From 9b1a8286bf0d11629a8f4c0f57d4c855721cd84a Mon Sep 17 00:00:00 2001 From: "dingbin.wang" Date: Fri, 6 Nov 2026 03:59:43 +0000 Subject: [PATCH] r42 r42_train_self.log --- logs/r42_train_self.log | 23 +++++++++++++++++++++++ 1 file changed, 23 insertions(+) create mode 100644 logs/r42_train_self.log diff --git a/logs/r42_train_self.log b/logs/r42_train_self.log new file mode 100644 index 0000000..a09033a --- /dev/null +++ b/logs/r42_train_self.log @@ -0,0 +1,23 @@ +[jev] Phase 1: env setup +[jev] laya.common OK +[jev] data OK (cached) +[jev] Phase 2: download mmBERT (ModelScope) +[jev] config.json: 0MB +[jev] tokenizer.json: 16MB +[jev] tokenizer_config.json: 0MB +[jev] special_tokens_map.json: 0MB +[jev] pytorch_model.bin: 1174MB +[jev] device=cuda tiny=False max_len=1024 head_len=160 win=800 aug_p=0.2 +[jev] Phase 3: build DecisionModel +[jev] encoder load: ckpt=138 matched=134 unexpected=4 core_missing=0 +[jev] model on cuda, trainable 17,267,203 +[jev] crosspair long negs: +3200, en_pos oversample: +2511 (total 26337) +[jev] train=26337 calib=5291 +[jev] ep-1 pretrain calib {'auc': 0.3813, 'ece': 0.0245, 'n': 5291} +[jev] saved best (auc=0.9481) +[jev] saved best (auc=0.9570) +[jev] TRAINING-DONE best_auc=0.9570 +[jev] CALIBRATION T=1.60 raw(ece=0.029) calibrated(ece=0.0225) auc=0.9948 +[jev] AGG-CALIB {'n': 300, 'auc_max_raw': 0.6232, 'ece_max_raw': 0.0787, 'T_agg': 0.6, 'ece_max_calibrated': 0.072} +[jev] Phase 5: push back +[jev] adapter params=17,267,206