train/sample_count_cache.json

{"file_count": 15, "total_samples": 1619475}