Use as your own risk.
conda create -n env python=3.7 anaconda
conda install pytorch torchvision cpuonly -c pytorch
conda install -c conda-forge opencv
<!-- pip install comet_ml -->
# conda install -c comet_ml -c conda-forge comet_ml # maybe bug
conda install ignite -c pytorch
pip install pytorch-ignite # if conda downgrade pytorch, then use pip
conda install -c conda-forge h5py
conda install scikit-learn
conda install -c anaconda pillow # consider conda-forge
conda install -c anaconda numpy #
conda install -c anaconda matplotlib #
conda install pandas
# pip install torchsummary
# pip install kornia # still buggy not recommended
for CUDA 9.0
shell script conda install pytorch==1.1.0 torchvision==0.3.0 cudatoolkit=9.0 -c pytorch conda install pytorch torchvision cudatoolkit=10.2 -c pytorch conda install pytorch torchvision cudatoolkit=10.1 -c pytorch
Let make /data
folder at root
cd /
sudo mkdir data
sudo chown -R tt /data/
When you use comet.ml
shell script conda install -c comet_ml comet_ml
"" TODO: https://github.com/kornia/kornia/blob/master/kornia/losses/ssim.py
add ssim from here
I have trouble using ssim, the model does not trained. If you are not sure about ssim, avoid it now. Or find big, well maintained open-source implementation of ssim.