Update README.md

README.md

@@ -1,3 +1,8 @@
+---
+tags:
+- super-resolution
+- Image-to-Image
+---
 # Accelerating Image Super-Resolution Networks with Pixel-Level Classification
 [![Project Page](https://img.shields.io/badge/Project-Page-green)](https://3587jjh.github.io/PCSR/)
 [![arXiv](https://img.shields.io/badge/arXiv-2407.21448-b31b1b)](https://arxiv.org/abs/2407.21448)
@@ -13,45 +18,6 @@
 - NVIDIA GPU + CUDA<br>
 - Python packages: `pip install numpy opencv-python pandas tqdm fast_pytorch_kmeans`
 
-## Demo
-We offer a demo that can simulate 4x super-resolution using a single input image:
-```
-python demo.py --img_path <image path> --k <value> [--adaptive] [--no_refinement]
-```
-For a detailed description of the arguments and types of output, please refer to `demo.py` and the `results` directory.
-<div style="display: flex;">
-    <figure style="margin: 0; text-align: center;">
-        <img src="results/PCSR_colored.png" style="width: 280px; height: 184px; object-fit: cover;"/>
-        <figcaption>Classification</figcaption>
-    </figure>
-    <figure style="margin: 0; text-align: center;">
-        <img src="results/PCSR.png" style="width: 280px; height: 184px; object-fit: cover;"/>
-        <figcaption>SR Result (X4)</figcaption>
-    </figure>
-</div>
-
-## Train
-We offer a variety of pretrained models available in the `save` directory.<br>
-
-### Data preprocessing
-We use the DIV2K training set as our base.
-```
-cd datasets/scripts
-python data_augmentation.py
-python extract_subimages.py
-python generate_mod_LR_bic.py
-```
-### Run training scripts (example usage)
-(Please move the pretrained model to a different directory to train a new model!)
-```
-# original model
-./dist.sh train.py --config configs/carn-x4.yaml --gpu 0
-
-# PCSR model (2-stage)
-./dist.sh train_pcsr.py --config configs/carn-pcsr-phase0.yaml --gpu 0,1
-./dist.sh train_pcsr.py --config configs/carn-pcsr-phase1.yaml --gpu 0
-```
-
 ## Test
 Please refer to `test.py` and `test_pcsr.py` for a detailed description of the arguments.
 ```
@@ -73,7 +39,4 @@ python test_pcsr.py --config <config path> --hr_data <hr foler> --scale <integer
       primaryClass={cs.CV},
       url={https://arxiv.org/abs/2407.21448}, 
 }
-```
-
-## Acknowledgement
-This repo is based on [LIIF](https://github.com/yinboc/liif) and [KAIR](https://github.com/cszn/KAIR). 
+```