Batch processing

ADAM supports batch inference so that you can infer the location of the lake breeze over multiple radar scans with one inference step. Batch inference will save computational time by loading the model only once and utilizing vectorization to perform the inference on multiple radar scans at a time.

import adam

# List of radar scans to process
radar_scans = [
    ('KLOT', '2025-07-15T18:00:00'),
    ('KLOT', '2025-07-15T19:00:00'),
    ('KLOT', '2025-07-15T20:00:00'),
]

# Preprocess all radar scans
preprocessed_scans = [adam.io.preprocess_radar_image(station, time) for station, time in radar_scans]

# Perform batch inference
lake_breeze_results = adam.model.infer_lake_breeze_batch(
    preprocessed_scans, model_name='lakebreeze_model_fcn_resnet50_no_augmentation')

Analyzing the mask data in custom workflows

The RadarImage class contains all of the information you need to perform custom analyses of the lake breeze front.