bgc-setnet / source /scripts /validate_data.py
whiteh4t's picture
Release final BGC retrieval checkpoints and model card
c87881a verified
Raw
History Blame Contribute Delete
694 Bytes
#!/usr/bin/env python3
from __future__ import annotations
import argparse
import json
from bgc_retrieval.config import load_config
from bgc_retrieval.quality import validate_frozen_dataset
def main() -> None:
parser = argparse.ArgumentParser()
parser.add_argument("--config", default="configs/main.yaml")
parser.add_argument("--split", default="data/manifests/silver_split.csv")
parser.add_argument("--output-dir", default="data/manifests")
args = parser.parse_args()
report = validate_frozen_dataset(
load_config(args.config), args.split, args.output_dir
)
print(json.dumps(report, indent=2, sort_keys=True))
if __name__ == "__main__":
main()