Conditional-Adversarial-Domain-Generalization-with-Single-Discriminator

Pytorch implementation of Conditional Adversarial Domain Generalization With a Single Discriminator for Bearing Fault Diagnosis [1].

Script execution

There is a configuration file named train_config.yml. This file contains the following variables that must me configured:

• train_set: Path to the training set .csv file.
• val_set: Path to the validation set .csv file.
• checkpoint: Checkpoint name to save the model.
• train_set: Name of the model to be used as feature extractor. It uses the timm package.

Once the file is properly configured, you must execute the following command:

python train.py -c train_config.yml

References

[1] Q. Zhang et al., "Conditional Adversarial Domain Generalization With a Single Discriminator for Bearing Fault Diagnosis," in IEEE Transactions on Instrumentation and Measurement, vol. 70, pp. 1-15, 2021, Art no. 3514515, doi: 10.1109/TIM.2021.3071350.