- Replace YOUR_TENSORRT_INCLUDE_DIR and YOUR_TENSORRT_LIB_DIR in CMakeLists.txt with your TensorRT include and lib directories.
- generate .wts from pytorch with .pt
git clone https://github.com/xinntao/Real-ESRGAN.git
cd Real-ESRGAN
# Install basicsr - https://github.com/xinntao/BasicSR
# We use BasicSR for both training and inference
pip install basicsr
# facexlib and gfpgan are for face enhancement
pip install facexlib
pip install gfpgan
pip install -r requirements.txt
python setup.py develop
download realesr-general-x4v3.pth (and realesr-general-wdn-x4v3.pth if needed) from https://github.com/xinntao/Real-ESRGAN/releases
cp {tensorrtx}/real-esrgan-general-x4v3/gen_wts.py {xinntao}/Real-ESRGAN
cd {xinntao}/Real-ESRGAN
python gen_wts.py
// a file 'real-esrgan.wts' will be generated.
Be aware that if you need both realesr-general-x4v3.pth and realesr-general-wdn-x4v3.pth, please write a Python script to average all weights of realesr-general-x4v3.pth and realesr-general-wdn-x4v3.pth (from {xinntao}/Real-ESRGAN), then save it as a .pth file, and use this new file to generate a .wts file.
- build tensorrtx/real-esrgan-general-x4v3 and run
cd {tensorrtx}/real-esrgan-general-x4v3/
mkdir build
cd build
cp {xinntao}/Real-ESRGAN/real-esrgan.wts {tensorrtx}/real-esrgan/weights/
cmake ..
make
./real-esrgan your_images_dir