For all subsequent operations, it's important to maintain the following file structure as indicated in doc_tree.txt, as the location of the dataset folder is crucial for the file-reading processes within main.py:
To test the model, download the model weights from Google Drive: Model Weights.
After downloading the weights to your local folder, you can test the model's performance on the test set with the following command:
python main.py --test-only --model-path model_parameters
Set parameters in config.py following the comments provided within the file. To train the model, run:
python main.py
This command will train the model and automatically save the parameters.
Use the following command for hyperparameter searching, allowing you to set the decay style and number of training epochs:
python hyperparameter.py --decay_style DS --epochs E
The default search parameters include learning rate, batch size, hidden layer sizes, regularization strength, and learning rate decay coefficient. For additional adjustments, modifications can be made directly to hyperparameter.py.
A detailed explanation of each file can be found in the report. And hyperparameter_searching_results include the logging information of the hyperparameter searching process.