Extending xpresso.ai FunctionalityΒΆ
Developers can use the xpresso.ai APIs to extend xpresso.ai functionality. Some examples include:
Scheduling a training pipeline
Checking on the availability of new training data, and automatically starting a training pipeline on the new data
Starting a training pipeline, monitoring its output metrics at regular intervals and pausing it if the metrics (e.g., training accuracy) are not satisfactory
Starting a training pipeline, monitoring its output metrics at regular intervals, pausing it if the metrics (e.g., training accuracy) are not satisfactory and restarting it with changed hyperparameters
Starting several training pipelines in parallel (with different algorithms / different hyperparameters), comparing them while running, and terminating unsatisfactory pipelines