ViperEkura b12b24eadc feat: rewrite evaluate_ppl with token-level loss and multi-file support ... - Support multiple input files, glob patterns, and directory input - Add --token_level flag: per-record token_ids + log_probs JSONL output - Add --max_samples for random subsampling per file - LossAccumulator: streaming mode (histogram-based percentiles, low memory) vs exact mode (full token list) - Token type analysis (ascii/cjk/non_ascii/special) when token_level=True - Fix token_ids/log_probs alignment (shift offset) - Cache frozenset(stop_ids) outside loop for performance - Aggregate stats: mean/median/ppl/p50/p90/p95/p99 - Summary JSON with all datasets in one file		2026-07-17 13:14:15 +08:00
..
analyze_weights.py	feat: SVD 有效秩/权重统计分析脚本	2026-06-29 21:39:22 +08:00
evaluate_humaneval.py	fix: reliable test timeout, separate generate/test phases, dynamic pass@k	2026-07-05 08:47:30 +08:00
evaluate_ifd.py	fix: resolve IFD token-set asymmetry and support single-token answers	2026-07-05 17:48:26 +08:00
evaluate_ifeval.py	refactor : metric_logger 改用事件类型 (type=step/validation/epoch)	2026-06-25 17:18:20 +08:00
evaluate_mmlu.py	feat: 新增 IFEval 指令遵循评测	2026-06-16 21:57:34 +08:00
evaluate_ppl.py	feat: rewrite evaluate_ppl with token-level loss and multi-file support	2026-07-17 13:14:15 +08:00
evaluate_rouge.py	feat : add ROUGE evaluation script (manual impl, no deps)	2026-07-05 01:15:01 +08:00