2025-12-01 03:03:28 - eval_codegen - INFO - Logging to: logs/codegen/eval_codegen_20251201_030328.log 2025-12-01 03:03:28 - eval_codegen - INFO - Monitor progress: tail -f logs/codegen/eval_codegen_20251201_030328.log 2025-12-01 03:03:28 - eval_codegen - INFO - ============================================================ 2025-12-01 03:03:28 - eval_codegen - INFO - CodeGen Evaluation 2025-12-01 03:03:28 - eval_codegen - INFO - ============================================================ 2025-12-01 03:03:28 - eval_codegen - INFO - Using GPU: NVIDIA GeForce RTX 5090 2025-12-01 03:03:28 - eval_codegen - INFO - Configuration: 2025-12-01 03:03:28 - eval_codegen - INFO - checkpoint: model/checkpoints/run1-java-codegen 2025-12-01 03:03:28 - eval_codegen - INFO - data: datasets/java 2025-12-01 03:03:28 - eval_codegen - INFO - k: 5 2025-12-01 03:03:28 - eval_codegen - INFO - batch_size: 8 2025-12-01 03:03:28 - eval_codegen - INFO - device: cuda 2025-12-01 03:03:28 - eval_codegen - INFO - Loading model from model/checkpoints/run1-java-codegen... 2025-12-01 03:03:29 - eval_codegen - INFO - Model loaded and set to eval mode 2025-12-01 03:03:29 - eval_codegen - INFO - Loaded 34496 test examples 2025-12-01 03:03:29 - eval_codegen - INFO - Starting evaluation with batch_size=8, k=5 2025-12-01 03:03:29 - eval_codegen - INFO - Total batches: 4312 2025-12-01 05:11:39 - eval_codegen - INFO - Evaluation completed in 7691.71 seconds 2025-12-01 05:11:39 - eval_codegen - INFO - Average speed: 4.48 samples/second 2025-12-01 05:11:39 - eval_codegen - INFO - Metrics saved to: model/metrics/run1-java-codegen/metrics.json 2025-12-01 05:11:40 - eval_codegen - INFO - Detailed results saved to: model/metrics/run1-java-codegen/detailed_results.jsonl 2025-12-01 05:11:40 - eval_codegen - INFO - ============================================================ 2025-12-01 05:11:40 - eval_codegen - INFO - EVALUATION RESULTS 2025-12-01 05:11:40 - eval_codegen - INFO - ============================================================ 2025-12-01 05:11:40 - eval_codegen - INFO - n: 34496 2025-12-01 05:11:40 - eval_codegen - INFO - exact_match: 0.4923 2025-12-01 05:11:40 - eval_codegen - INFO - exact_match_case_insensitive: 0.4934 2025-12-01 05:11:40 - eval_codegen - INFO - topk_accuracy: 0.5651 2025-12-01 05:11:40 - eval_codegen - INFO - avg_levenshtein: 7.1085 2025-12-01 05:11:40 - eval_codegen - INFO - k: 5 2025-12-01 05:11:40 - eval_codegen - INFO - device: cuda 2025-12-01 05:11:40 - eval_codegen - INFO - batch_size: 8 2025-12-01 05:11:40 - eval_codegen - INFO - elapsed_time_seconds: 7691.7084 2025-12-01 05:11:40 - eval_codegen - INFO - samples_per_second: 4.4848 2025-12-01 05:11:40 - eval_codegen - INFO - Total evaluation time: 7691.71 seconds (128.20 minutes)