Upload 8 files

config.json

@@ -0,0 +1,37 @@
+{
+  "_name_or_path": "/home.local/jianwei/workspace/archive/SparseOptimizer/output/Layer_7_12_Hid_160_768_Head_10_12_IMRatio_3.5",
+  "architectures": [
+    "BertForSequenceClassification"
+  ],
+  "attention_probs_dropout_prob": 0.1,
+  "classifier_dropout": null,
+  "embedding_size": 160,
+  "finetuning_task": "stsb",
+  "gradient_checkpointing": false,
+  "hidden_act": "gelu",
+  "hidden_dropout_prob": 0.1,
+  "hidden_size": 160,
+  "id2label": {
+    "0": "LABEL_0"
+  },
+  "initializer_range": 0.02,
+  "intermediate_size": 560,
+  "label2id": {
+    "LABEL_0": 0
+  },
+  "layer_norm_eps": 1e-12,
+  "max_position_embeddings": 512,
+  "model_type": "bert",
+  "num_attention_heads": 10,
+  "num_hidden_layers": 7,
+  "output_intermediate": true,
+  "output_past": true,
+  "pad_token_id": 0,
+  "position_embedding_type": "absolute",
+  "problem_type": "regression",
+  "torch_dtype": "float32",
+  "transformers_version": "4.17.0",
+  "type_vocab_size": 2,
+  "use_cache": true,
+  "vocab_size": 30522
+}

log_bs32_lr3e-05_20221124_035004_897265.txt

@@ -0,0 +1,1294 @@
+------------> log file ==runs2/stsb/OUTPUT_ID/log_bs32_lr3e-05_20221124_035004_897265.txt
+Namespace(aug_train=False, data_dir='/home.local/jianwei/datasets/nlp/glue_data/STS-B', do_eval=False, early_stop=False, early_stop_metric='accuracy', eval_step=120, gradient_accumulation_steps=1, learning_rate=3e-05, local_rank=0, lr_scheduler_type=<SchedulerType.CONSTANT_WITH_WARMUP: 'constant_with_warmup'>, max_length=128, max_train_steps=None, model_name_or_path='/home.local/jianwei/workspace/archive/SparseOptimizer/output/Layer_7_12_Hid_160_768_Head_10_12_IMRatio_3.5', num_train_epochs=30, num_warmup_steps=0, output_dir='runs2/stsb/OUTPUT_ID', pad_to_max_length=False, per_device_eval_batch_size=32, per_device_train_batch_size=32, print_step=5, save_last=False, seed=None, task_name='stsb', train_file=None, use_slow_tokenizer=False, validation_file=None, weight_decay=0.0)
+Distributed environment: NO
+Num processes: 1
+Process index: 0
+Local process index: 0
+Device: cuda
+Mixed precision type: fp16
+
+Sample 4674 of the training set: (tensor([  101, 10079,  3629,  3102,  2048, 12632,  2336,   102, 10079,  4894,
+         8563,  2340, 12632,  2336,   102,     0,     0,     0,     0,     0,
+            0,     0,     0,     0,     0,     0,     0,     0,     0,     0,
+            0,     0,     0,     0,     0,     0,     0,     0,     0,     0,
+            0,     0,     0,     0,     0,     0,     0,     0,     0,     0,
+            0,     0,     0,     0,     0,     0,     0,     0,     0,     0,
+            0,     0,     0,     0,     0,     0,     0,     0,     0,     0,
+            0,     0,     0,     0,     0,     0,     0,     0,     0,     0,
+            0,     0,     0,     0,     0,     0,     0,     0,     0,     0,
+            0,     0,     0,     0,     0,     0,     0,     0,     0,     0,
+            0,     0,     0,     0,     0,     0,     0,     0,     0,     0,
+            0,     0,     0,     0,     0,     0,     0,     0,     0,     0,
+            0,     0,     0,     0,     0,     0,     0,     0]), tensor([1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+        0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+        0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+        0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+        0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+        0, 0, 0, 0, 0, 0, 0, 0]), tensor([0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+        0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+        0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+        0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+        0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+        0, 0, 0, 0, 0, 0, 0, 0]), tensor(3.)).
+Sample 112 of the training set: (tensor([  101,  1037,  2879,  2003,  9361,  1037, 21854, 11563,  1012,   102,
+         1037,  2879,  2003,  2559,  2012,  1037,  8094,  1012,   102,     0,
+            0,     0,     0,     0,     0,     0,     0,     0,     0,     0,
+            0,     0,     0,     0,     0,     0,     0,     0,     0,     0,
+            0,     0,     0,     0,     0,     0,     0,     0,     0,     0,
+            0,     0,     0,     0,     0,     0,     0,     0,     0,     0,
+            0,     0,     0,     0,     0,     0,     0,     0,     0,     0,
+            0,     0,     0,     0,     0,     0,     0,     0,     0,     0,
+            0,     0,     0,     0,     0,     0,     0,     0,     0,     0,
+            0,     0,     0,     0,     0,     0,     0,     0,     0,     0,
+            0,     0,     0,     0,     0,     0,     0,     0,     0,     0,
+            0,     0,     0,     0,     0,     0,     0,     0,     0,     0,
+            0,     0,     0,     0,     0,     0,     0,     0]), tensor([1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0,
+        0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+        0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+        0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+        0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+        0, 0, 0, 0, 0, 0, 0, 0]), tensor([0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0,
+        0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+        0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+        0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+        0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+        0, 0, 0, 0, 0, 0, 0, 0]), tensor(3.8000)).
+Sample 4529 of the training set: (tensor([  101,  3725,  3844,  2015,  2091,  4264,  2004,  3586,  6240,  9446,
+         6561,  2605,   102,  7327,  7767,  1005,  4340,  1005,  2000, 10663,
+         3586,  4168,  4017,  9446,   102,     0,     0,     0,     0,     0,
+            0,     0,     0,     0,     0,     0,     0,     0,     0,     0,
+            0,     0,     0,     0,     0,     0,     0,     0,     0,     0,
+            0,     0,     0,     0,     0,     0,     0,     0,     0,     0,
+            0,     0,     0,     0,     0,     0,     0,     0,     0,     0,
+            0,     0,     0,     0,     0,     0,     0,     0,     0,     0,
+            0,     0,     0,     0,     0,     0,     0,     0,     0,     0,
+            0,     0,     0,     0,     0,     0,     0,     0,     0,     0,
+            0,     0,     0,     0,     0,     0,     0,     0,     0,     0,
+            0,     0,     0,     0,     0,     0,     0,     0,     0,     0,
+            0,     0,     0,     0,     0,     0,     0,     0]), tensor([1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
+        1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+        0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+        0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+        0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+        0, 0, 0, 0, 0, 0, 0, 0]), tensor([0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
+        1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+        0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+        0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+        0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+        0, 0, 0, 0, 0, 0, 0, 0]), tensor(3.)).
