Metadata-Version: 2.3
Name: vsfemasr
Version: 2.0.0
Summary: FeMaSR function for VapourSynth
Project-URL: Homepage, https://github.com/HolyWu/vs-femasr
Project-URL: Issues, https://github.com/HolyWu/vs-femasr/issues
Author-email: HolyWu <holywu@gmail.com>
License-File: LICENSE
Keywords: FeMaSR,PyTorch,TensorRT,VapourSynth
Classifier: Environment :: GPU :: NVIDIA CUDA
Classifier: License :: CC0 1.0 Universal (CC0 1.0) Public Domain Dedication
Classifier: Operating System :: OS Independent
Classifier: Programming Language :: Python :: 3
Classifier: Topic :: Multimedia :: Video
Requires-Python: >=3.10
Requires-Dist: numpy
Requires-Dist: requests
Requires-Dist: timm
Requires-Dist: torch>=2.4.0.dev
Requires-Dist: tqdm
Requires-Dist: vapoursynth>=66
Description-Content-Type: text/markdown

# FeMaSR
Real-World Blind Super-Resolution via Feature Matching with Implicit High-Resolution Priors, based on https://github.com/chaofengc/FeMaSR.


## Dependencies
- [PyTorch](https://pytorch.org/get-started/) 2.4.0.dev or later
- [VapourSynth](http://www.vapoursynth.com/) R66 or later

`trt` requires additional Python packages:
- [TensorRT](https://developer.nvidia.com/tensorrt/) 10.0.1
- [Torch-TensorRT](https://pytorch.org/TensorRT/) 2.4.0.dev

To install TensorRT, run `pip install tensorrt==10.0.1 tensorrt-cu12_bindings==10.0.1 tensorrt-cu12_libs==10.0.1 --extra-index-url https://pypi.nvidia.com`

To install Torch-TensorRT, Windows users can pip install the whl file on [Releases](https://github.com/HolyWu/vs-femasr/releases). Linux users can run `pip install --pre torch_tensorrt --index-url https://download.pytorch.org/whl/nightly/cu124` (requires PyTorch nightly build).


## Installation
```
pip install -U vsfemasr
python -m vsfemasr
```


## Usage
```python
from vsfemasr import femasr

ret = femasr(clip)
```

See `__init__.py` for the description of the parameters.
