def post_process(seg_config_filename, clas_config_filename):
seg_config = load_config(seg_config_filename)
clas_config = load_config(clas_config_filename)
setup_logging(seg_config)
h5.PostProcessor(seg_config["output"]["N"], verbose=2).process(
seg_config["output"]["avg"],
clas_config["output"]["avg"],
seg_config["output"]["img"],
seg_config["output"]["sub"],
)
def average(config_filename):
config = load_config(config_filename)
setup_logging(config)
getattr(h5, config["average"])(config["output"]["N"],
verbose=config["verbose"]).average(
config["output"]["raw"],
config["output"]["avg"],
config["group_weights"])
def predict_all(folder, config_filename):
config = load_config(config_filename)
checkpoints = sorted(list(Path(folder).glob("**/model_best.pth")))
print(f"Performing predictions for {checkpoints}")
for checkpoint in checkpoints:
try:
InferenceManager(config).run(checkpoint)
except Exception as ex:
print(f"Caught exception: {ex}")
def __init__(self, base_config):
self._base = load_config(base_config)
def predict(config_filename, model_checkpoint):
"""
Perform inference using saved model weights, and save to HDF5 database.
"""
config = load_config(config_filename)
InferenceManager(config).run(model_checkpoint)
def train_ensemble(config_filename, num_models):
config = load_config(config_filename)
EnsembleManager(config).start(num_models)
def analyse(config_filename):
seg_config = load_config(config_filename)
setup_logging(seg_config)
h5.HDF5SegAnalyser(seg_config["output"]["N"],
verbose=2).analyse(seg_config["output"]["avg"],
seg_config["output"]["ana"])