+***** Running training *****
+  Num examples = 5749
+  Num Epochs = 30
+  Instantaneous batch size per device = 32
+  Total train batch size (w. parallel, distributed & accumulation) = 32
+  Gradient Accumulation steps = 1
+  Total optimization steps = 5400
+000005/005400, loss: 10.919802, avg_loss: 10.361012
+000010/005400, loss: 10.181771, avg_loss: 9.698319
+000015/005400, loss: 7.925676, avg_loss: 9.470798
+000020/005400, loss: 9.417774, avg_loss: 9.406657
+000025/005400, loss: 11.084503, avg_loss: 9.472598
+000030/005400, loss: 7.033692, avg_loss: 9.447541
+000035/005400, loss: 9.298050, avg_loss: 9.386552
+000040/005400, loss: 8.342388, avg_loss: 9.386284
+000045/005400, loss: 9.821406, avg_loss: 9.483932
+000050/005400, loss: 9.257509, avg_loss: 9.586469
+000055/005400, loss: 8.752683, avg_loss: 9.634466
+000060/005400, loss: 6.560993, avg_loss: 9.559579
+000065/005400, loss: 9.872775, avg_loss: 9.555094
+000070/005400, loss: 9.549786, avg_loss: 9.570707
+000075/005400, loss: 9.400767, avg_loss: 9.534591
+000080/005400, loss: 9.152719, avg_loss: 9.532532
+000085/005400, loss: 10.023327, avg_loss: 9.548860
+000090/005400, loss: 8.150848, avg_loss: 9.539588
+000095/005400, loss: 8.193304, avg_loss: 9.481100
+000100/005400, loss: 8.688814, avg_loss: 9.431278
+000105/005400, loss: 9.266927, avg_loss: 9.390692
+000110/005400, loss: 7.621550, avg_loss: 9.346891
+000115/005400, loss: 6.959364, avg_loss: 9.281569
+000120/005400, loss: 9.679270, avg_loss: 9.291935
+000125/005400, loss: 8.002371, avg_loss: 9.251488
+000130/005400, loss: 8.983469, avg_loss: 9.210261
+000135/005400, loss: 7.914767, avg_loss: 9.138828
+000140/005400, loss: 9.368698, avg_loss: 9.096162
+000145/005400, loss: 7.681985, avg_loss: 9.041475
+000150/005400, loss: 7.530379, avg_loss: 8.976686
+000155/005400, loss: 9.263411, avg_loss: 8.941676
+000160/005400, loss: 7.710734, avg_loss: 8.895261
+000165/005400, loss: 8.456438, avg_loss: 8.831314
+000170/005400, loss: 6.155419, avg_loss: 8.772081
+000175/005400, loss: 8.032525, avg_loss: 8.692823
+000180/005400, loss: 4.489757, avg_loss: 8.613270
+***** Running dev evaluation *****
+  Num examples = 1500
+  Instantaneous batch size per device = 32
+epoch 0, step 180/5400: {'pearson': 0.21495300918671972, 'spearmanr': 0.18778433070729544}
+000185/005400, loss: 7.743507, avg_loss: 8.577872
+000190/005400, loss: 6.030101, avg_loss: 8.510907
+000195/005400, loss: 3.536020, avg_loss: 8.447573
+000200/005400, loss: 6.482443, avg_loss: 8.383874
+000205/005400, loss: 7.495704, avg_loss: 8.331849
+000210/005400, loss: 7.830889, avg_loss: 8.285714
+000215/005400, loss: 7.473868, avg_loss: 8.219836
+000220/005400, loss: 6.685350, avg_loss: 8.164707
+000225/005400, loss: 4.961877, avg_loss: 8.111880
+000230/005400, loss: 5.369789, avg_loss: 8.067034
+000235/005400, loss: 4.157079, avg_loss: 8.007677
+000240/005400, loss: 6.113519, avg_loss: 7.971732
+000245/005400, loss: 4.965279, avg_loss: 7.912654
+000250/005400, loss: 3.810572, avg_loss: 7.868684
+000255/005400, loss: 5.212838, avg_loss: 7.827637
+000260/005400, loss: 5.060454, avg_loss: 7.780213
+000265/005400, loss: 3.830095, avg_loss: 7.720051
+000270/005400, loss: 5.186792, avg_loss: 7.666411
+000275/005400, loss: 5.622235, avg_loss: 7.623420
+000280/005400, loss: 4.717276, avg_loss: 7.579897
+000285/005400, loss: 4.819950, avg_loss: 7.529627
+000290/005400, loss: 5.464397, avg_loss: 7.489963
+000295/005400, loss: 5.470286, avg_loss: 7.442070
+000300/005400, loss: 3.843780, avg_loss: 7.396591
+000305/005400, loss: 3.396843, avg_loss: 7.349362
+000310/005400, loss: 4.573213, avg_loss: 7.293336
+000315/005400, loss: 4.345067, avg_loss: 7.247148
+000320/005400, loss: 4.538530, avg_loss: 7.205413
+000325/005400, loss: 3.374168, avg_loss: 7.165096
+000330/005400, loss: 3.680195, avg_loss: 7.118695
+000335/005400, loss: 3.798603, avg_loss: 7.071580
+000340/005400, loss: 4.418723, avg_loss: 7.028340
+000345/005400, loss: 2.651713, avg_loss: 6.979162
+000350/005400, loss: 4.138247, avg_loss: 6.931522
+000355/005400, loss: 4.034257, avg_loss: 6.891976
+000360/005400, loss: 3.947625, avg_loss: 6.853448
+***** Running dev evaluation *****
+  Num examples = 1500
+  Instantaneous batch size per device = 32
+epoch 1, step 360/5400: {'pearson': 0.1693196142024497, 'spearmanr': 0.1520939753827761}
+000365/005400, loss: 4.066084, avg_loss: 6.813658
+000370/005400, loss: 2.446641, avg_loss: 6.765847
+000375/005400, loss: 3.652923, avg_loss: 6.724291
+000380/005400, loss: 2.925441, avg_loss: 6.683929
+000385/005400, loss: 3.510277, avg_loss: 6.641427
+000390/005400, loss: 3.712820, avg_loss: 6.597568
+000395/005400, loss: 2.864999, avg_loss: 6.559540
+000400/005400, loss: 2.363536, avg_loss: 6.515924
+000405/005400, loss: 3.202157, avg_loss: 6.472551
+000410/005400, loss: 2.507275, avg_loss: 6.427726
+000415/005400, loss: 2.655454, avg_loss: 6.383974
+000420/005400, loss: 3.361968, avg_loss: 6.343995
+000425/005400, loss: 2.212350, avg_loss: 6.302114
