Super-Resolution Implicit Model (SRIM) is a multi-modal image super-resoluition model. This repository contains two implementations of the model:
- One deep neural network featuring Residual-in-Residual Dense Block (RRDB)
- One legacy CNN based network derived from Caffe implementation
Code organization copied from [BasicSR]
- Python 3 (Recommend to use Anaconda)
- PyTorch >= 1.1
- NVIDIA GPU + CUDA
- Python packages:
pip install -r requirements.txt
We use datasets in LDMB format for faster IO speed. Please refer to create_hr_lr_lmdb.py for more details.
To train model with RRDB, please refer to train_srim.sh To train legacy Caffe model, please refer to train_caffe.sh