/CS_T0828_HW4-Super_Resolution

super resolution

Primary LanguagePython

CS_T0828_HW4

Code for Selected Topics in Visual Recognition using Deep Learning(2020 Autumn) HW4

Hardware

ubuntu 16.04 LTS Intel® Core™ i9-10900 CPU @ 3.70GHz x 20 RTX 2080 Ti

Requirements

My Environment settings are below:

  • Python = 3.7.9
  • pandas = 1.1.3
  • opencv = 4.4.0

In this work, I used RCAN model from thstkdgus35/EDSR-PyTorch, so you should also followed the requirements:

Python 3.6
PyTorch >= 1.0.0
numpy
skimage
imageio
matplotlib
tqdm
cv2 >= 3.xx (Only if you want to use video input/output)

Reproducing Submission

To Reproduct the submission, do the folowed steps

  1. Framework Download and Setting
  2. Dataset Preparation
  3. Training
  4. Testing

Framework Download and Setting

$ git clone https://github.com/thstkdgus35/EDSR-PyTorch.git

$ git clone https://github.com/LCA-0907/CS_T0828_HW4.git

copy my code to the corresponding src dir in EDSR-Pytorch

It should located like:

+EDSR_PyTorch
| +src
| | +data
| | | custom.py
| | | .....
| | option.py
| | prepare_data.py
| | template.py
| | run.sh
| | .....

Dataset Preparation

Download dataset from google drive Use prepare_data.py to generate LR training images.

Training

To train the model, use command line as python3 main.py --template RCAN --save RCAN_BIX3_G10R20P48 --scale 3 --reset --save_results --patch_size 72 --n_threads 1 --ext img --batch_size 64 or run run.sh

Testing

To test the model, use command python3 main.py --model RCAN --test_only --scale 3 --pre_train ../experiment/RCAN_BIX3_G10R20P48/model/model_best.pt --save_results --ext img --data_test Demo --template RCAN or run run.sh

tags: Deep Learning CV Super resolution