Metadata-Version: 2.1
Name: torchsde-brownian
Version: 0.2.5
Summary: Brownian motion implementations extracted from torchsde
Project-URL: Homepage, https://github.com/akx/torchsde-brownian
Author-email: Xuechen Li <lxuechen@cs.stanford.edu>, Patrick Kidger <hello@kidger.site>, Aarni Koskela <akx@iki.fi>
License-Expression: Apache-2.0
License-File: LICENSE
Classifier: License :: OSI Approved :: Apache Software License
Classifier: Programming Language :: Python :: 3
Requires-Python: >=3.6
Requires-Dist: numpy>=1.19
Requires-Dist: torch>=1.6.0
Requires-Dist: trampoline>=0.1.2
Provides-Extra: test
Requires-Dist: pytest; extra == 'test'
Requires-Dist: scipy>=1.5; extra == 'test'
Description-Content-Type: text/markdown

# torchsde_brownian

This library provides the Brownian motion generation code from [torchsde] version 0.2.5
as a standalone package `torchsde_brownian`.

It is mainly meant for use in downstream packages such as `k-diffusion` and `diffusers`,
for where it is a drop-in replacement: the dependency just needs to be changed,
and the import for `torchsde` to `torchsde_brownian`.

According to the authors of `torchsde`, that package is currently in maintenance mode,
so no further development, aside from possible compatibility fixes, is expected.

All credit for the original implementation goes to the authors of [torchsde],
Xuechen Li and Patrick Kidger.

The commit history of this repository has been kept intact so as to preserve
the provenance of the code and their contributions.

## Installation

```shell script
pip install torchsde_brownian
```

**Requirements:** Python >=3.8 and PyTorch >=1.6.0.

## Documentation

Available [here](./DOCUMENTATION.md).

## Citation and references

Please find citations and references in [torchsde]'s readme.

[torchsde]: https://github.com/google-research/torchsde
