Saumya-svm / WindPMF

Source code for spatiotemporal learning for forecasting of wind frequency maps.

WindPMF

To run : python3 main.py --eval 1 --weights_path <path> --params_apth <params_path>

Command Line Arguments

• --train (int, default=0): Flag to indicate whether to train the model. Set to 1 to enable training.
• --eval (int, default=1): Flag to indicate whether to evaluate the model. Set to 1 to enable evaluation.
• --opt (int, default=0): Flag to indicate whether to optimize the model hyperparameters. Set to 1 to enable optimization.
• --weights_path (str, default=None): Path to the pre-trained model weights. If not provided, the model will be initialized with random weights.
• --params_path (str, default=None): Path to the file containing optimized hyperparameters. If not provided, default parameters will be used.
• --ini_learning_rate (float, default=8e-3): Initial learning rate for the optimizer.
• --decay_rate (float, default=0.85): Decay rate for the learning rate scheduler.
• --step_size (int, default=75): Step size for the learning rate scheduler.
• --seq_len (int, default=16): Sequence length for the input data.
• --numBlocks (int, default=3): Number of blocks in the model architecture.
• --numLayers (int, default=3): Number of layers in each block of the model.
• --filters (int, default=64): Number of filters in the convolutional layers.