTorch website. (#14100)
## Summary Just to sync the documentation with PyTorch's officially recommended installation method.  **Change:** * CUDA 12.1 to CUDA 12.6 * CUDA 12.4 to CUDA 12.8 * ROCm 6.2 to ROCm 6.3 ## Test Plan Run doc server in local. Result: <details><summary>CUDA 12.6</summary> <p>   </p> </details> <details><summary>CUDA 12.8</summary> <p>   </p> </details> <details><summary>ROCm6</summary> <p>   </p> </details>
This commit is contained in:
parent
e02cd74e64
commit
d653fbb133
1 changed files with 13 additions and 13 deletions
|
@ -85,21 +85,21 @@ In such cases, the first step is to add the relevant PyTorch index to your `pypr
|
|||
explicit = true
|
||||
```
|
||||
|
||||
=== "CUDA 12.1"
|
||||
=== "CUDA 12.6"
|
||||
|
||||
```toml
|
||||
[[tool.uv.index]]
|
||||
name = "pytorch-cu121"
|
||||
url = "https://download.pytorch.org/whl/cu121"
|
||||
name = "pytorch-cu126"
|
||||
url = "https://download.pytorch.org/whl/cu126"
|
||||
explicit = true
|
||||
```
|
||||
|
||||
=== "CUDA 12.4"
|
||||
=== "CUDA 12.8"
|
||||
|
||||
```toml
|
||||
[[tool.uv.index]]
|
||||
name = "pytorch-cu124"
|
||||
url = "https://download.pytorch.org/whl/cu124"
|
||||
name = "pytorch-cu128"
|
||||
url = "https://download.pytorch.org/whl/cu128"
|
||||
explicit = true
|
||||
```
|
||||
|
||||
|
@ -108,7 +108,7 @@ In such cases, the first step is to add the relevant PyTorch index to your `pypr
|
|||
```toml
|
||||
[[tool.uv.index]]
|
||||
name = "pytorch-rocm"
|
||||
url = "https://download.pytorch.org/whl/rocm6.2"
|
||||
url = "https://download.pytorch.org/whl/rocm6.3"
|
||||
explicit = true
|
||||
```
|
||||
|
||||
|
@ -154,7 +154,7 @@ Next, update the `pyproject.toml` to point `torch` and `torchvision` to the desi
|
|||
]
|
||||
```
|
||||
|
||||
=== "CUDA 12.1"
|
||||
=== "CUDA 12.6"
|
||||
|
||||
PyTorch doesn't publish CUDA builds for macOS. As such, we gate on `sys_platform` to instruct uv to limit
|
||||
the PyTorch index to Linux and Windows, falling back to PyPI on macOS:
|
||||
|
@ -162,14 +162,14 @@ Next, update the `pyproject.toml` to point `torch` and `torchvision` to the desi
|
|||
```toml
|
||||
[tool.uv.sources]
|
||||
torch = [
|
||||
{ index = "pytorch-cu121", marker = "sys_platform == 'linux' or sys_platform == 'win32'" },
|
||||
{ index = "pytorch-cu126", marker = "sys_platform == 'linux' or sys_platform == 'win32'" },
|
||||
]
|
||||
torchvision = [
|
||||
{ index = "pytorch-cu121", marker = "sys_platform == 'linux' or sys_platform == 'win32'" },
|
||||
{ index = "pytorch-cu126", marker = "sys_platform == 'linux' or sys_platform == 'win32'" },
|
||||
]
|
||||
```
|
||||
|
||||
=== "CUDA 12.4"
|
||||
=== "CUDA 12.8"
|
||||
|
||||
PyTorch doesn't publish CUDA builds for macOS. As such, we gate on `sys_platform` to instruct uv to limit
|
||||
the PyTorch index to Linux and Windows, falling back to PyPI on macOS:
|
||||
|
@ -177,10 +177,10 @@ Next, update the `pyproject.toml` to point `torch` and `torchvision` to the desi
|
|||
```toml
|
||||
[tool.uv.sources]
|
||||
torch = [
|
||||
{ index = "pytorch-cu124", marker = "sys_platform == 'linux' or sys_platform == 'win32'" },
|
||||
{ index = "pytorch-cu128", marker = "sys_platform == 'linux' or sys_platform == 'win32'" },
|
||||
]
|
||||
torchvision = [
|
||||
{ index = "pytorch-cu124", marker = "sys_platform == 'linux' or sys_platform == 'win32'" },
|
||||
{ index = "pytorch-cu128", marker = "sys_platform == 'linux' or sys_platform == 'win32'" },
|
||||
]
|
||||
```
|
||||
|
||||
|
|
Loading…
Add table
Add a link
Reference in a new issue