Metadata-Version: 2.4
Name: rgpu
Version: 0.1.0
Summary: A PyTorch device whose tensors live on a remote GPU
Author: Yan Michalevsky
License-Expression: Apache-2.0
Project-URL: Documentation, https://rgpu.pages.dev/docs/
Project-URL: Source, https://github.com/ymcrcat/rgpu
Classifier: Programming Language :: Python :: 3
Classifier: Operating System :: OS Independent
Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
Requires-Python: >=3.10
Description-Content-Type: text/markdown
License-File: LICENSE
Requires-Dist: torch>=2.14
Provides-Extra: test
Requires-Dist: pytest>=8; extra == "test"
Requires-Dist: torchvision; extra == "test"
Dynamic: license-file

# rgpu

A PyTorch device whose tensors live on a remote GPU. Train on a laptop with
`device="rgpu"` while the work runs on a GPU host you reach over SSH.

```sh
pip install rgpu
```

The GPU host runs `rgpu-opserver`; the laptop runs your script through
`rgpu-run --server 127.0.0.1:9720 python train.py`. The protocol has no
authentication, so keep the server on localhost and reach it through an SSH
tunnel.

Documentation: https://rgpu.pages.dev/docs/
