This repository is to do convolutional autoencoder with SetNet for solar flux prediction
- Python 3.5
- PyTorch 1.0.0
- Torchvision 0.2.1
- Sklearn 0.20.2
- PIL 5.4.1
- Panda 0.24.0
- matplotlib 3.0.2
pip install --upgrade fastai torch==1.0.0 pip install torchvision==0.2.1
Download the dataset unzip it and put the folder SDOBenchmark-data-full in the project folder
$ python train_autoencoder.py$ python train_regression.pyDownload pre-trained model weights into the project folder
then open:
demo.ipynb
the files in the project folder should be organized like this:
