SR results (up to 64 $\times$ ) for the example climate datasets.

Solar irradiance (DNI & DHI)

Wind velocity (U & V)

Table of Contents

• Requirements
• Pre-trained weights
• Model training
• Model testing

Here is the list of libraries you need to install to execute the code:

• python = 3.6.13
• pytorch = 1.10.2
• scipy
• numpy
• matplotlib
• torchvision

Pre-trained weights and corresponding model configures can be found in Results/**/**/args and Results/**/**/checkpoint. Checkpoints of models at different SR scales can be downloaded from OneDrive.

python train.py --dataset_name Solar_07-14 --lr 0.004 --batch_size 1 --epoch 15,15 --report_step 140000 --trial_name Solar/Solar_07-14_bs1_epoch15_lr4e-3_64X --reset True

python train.py --dataset_name Wind_07-14 --lr 0.002 --batch_size 16 --epoch 30,30 --report_step 7000 --trial_name Wind/Wind_07-14_bs16_epoch30_lr2e-3_64X --reset True

python test.py