+000430/005400, loss: 2.654854, avg_loss: 6.260254
+000435/005400, loss: 2.469006, avg_loss: 6.219799
+000440/005400, loss: 2.423651, avg_loss: 6.179832
+000445/005400, loss: 1.999993, avg_loss: 6.138898
+000450/005400, loss: 3.104252, avg_loss: 6.101511
+000455/005400, loss: 2.722913, avg_loss: 6.065750
+000460/005400, loss: 2.014916, avg_loss: 6.027135
+000465/005400, loss: 2.255650, avg_loss: 5.989807
+000470/005400, loss: 2.582577, avg_loss: 5.953727
+000475/005400, loss: 2.268125, avg_loss: 5.922364
+000480/005400, loss: 2.132411, avg_loss: 5.883134
+000485/005400, loss: 2.092988, avg_loss: 5.850251
+000490/005400, loss: 2.469923, avg_loss: 5.816648
+000495/005400, loss: 1.907046, avg_loss: 5.776493
+000500/005400, loss: 2.188262, avg_loss: 5.740701
+000505/005400, loss: 1.522663, avg_loss: 5.703157
+000510/005400, loss: 1.982296, avg_loss: 5.667968
+000515/005400, loss: 2.409446, avg_loss: 5.635783
+000520/005400, loss: 1.887568, avg_loss: 5.603417
+000525/005400, loss: 2.210217, avg_loss: 5.572377
+000530/005400, loss: 2.381753, avg_loss: 5.541968
+000535/005400, loss: 2.081358, avg_loss: 5.511043
+000540/005400, loss: 2.770565, avg_loss: 5.483432
+***** Running dev evaluation *****
+  Num examples = 1500
+  Instantaneous batch size per device = 32
+epoch 2, step 540/5400: {'pearson': 0.5585231671416229, 'spearmanr': 0.5968823171253705}
+000545/005400, loss: 2.294569, avg_loss: 5.453700
+000550/005400, loss: 1.879893, avg_loss: 5.425492
+000555/005400, loss: 2.054521, avg_loss: 5.396323
+000560/005400, loss: 2.426673, avg_loss: 5.367306
+000565/005400, loss: 1.785937, avg_loss: 5.339570
+000570/005400, loss: 2.125966, avg_loss: 5.312624
+000575/005400, loss: 2.204447, avg_loss: 5.285309
+000580/005400, loss: 1.977976, avg_loss: 5.257472
+000585/005400, loss: 1.667451, avg_loss: 5.227795
+000590/005400, loss: 2.013373, avg_loss: 5.196637
+000595/005400, loss: 1.661575, avg_loss: 5.165030
+000600/005400, loss: 1.761523, avg_loss: 5.134732
+000605/005400, loss: 1.165827, avg_loss: 5.104335
+000610/005400, loss: 1.423938, avg_loss: 5.073547
+000615/005400, loss: 1.275937, avg_loss: 5.045039
+000620/005400, loss: 1.456807, avg_loss: 5.016568
+000625/005400, loss: 1.428447, avg_loss: 4.988611
+000630/005400, loss: 1.340862, avg_loss: 4.960019
+000635/005400, loss: 1.158772, avg_loss: 4.931690
+000640/005400, loss: 1.279753, avg_loss: 4.903349
+000645/005400, loss: 1.495990, avg_loss: 4.875799
+000650/005400, loss: 1.418819, avg_loss: 4.847696
+000655/005400, loss: 1.233781, avg_loss: 4.819313
+000660/005400, loss: 0.825644, avg_loss: 4.790388
+000665/005400, loss: 1.236975, avg_loss: 4.763828
+000670/005400, loss: 1.427844, avg_loss: 4.737722
+000675/005400, loss: 1.194959, avg_loss: 4.710323
+000680/005400, loss: 1.298458, avg_loss: 4.683432
+000685/005400, loss: 1.250220, avg_loss: 4.658527
+000690/005400, loss: 1.528629, avg_loss: 4.632631
+000695/005400, loss: 0.912524, avg_loss: 4.608270
+000700/005400, loss: 0.927178, avg_loss: 4.583745
+000705/005400, loss: 1.425212, avg_loss: 4.560665
+000710/005400, loss: 1.385559, avg_loss: 4.537204
+000715/005400, loss: 1.303016, avg_loss: 4.512879
+000720/005400, loss: 1.559370, avg_loss: 4.489179
+***** Running dev evaluation *****
+  Num examples = 1500
+  Instantaneous batch size per device = 32
+epoch 3, step 720/5400: {'pearson': 0.7538161883822286, 'spearmanr': 0.7339178388810693}
+000725/005400, loss: 1.238668, avg_loss: 4.465238
+000730/005400, loss: 0.980286, avg_loss: 4.441211
+000735/005400, loss: 1.036360, avg_loss: 4.417511
+000740/005400, loss: 0.733856, avg_loss: 4.392772
+000745/005400, loss: 0.911818, avg_loss: 4.369401
+000750/005400, loss: 0.923820, avg_loss: 4.346304
+000755/005400, loss: 1.057140, avg_loss: 4.322930
+000760/005400, loss: 0.671852, avg_loss: 4.300414
+000765/005400, loss: 1.206062, avg_loss: 4.278674
+000770/005400, loss: 1.131734, avg_loss: 4.256801
+000775/005400, loss: 0.727703, avg_loss: 4.235108
+000780/005400, loss: 0.807117, avg_loss: 4.213659
+000785/005400, loss: 0.656183, avg_loss: 4.193212
+000790/005400, loss: 1.307819, avg_loss: 4.173065
+000795/005400, loss: 1.318512, avg_loss: 4.152132
+000800/005400, loss: 0.923019, avg_loss: 4.130066
+000805/005400, loss: 0.358175, avg_loss: 4.109530
+000810/005400, loss: 0.568605, avg_loss: 4.090200
+000815/005400, loss: 0.538159, avg_loss: 4.070320
+000820/005400, loss: 0.791279, avg_loss: 4.050106
+000825/005400, loss: 0.646954, avg_loss: 4.029779
+000830/005400, loss: 0.696995, avg_loss: 4.010299
+000835/005400, loss: 0.851315, avg_loss: 3.991444
+000840/005400, loss: 0.953209, avg_loss: 3.972254
+000845/005400, loss: 0.639867, avg_loss: 3.953383
+000850/005400, loss: 0.828691, avg_loss: 3.934936
+000855/005400, loss: 0.851312, avg_loss: 3.916689
+000860/005400, loss: 0.913527, avg_loss: 3.898366
+000865/005400, loss: 0.769578, avg_loss: 3.880448
+000870/005400, loss: 0.780589, avg_loss: 3.862448
+000875/005400, loss: 0.973308, avg_loss: 3.845167
+000880/005400, loss: 0.829930, avg_loss: 3.827655
