Code for Selected Topics in Visual Recognition using Deep Learning(2020 Autumn) HW4
ubuntu 16.04 LTS Intel® Core™ i9-10900 CPU @ 3.70GHz x 20 RTX 2080 Ti
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)
To Reproduct the submission, do the folowed steps
$ 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
| | .....
Download dataset from google drive
Use prepare_data.py
to generate LR training images.
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
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