diff --git a/logs/train-final-v05.log b/logs/train-final-v05.log index 42b9890..e5bf5bf 100644 --- a/logs/train-final-v05.log +++ b/logs/train-final-v05.log @@ -3,34 +3,32 @@ Some weights of ModernBertForSequenceClassification were not initialized from th You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference. [jev] LoRA targets: ['Wqkv', 'Wo'] trainable params: 2,298,625 || all params: 309,829,634 || trainable%: 0.7419 -[jev] train=37182 calib=5291 -[jev] ep0 step100/2322 loss=0.2687 -[jev] ep0 step200/2322 loss=0.3022 -[jev] ep0 step300/2322 loss=0.3244 -[jev] ep0 step400/2322 loss=0.2449 -[jev] ep0 step500/2322 loss=0.2033 -[jev] ep0 step600/2322 loss=0.2433 -[jev] ep0 step700/2322 loss=0.2197 -[jev] ep0 step800/2322 loss=0.1367 -[jev] ep0 step900/2322 loss=0.1337 -[jev] ep0 step1000/2322 loss=0.1400 -[jev] ep0 step1100/2322 loss=0.1538 +[jev] train=27954 calib=5291 +[jev] canaries: ['en', 'zh'] +[jev] ep0 step100/1746 loss=0.2554 +[jev] ep0 step200/1746 loss=0.4286 +[jev] ep0 step300/1746 loss=0.4042 +[jev] ep0 step400/1746 loss=0.3410 +[jev] ep0 step500/1746 loss=0.3232 +[jev] ep0 step600/1746 loss=0.2050 +[jev] ep0 step700/1746 loss=0.2220 +[jev] ep0 step800/1746 loss=0.2669 /usr/local/lib/python3.10/dist-packages/torch/_inductor/compile_fx.py:282: UserWarning: TensorFloat32 tensor cores for float32 matrix multiplication available but not enabled. Consider setting `torch.set_float32_matmul_precision('high')` for better performance. warnings.warn( -[jev] epoch 0 calibration: {'auc': 0.9947, 'ece': 0.0182, 'n': 5291} +[jev] epoch 0 calibration: {'auc': 0.9936, 'ece': 0.0129, 'n': 5291} +[jev] epoch 0 canary_en: {'auc': 0.8138, 'ece': 0.075, 'n': 300} +[jev] epoch 0 canary_zh: {'auc': 1.0, 'ece': 0.0038, 'n': 300} [jev] live-checkpoint 已写 /tmp/jd/model_live -[jev] ep1 step1200/2322 loss=0.2625 -[jev] ep1 step1300/2322 loss=0.0523 -[jev] ep1 step1400/2322 loss=0.0902 -[jev] ep1 step1500/2322 loss=0.0737 -[jev] ep1 step1600/2322 loss=0.1239 -[jev] ep1 step1700/2322 loss=0.0809 -[jev] ep1 step1800/2322 loss=0.2293 -[jev] ep1 step1900/2322 loss=0.2075 -[jev] ep1 step2000/2322 loss=0.1444 -[jev] ep1 step2100/2322 loss=0.1464 -[jev] ep1 step2200/2322 loss=0.1202 -[jev] ep1 step2300/2322 loss=0.2464 -[jev] epoch 1 calibration: {'auc': 0.9959, 'ece': 0.0082, 'n': 5291} -[jev] live-checkpoint 已写 /tmp/jd/model_live -[jev] DONE best={'auc': 0.9959, 'ece': 0.0082, 'epoch': 1} +[jev] ep1 step900/1746 loss=0.2454 +[jev] ep1 step1000/1746 loss=0.2127 +[jev] ep1 step1100/1746 loss=0.1612 +[jev] ep1 step1200/1746 loss=0.1135 +[jev] ep1 step1300/1746 loss=0.1401 +[jev] ep1 step1400/1746 loss=0.1381 +[jev] ep1 step1500/1746 loss=0.3087 +[jev] ep1 step1600/1746 loss=0.1147 +[jev] ep1 step1700/1746 loss=0.1655 +[jev] epoch 1 calibration: {'auc': 0.9949, 'ece': 0.0105, 'n': 5291} +[jev] epoch 1 canary_en: {'auc': 0.8466, 'ece': 0.066, 'n': 300} +[jev] epoch 1 canary_zh: {'auc': 1.0, 'ece': 0.0032, 'n': 300} +[jev] DONE best={'auc': 0.9936, 'ece': 0.0129, 'epoch': 0} diff --git a/model_parts/metrics.json b/model_parts/metrics.json index f76e35a..c505a7d 100644 --- a/model_parts/metrics.json +++ b/model_parts/metrics.json @@ -1,9 +1,9 @@ { "best": { - "auc": 0.9959, - "ece": 0.0082, - "epoch": 1 + "auc": 0.9936, + "ece": 0.0129, + "epoch": 0 }, "base": "/tmp/base", - "train_pairs": 37182 + "train_pairs": 27954 } \ No newline at end of file diff --git a/model_parts/model__adapter_model.safetensors b/model_parts/model__adapter_model.safetensors index 1f551a2..65e2106 100644 Binary files a/model_parts/model__adapter_model.safetensors and b/model_parts/model__adapter_model.safetensors differ