+000885/005400, loss: 0.766140, avg_loss: 3.810577
+000890/005400, loss: 0.563716, avg_loss: 3.793318
+000895/005400, loss: 1.047082, avg_loss: 3.776392
+000900/005400, loss: 0.866974, avg_loss: 3.759473
+***** Running dev evaluation *****
+  Num examples = 1500
+  Instantaneous batch size per device = 32
+epoch 4, step 900/5400: {'pearson': 0.8037387020413668, 'spearmanr': 0.8107612065966875}
+000905/005400, loss: 0.766670, avg_loss: 3.742716
+000910/005400, loss: 0.644619, avg_loss: 3.725242
+000915/005400, loss: 0.794441, avg_loss: 3.709127
+000920/005400, loss: 0.583735, avg_loss: 3.693077
+000925/005400, loss: 0.467945, avg_loss: 3.676897
+000930/005400, loss: 0.635556, avg_loss: 3.661061
+000935/005400, loss: 0.546880, avg_loss: 3.644815
+000940/005400, loss: 0.442663, avg_loss: 3.628318
+000945/005400, loss: 0.683668, avg_loss: 3.612901
+000950/005400, loss: 0.656306, avg_loss: 3.597124
+000955/005400, loss: 0.710459, avg_loss: 3.582143
+000960/005400, loss: 0.503140, avg_loss: 3.567162
+000965/005400, loss: 0.659339, avg_loss: 3.552504
+000970/005400, loss: 0.707433, avg_loss: 3.537561
+000975/005400, loss: 0.965483, avg_loss: 3.523352
+000980/005400, loss: 0.855915, avg_loss: 3.508989
+000985/005400, loss: 0.649465, avg_loss: 3.494453
+000990/005400, loss: 0.513151, avg_loss: 3.480050
+000995/005400, loss: 0.907288, avg_loss: 3.465789
+001000/005400, loss: 0.461537, avg_loss: 3.451433
+001005/005400, loss: 0.496157, avg_loss: 3.437648
+001010/005400, loss: 0.989706, avg_loss: 3.424380
+001015/005400, loss: 0.754088, avg_loss: 3.410539
+001020/005400, loss: 0.731938, avg_loss: 3.396362
+001025/005400, loss: 0.844449, avg_loss: 3.382560
+001030/005400, loss: 0.346046, avg_loss: 3.368838
+001035/005400, loss: 0.518788, avg_loss: 3.355767
+001040/005400, loss: 0.714191, avg_loss: 3.342353
+001045/005400, loss: 0.800863, avg_loss: 3.329591
+001050/005400, loss: 0.538331, avg_loss: 3.316277
+001055/005400, loss: 0.645015, avg_loss: 3.303465
+001060/005400, loss: 0.451743, avg_loss: 3.290423
+001065/005400, loss: 0.482815, avg_loss: 3.277498
+001070/005400, loss: 0.428583, avg_loss: 3.264993
+001075/005400, loss: 0.905002, avg_loss: 3.253403
+001080/005400, loss: 0.423076, avg_loss: 3.241331
+***** Running dev evaluation *****
+  Num examples = 1500
+  Instantaneous batch size per device = 32
+epoch 5, step 1080/5400: {'pearson': 0.8115941618503355, 'spearmanr': 0.8282434089896973}
+001085/005400, loss: 0.671325, avg_loss: 3.228906
+001090/005400, loss: 0.510681, avg_loss: 3.216880
+001095/005400, loss: 0.503297, avg_loss: 3.204282
+001100/005400, loss: 0.476207, avg_loss: 3.192340
+001105/005400, loss: 0.287384, avg_loss: 3.180409
+001110/005400, loss: 0.838371, avg_loss: 3.168479
+001115/005400, loss: 0.561100, avg_loss: 3.157175
+001120/005400, loss: 0.461640, avg_loss: 3.145364
+001125/005400, loss: 0.672549, avg_loss: 3.133893
+001130/005400, loss: 0.443830, avg_loss: 3.122226
+001135/005400, loss: 0.465307, avg_loss: 3.110867
+001140/005400, loss: 0.763562, avg_loss: 3.099965
+001145/005400, loss: 0.561359, avg_loss: 3.088965
+001150/005400, loss: 0.411171, avg_loss: 3.077866
+001155/005400, loss: 0.406792, avg_loss: 3.066446
+001160/005400, loss: 0.503313, avg_loss: 3.055675
+001165/005400, loss: 0.475825, avg_loss: 3.045274
+001170/005400, loss: 0.584800, avg_loss: 3.034442
+001175/005400, loss: 0.465069, avg_loss: 3.023891
+001180/005400, loss: 0.494697, avg_loss: 3.013498
+001185/005400, loss: 0.544740, avg_loss: 3.003423
+001190/005400, loss: 0.406965, avg_loss: 2.992763
+001195/005400, loss: 0.268987, avg_loss: 2.982255
+001200/005400, loss: 0.495571, avg_loss: 2.972160
+001205/005400, loss: 0.538762, avg_loss: 2.961849
+001210/005400, loss: 0.478300, avg_loss: 2.952011
+001215/005400, loss: 0.338071, avg_loss: 2.942186
+001220/005400, loss: 0.505288, avg_loss: 2.932113
+001225/005400, loss: 0.570436, avg_loss: 2.922221
+001230/005400, loss: 0.523959, avg_loss: 2.912277
+001235/005400, loss: 0.491099, avg_loss: 2.902272
+001240/005400, loss: 0.559447, avg_loss: 2.892549
+001245/005400, loss: 0.753339, avg_loss: 2.883637
+001250/005400, loss: 0.602193, avg_loss: 2.874184
+001255/005400, loss: 0.302810, avg_loss: 2.864446
+001260/005400, loss: 0.636528, avg_loss: 2.855367
+***** Running dev evaluation *****
+  Num examples = 1500
+  Instantaneous batch size per device = 32
+epoch 6, step 1260/5400: {'pearson': 0.8244669741341696, 'spearmanr': 0.8347289521968146}
+001265/005400, loss: 0.576356, avg_loss: 2.845884
+001270/005400, loss: 0.356003, avg_loss: 2.836981
+001275/005400, loss: 0.282959, avg_loss: 2.827679
+001280/005400, loss: 0.471389, avg_loss: 2.818289
+001285/005400, loss: 0.291599, avg_loss: 2.809166
+001290/005400, loss: 0.309215, avg_loss: 2.799846
+001295/005400, loss: 0.440720, avg_loss: 2.790764
+001300/005400, loss: 0.452717, avg_loss: 2.781574
+001305/005400, loss: 0.379403, avg_loss: 2.772831
+001310/005400, loss: 0.740967, avg_loss: 2.764373
+001315/005400, loss: 0.554469, avg_loss: 2.755583
+001320/005400, loss: 0.422943, avg_loss: 2.747635
+001325/005400, loss: 0.613703, avg_loss: 2.739164
+001330/005400, loss: 0.333465, avg_loss: 2.730182
+001335/005400, loss: 0.531835, avg_loss: 2.721662
+001340/005400, loss: 0.447510, avg_loss: 2.713335
+001345/005400, loss: 0.487799, avg_loss: 2.705467
+001350/005400, loss: 0.629011, avg_loss: 2.697427
+001355/005400, loss: 0.316717, avg_loss: 2.688931
+001360/005400, loss: 0.483824, avg_loss: 2.680822
+001365/005400, loss: 0.420798, avg_loss: 2.672428
+001370/005400, loss: 0.312988, avg_loss: 2.664160
+001375/005400, loss: 0.253772, avg_loss: 2.655796
+001380/005400, loss: 0.507312, avg_loss: 2.648081
+001385/005400, loss: 0.423927, avg_loss: 2.640514
+001390/005400, loss: 0.488432, avg_loss: 2.632712
+001395/005400, loss: 0.496802, avg_loss: 2.624703
+001400/005400, loss: 0.411566, avg_loss: 2.617226
+001405/005400, loss: 0.620914, avg_loss: 2.609520
+001410/005400, loss: 0.529554, avg_loss: 2.602186
+001415/005400, loss: 0.377586, avg_loss: 2.594550
+001420/005400, loss: 0.537113, avg_loss: 2.587398
+001425/005400, loss: 0.502925, avg_loss: 2.579730
+001430/005400, loss: 0.501363, avg_loss: 2.572518
+001435/005400, loss: 0.523148, avg_loss: 2.564881
+001440/005400, loss: 0.283889, avg_loss: 2.557591
+***** Running dev evaluation *****
+  Num examples = 1500
+  Instantaneous batch size per device = 32
+epoch 7, step 1440/5400: {'pearson': 0.8356315632016451, 'spearmanr': 0.8428067774651329}
+001445/005400, loss: 0.283461, avg_loss: 2.549994
+001450/005400, loss: 0.473319, avg_loss: 2.542806
+001455/005400, loss: 0.465852, avg_loss: 2.535222
+001460/005400, loss: 0.452470, avg_loss: 2.528097
+001465/005400, loss: 0.528226, avg_loss: 2.521023
+001470/005400, loss: 0.372980, avg_loss: 2.513948
+001475/005400, loss: 0.580186, avg_loss: 2.507289
+001480/005400, loss: 0.250609, avg_loss: 2.500083
+001485/005400, loss: 0.373619, avg_loss: 2.492741
+001490/005400, loss: 0.313954, avg_loss: 2.485812
+001495/005400, loss: 0.421009, avg_loss: 2.478890
+001500/005400, loss: 0.417312, avg_loss: 2.472097
+001505/005400, loss: 0.419549, avg_loss: 2.465457
+001510/005400, loss: 0.567841, avg_loss: 2.458859
+001515/005400, loss: 0.221651, avg_loss: 2.452159
+001520/005400, loss: 0.323677, avg_loss: 2.445824
+001525/005400, loss: 0.563059, avg_loss: 2.439246
+001530/005400, loss: 0.273469, avg_loss: 2.432506
+001535/005400, loss: 0.230308, avg_loss: 2.425987
+001540/005400, loss: 0.275917, avg_loss: 2.419360
+001545/005400, loss: 0.490302, avg_loss: 2.412818
+001550/005400, loss: 0.171527, avg_loss: 2.406091
+001555/005400, loss: 0.499564, avg_loss: 2.399561
+001560/005400, loss: 0.583477, avg_loss: 2.393275
+001565/005400, loss: 0.422795, avg_loss: 2.387004
+001570/005400, loss: 0.356273, avg_loss: 2.380570
+001575/005400, loss: 0.442116, avg_loss: 2.374079
+001580/005400, loss: 0.380964, avg_loss: 2.367966
+001585/005400, loss: 0.454051, avg_loss: 2.361857
+001590/005400, loss: 0.292075, avg_loss: 2.355417
+001595/005400, loss: 0.433962, avg_loss: 2.349358
+001600/005400, loss: 0.253748, avg_loss: 2.343178
+001605/005400, loss: 0.277990, avg_loss: 2.337058
+001610/005400, loss: 0.658840, avg_loss: 2.331389
+001615/005400, loss: 0.284291, avg_loss: 2.325417
+001620/005400, loss: 0.347131, avg_loss: 2.319557
+***** Running dev evaluation *****
+  Num examples = 1500
+  Instantaneous batch size per device = 32
+epoch 8, step 1620/5400: {'pearson': 0.840875635131036, 'spearmanr': 0.8391187190190564}
+001625/005400, loss: 0.468040, avg_loss: 2.313795
+001630/005400, loss: 0.377569, avg_loss: 2.307781
+001635/005400, loss: 0.373161, avg_loss: 2.301947
+001640/005400, loss: 0.542144, avg_loss: 2.296237
+001645/005400, loss: 0.394721, avg_loss: 2.290366
+001650/005400, loss: 0.313285, avg_loss: 2.284578
+001655/005400, loss: 0.458701, avg_loss: 2.278912
+001660/005400, loss: 0.294037, avg_loss: 2.273092
+001665/005400, loss: 0.288020, avg_loss: 2.267503
+001670/005400, loss: 0.372206, avg_loss: 2.261890
+001675/005400, loss: 0.439113, avg_loss: 2.256269
+001680/005400, loss: 0.265594, avg_loss: 2.250567
+001685/005400, loss: 0.307823, avg_loss: 2.244832
+001690/005400, loss: 0.214900, avg_loss: 2.239233
+001695/005400, loss: 0.430367, avg_loss: 2.234019
+001700/005400, loss: 0.428587, avg_loss: 2.228347
+001705/005400, loss: 0.466478, avg_loss: 2.223007
+001710/005400, loss: 0.406999, avg_loss: 2.217425
+001715/005400, loss: 0.249302, avg_loss: 2.211718
+001720/005400, loss: 0.449824, avg_loss: 2.206581
+001725/005400, loss: 0.200499, avg_loss: 2.201121
+001730/005400, loss: 0.528394, avg_loss: 2.196022
+001735/005400, loss: 0.420790, avg_loss: 2.190833
+001740/005400, loss: 0.393591, avg_loss: 2.185567
+001745/005400, loss: 0.292256, avg_loss: 2.180424
+001750/005400, loss: 0.401385, avg_loss: 2.175266
+001755/005400, loss: 0.294124, avg_loss: 2.169960
+001760/005400, loss: 0.363119, avg_loss: 2.164699
+001765/005400, loss: 0.390154, avg_loss: 2.159830
+001770/005400, loss: 0.313013, avg_loss: 2.154815
+001775/005400, loss: 0.308711, avg_loss: 2.149686
+001780/005400, loss: 0.483320, avg_loss: 2.144812
+001785/005400, loss: 0.379410, avg_loss: 2.139796
+001790/005400, loss: 0.422236, avg_loss: 2.134915
+001795/005400, loss: 0.511399, avg_loss: 2.130093
+001800/005400, loss: 0.423039, avg_loss: 2.125146
+***** Running dev evaluation *****
+  Num examples = 1500
+  Instantaneous batch size per device = 32
+epoch 9, step 1800/5400: {'pearson': 0.8342714757320445, 'spearmanr': 0.8376185602281018}
+001805/005400, loss: 0.486487, avg_loss: 2.120132
+001810/005400, loss: 0.270155, avg_loss: 2.115208
+001815/005400, loss: 0.227492, avg_loss: 2.110093
+001820/005400, loss: 0.346458, avg_loss: 2.105187
+001825/005400, loss: 0.426929, avg_loss: 2.100322
+001830/005400, loss: 0.117478, avg_loss: 2.095436
+001835/005400, loss: 0.279193, avg_loss: 2.090488
+001840/005400, loss: 0.387577, avg_loss: 2.085845
+001845/005400, loss: 0.250648, avg_loss: 2.081071
+001850/005400, loss: 0.303584, avg_loss: 2.076289
+001855/005400, loss: 0.405041, avg_loss: 2.071732
+001860/005400, loss: 0.166183, avg_loss: 2.066910
+001865/005400, loss: 0.319343, avg_loss: 2.062141
+001870/005400, loss: 0.317750, avg_loss: 2.057461
+001875/005400, loss: 0.315497, avg_loss: 2.052864
+001880/005400, loss: 0.338883, avg_loss: 2.048301
+001885/005400, loss: 0.322422, avg_loss: 2.043658
+001890/005400, loss: 0.136494, avg_loss: 2.038912
+001895/005400, loss: 0.384212, avg_loss: 2.034422
+001900/005400, loss: 0.386642, avg_loss: 2.029817
+001905/005400, loss: 0.336843, avg_loss: 2.025262
+001910/005400, loss: 0.378603, avg_loss: 2.020888
+001915/005400, loss: 0.244922, avg_loss: 2.016456
+001920/005400, loss: 0.388475, avg_loss: 2.012008
+001925/005400, loss: 0.275199, avg_loss: 2.007573
+001930/005400, loss: 0.286381, avg_loss: 2.003031
+001935/005400, loss: 0.408020, avg_loss: 1.998613
+001940/005400, loss: 0.296814, avg_loss: 1.994459
+001945/005400, loss: 0.221215, avg_loss: 1.990418
+001950/005400, loss: 0.386474, avg_loss: 1.986272
+001955/005400, loss: 0.186999, avg_loss: 1.981974
+001960/005400, loss: 0.353515, avg_loss: 1.977982
+001965/005400, loss: 0.220710, avg_loss: 1.973756
+001970/005400, loss: 0.522696, avg_loss: 1.969660
+001975/005400, loss: 0.318528, avg_loss: 1.965668
+001980/005400, loss: 0.256884, avg_loss: 1.961408
+***** Running dev evaluation *****
+  Num examples = 1500
+  Instantaneous batch size per device = 32
+epoch 10, step 1980/5400: {'pearson': 0.8390370712384592, 'spearmanr': 0.8380421225427299}
+001985/005400, loss: 0.339906, avg_loss: 1.957260
+001990/005400, loss: 0.177573, avg_loss: 1.953061
+001995/005400, loss: 0.434594, avg_loss: 1.949158
+002000/005400, loss: 0.394058, avg_loss: 1.945127
+002005/005400, loss: 0.284734, avg_loss: 1.941044
+002010/005400, loss: 0.441842, avg_loss: 1.937158
+002015/005400, loss: 0.370813, avg_loss: 1.933077
+002020/005400, loss: 0.231465, avg_loss: 1.929090
+002025/005400, loss: 0.401823, avg_loss: 1.925187
+002030/005400, loss: 0.417580, avg_loss: 1.921148
+002035/005400, loss: 0.233858, avg_loss: 1.917078
+002040/005400, loss: 0.179666, avg_loss: 1.913157
+002045/005400, loss: 0.260741, avg_loss: 1.909101
+002050/005400, loss: 0.221551, avg_loss: 1.905037
+002055/005400, loss: 0.234906, avg_loss: 1.901112
+002060/005400, loss: 0.170529, avg_loss: 1.897019
+002065/005400, loss: 0.246520, avg_loss: 1.893189
+002070/005400, loss: 0.221311, avg_loss: 1.889234
+002075/005400, loss: 0.181704, avg_loss: 1.885389
+002080/005400, loss: 0.418144, avg_loss: 1.881511
+002085/005400, loss: 0.207121, avg_loss: 1.877616
+002090/005400, loss: 0.250038, avg_loss: 1.873798
+002095/005400, loss: 0.266151, avg_loss: 1.869941
+002100/005400, loss: 0.329553, avg_loss: 1.866257
+002105/005400, loss: 0.316394, avg_loss: 1.862574
+002110/005400, loss: 0.202054, avg_loss: 1.858893
+002115/005400, loss: 0.558679, avg_loss: 1.855374
+002120/005400, loss: 0.305135, avg_loss: 1.851792
+002125/005400, loss: 0.306204, avg_loss: 1.848025
+002130/005400, loss: 0.354196, avg_loss: 1.844382
+002135/005400, loss: 0.513295, avg_loss: 1.840886
+002140/005400, loss: 0.338046, avg_loss: 1.837288
+002145/005400, loss: 0.233815, avg_loss: 1.833621
+002150/005400, loss: 0.303081, avg_loss: 1.830035
+002155/005400, loss: 0.217688, avg_loss: 1.826318
+002160/005400, loss: 0.223059, avg_loss: 1.822730
+***** Running dev evaluation *****
+  Num examples = 1500
+  Instantaneous batch size per device = 32
+epoch 11, step 2160/5400: {'pearson': 0.8434982902424131, 'spearmanr': 0.8445651086908786}
+002165/005400, loss: 0.237432, avg_loss: 1.819061
+002170/005400, loss: 0.283776, avg_loss: 1.815507
+002175/005400, loss: 0.309928, avg_loss: 1.811960
+002180/005400, loss: 0.256525, avg_loss: 1.808401
+002185/005400, loss: 0.282268, avg_loss: 1.804922
+002190/005400, loss: 0.277528, avg_loss: 1.801368
+002195/005400, loss: 0.345856, avg_loss: 1.797885
+002200/005400, loss: 0.393328, avg_loss: 1.794652
+002205/005400, loss: 0.224377, avg_loss: 1.791248
+002210/005400, loss: 0.219291, avg_loss: 1.787713
+002215/005400, loss: 0.147671, avg_loss: 1.784197
+002220/005400, loss: 0.339344, avg_loss: 1.780853
+002225/005400, loss: 0.219361, avg_loss: 1.777467
+002230/005400, loss: 0.280020, avg_loss: 1.774036
+002235/005400, loss: 0.261592, avg_loss: 1.770745
+002240/005400, loss: 0.293255, avg_loss: 1.767543
+002245/005400, loss: 0.260899, avg_loss: 1.764155
+002250/005400, loss: 0.251379, avg_loss: 1.760734
+002255/005400, loss: 0.180517, avg_loss: 1.757394
+002260/005400, loss: 0.237342, avg_loss: 1.754018
+002265/005400, loss: 0.348091, avg_loss: 1.750775
+002270/005400, loss: 0.169205, avg_loss: 1.747420
+002275/005400, loss: 0.308270, avg_loss: 1.744165
+002280/005400, loss: 0.265926, avg_loss: 1.740912
+002285/005400, loss: 0.269741, avg_loss: 1.737594
+002290/005400, loss: 0.368088, avg_loss: 1.734481
+002295/005400, loss: 0.288817, avg_loss: 1.731510
+002300/005400, loss: 0.151223, avg_loss: 1.728326
+002305/005400, loss: 0.314602, avg_loss: 1.725295
+002310/005400, loss: 0.204679, avg_loss: 1.722112
+002315/005400, loss: 0.288287, avg_loss: 1.718930
+002320/005400, loss: 0.245926, avg_loss: 1.715852
+002325/005400, loss: 0.204663, avg_loss: 1.712662
+002330/005400, loss: 0.215070, avg_loss: 1.709556
+002335/005400, loss: 0.190882, avg_loss: 1.706442
+002340/005400, loss: 0.224660, avg_loss: 1.703429
+***** Running dev evaluation *****
+  Num examples = 1500
+  Instantaneous batch size per device = 32
+epoch 12, step 2340/5400: {'pearson': 0.8415414818553372, 'spearmanr': 0.8425621296013649}
+002345/005400, loss: 0.207369, avg_loss: 1.700278
+002350/005400, loss: 0.261497, avg_loss: 1.697250
+002355/005400, loss: 0.230280, avg_loss: 1.694103
+002360/005400, loss: 0.262285, avg_loss: 1.690920
+002365/005400, loss: 0.151266, avg_loss: 1.687904
+002370/005400, loss: 0.269719, avg_loss: 1.684892
+002375/005400, loss: 0.354083, avg_loss: 1.681934
+002380/005400, loss: 0.237291, avg_loss: 1.678996
+002385/005400, loss: 0.186130, avg_loss: 1.676010
+002390/005400, loss: 0.260663, avg_loss: 1.673000
+002395/005400, loss: 0.203245, avg_loss: 1.669989
+002400/005400, loss: 0.309466, avg_loss: 1.667078
+002405/005400, loss: 0.167727, avg_loss: 1.664065
+002410/005400, loss: 0.180444, avg_loss: 1.661110
+002415/005400, loss: 0.205075, avg_loss: 1.658129
+002420/005400, loss: 0.251971, avg_loss: 1.655157
+002425/005400, loss: 0.503691, avg_loss: 1.652340
+002430/005400, loss: 0.361796, avg_loss: 1.649719
+002435/005400, loss: 0.220655, avg_loss: 1.646866
+002440/005400, loss: 0.364590, avg_loss: 1.644123
+002445/005400, loss: 0.387156, avg_loss: 1.641263
+002450/005400, loss: 0.321079, avg_loss: 1.638517
+002455/005400, loss: 0.165761, avg_loss: 1.635770
+002460/005400, loss: 0.270390, avg_loss: 1.632963
+002465/005400, loss: 0.202102, avg_loss: 1.630213
+002470/005400, loss: 0.162662, avg_loss: 1.627334
+002475/005400, loss: 0.141903, avg_loss: 1.624407
+002480/005400, loss: 0.130965, avg_loss: 1.621656
+002485/005400, loss: 0.185001, avg_loss: 1.618813
+002490/005400, loss: 0.237992, avg_loss: 1.616033
+002495/005400, loss: 0.158510, avg_loss: 1.613217
+002500/005400, loss: 0.259753, avg_loss: 1.610477
+002505/005400, loss: 0.108687, avg_loss: 1.607702
+002510/005400, loss: 0.179495, avg_loss: 1.604972
+002515/005400, loss: 0.267883, avg_loss: 1.602195
+002520/005400, loss: 0.205575, avg_loss: 1.599474
+***** Running dev evaluation *****
+  Num examples = 1500
+  Instantaneous batch size per device = 32
+epoch 13, step 2520/5400: {'pearson': 0.8425599117367437, 'spearmanr': 0.8414850205786223}
+002525/005400, loss: 0.199653, avg_loss: 1.596711
+002530/005400, loss: 0.201341, avg_loss: 1.593993
+002535/005400, loss: 0.203724, avg_loss: 1.591213
+002540/005400, loss: 0.254623, avg_loss: 1.588562
+002545/005400, loss: 0.369073, avg_loss: 1.585980
+002550/005400, loss: 0.106891, avg_loss: 1.583355
+002555/005400, loss: 0.136818, avg_loss: 1.580702
+002560/005400, loss: 0.231878, avg_loss: 1.577973
+002565/005400, loss: 0.156474, avg_loss: 1.575269
+002570/005400, loss: 0.236511, avg_loss: 1.572622
+002575/005400, loss: 0.257811, avg_loss: 1.570007
+002580/005400, loss: 0.468576, avg_loss: 1.567428
+002585/005400, loss: 0.163139, avg_loss: 1.564778
+002590/005400, loss: 0.436930, avg_loss: 1.562216
+002595/005400, loss: 0.196596, avg_loss: 1.559604
+002600/005400, loss: 0.232763, avg_loss: 1.557100
+002605/005400, loss: 0.164102, avg_loss: 1.554545
+002610/005400, loss: 0.258984, avg_loss: 1.551967
+002615/005400, loss: 0.188581, avg_loss: 1.549408
+002620/005400, loss: 0.215384, avg_loss: 1.546768
+002625/005400, loss: 0.165978, avg_loss: 1.544174
+002630/005400, loss: 0.254275, avg_loss: 1.541621
+002635/005400, loss: 0.260447, avg_loss: 1.539074
+002640/005400, loss: 0.257019, avg_loss: 1.536569
+002645/005400, loss: 0.304152, avg_loss: 1.534171
+002650/005400, loss: 0.172311, avg_loss: 1.531694
+002655/005400, loss: 0.217652, avg_loss: 1.529228
+002660/005400, loss: 0.431580, avg_loss: 1.526855
+002665/005400, loss: 0.342930, avg_loss: 1.524416
+002670/005400, loss: 0.281481, avg_loss: 1.521985
+002675/005400, loss: 0.115055, avg_loss: 1.519483
+002680/005400, loss: 0.190243, avg_loss: 1.517189
+002685/005400, loss: 0.173296, avg_loss: 1.514757
+002690/005400, loss: 0.374071, avg_loss: 1.512512
+002695/005400, loss: 0.322947, avg_loss: 1.510223
+002700/005400, loss: 0.133452, avg_loss: 1.507823
+***** Running dev evaluation *****
+  Num examples = 1500
+  Instantaneous batch size per device = 32
+epoch 14, step 2700/5400: {'pearson': 0.8428262938537643, 'spearmanr': 0.8418967117492774}
+002705/005400, loss: 0.165615, avg_loss: 1.505401
+002710/005400, loss: 0.191277, avg_loss: 1.503067
+002715/005400, loss: 0.186724, avg_loss: 1.500670
+002720/005400, loss: 0.166687, avg_loss: 1.498308
+002725/005400, loss: 0.173368, avg_loss: 1.495943
+002730/005400, loss: 0.182292, avg_loss: 1.493571
+002735/005400, loss: 0.094817, avg_loss: 1.491132
+002740/005400, loss: 0.151966, avg_loss: 1.488704
+002745/005400, loss: 0.118933, avg_loss: 1.486331
+002750/005400, loss: 0.150439, avg_loss: 1.484025
+002755/005400, loss: 0.220458, avg_loss: 1.481835
+002760/005400, loss: 0.165892, avg_loss: 1.479519
+002765/005400, loss: 0.226839, avg_loss: 1.477155
+002770/005400, loss: 0.181736, avg_loss: 1.474844
+002775/005400, loss: 0.103294, avg_loss: 1.472493
+002780/005400, loss: 0.152098, avg_loss: 1.470169
+002785/005400, loss: 0.210727, avg_loss: 1.467948
+002790/005400, loss: 0.218008, avg_loss: 1.465678
+002795/005400, loss: 0.303881, avg_loss: 1.463492
+002800/005400, loss: 0.149363, avg_loss: 1.461267
+002805/005400, loss: 0.278521, avg_loss: 1.459031
+002810/005400, loss: 0.177459, avg_loss: 1.456765
+002815/005400, loss: 0.147072, avg_loss: 1.454549
+002820/005400, loss: 0.154193, avg_loss: 1.452240
+002825/005400, loss: 0.118995, avg_loss: 1.450022
+002830/005400, loss: 0.306946, avg_loss: 1.447801
+002835/005400, loss: 0.203090, avg_loss: 1.445593
+002840/005400, loss: 0.196348, avg_loss: 1.443464
+002845/005400, loss: 0.113525, avg_loss: 1.441222
+002850/005400, loss: 0.305031, avg_loss: 1.439138
+002855/005400, loss: 0.179518, avg_loss: 1.436929
+002860/005400, loss: 0.317867, avg_loss: 1.434791
+002865/005400, loss: 0.244391, avg_loss: 1.432654
+002870/005400, loss: 0.201873, avg_loss: 1.430598
+002875/005400, loss: 0.332513, avg_loss: 1.428486
+002880/005400, loss: 0.174545, avg_loss: 1.426279
+***** Running dev evaluation *****
+  Num examples = 1500
+  Instantaneous batch size per device = 32
+epoch 15, step 2880/5400: {'pearson': 0.8465462185651544, 'spearmanr': 0.8451574856196069}
+002885/005400, loss: 0.140292, avg_loss: 1.424148
+002890/005400, loss: 0.180590, avg_loss: 1.422041
+002895/005400, loss: 0.276235, avg_loss: 1.419968
+002900/005400, loss: 0.079708, avg_loss: 1.417818
+002905/005400, loss: 0.178860, avg_loss: 1.415680
+002910/005400, loss: 0.191974, avg_loss: 1.413542
+002915/005400, loss: 0.160231, avg_loss: 1.411524
+002920/005400, loss: 0.179065, avg_loss: 1.409382
+002925/005400, loss: 0.261529, avg_loss: 1.407299
+002930/005400, loss: 0.196875, avg_loss: 1.405278
+002935/005400, loss: 0.172792, avg_loss: 1.403193
+002940/005400, loss: 0.132129, avg_loss: 1.401091
+002945/005400, loss: 0.143233, avg_loss: 1.398991
+002950/005400, loss: 0.098005, avg_loss: 1.396972
+002955/005400, loss: 0.216378, avg_loss: 1.394936
+002960/005400, loss: 0.168641, avg_loss: 1.392847
+002965/005400, loss: 0.200968, avg_loss: 1.390786
+002970/005400, loss: 0.125896, avg_loss: 1.388788
+002975/005400, loss: 0.244486, avg_loss: 1.386788
+002980/005400, loss: 0.157024, avg_loss: 1.384753
+002985/005400, loss: 0.131733, avg_loss: 1.382739
+002990/005400, loss: 0.180723, avg_loss: 1.380701
+002995/005400, loss: 0.213533, avg_loss: 1.378717
+003000/005400, loss: 0.149431, avg_loss: 1.376713
+003005/005400, loss: 0.145573, avg_loss: 1.374738
+003010/005400, loss: 0.142425, avg_loss: 1.372738
+003015/005400, loss: 0.273710, avg_loss: 1.370737
+003020/005400, loss: 0.164532, avg_loss: 1.368793
+003025/005400, loss: 0.354658, avg_loss: 1.366944
+003030/005400, loss: 0.162812, avg_loss: 1.365036
+003035/005400, loss: 0.225085, avg_loss: 1.363150
+003040/005400, loss: 0.237793, avg_loss: 1.361249
+003045/005400, loss: 0.175477, avg_loss: 1.359304
+003050/005400, loss: 0.220884, avg_loss: 1.357379
+003055/005400, loss: 0.116397, avg_loss: 1.355440
+003060/005400, loss: 0.180262, avg_loss: 1.353549
+***** Running dev evaluation *****
+  Num examples = 1500
+  Instantaneous batch size per device = 32
+epoch 16, step 3060/5400: {'pearson': 0.8475945534372652, 'spearmanr': 0.8462737598699491}
+003065/005400, loss: 0.208348, avg_loss: 1.351671
+003070/005400, loss: 0.162787, avg_loss: 1.349760
+003075/005400, loss: 0.204459, avg_loss: 1.347873
+003080/005400, loss: 0.243172, avg_loss: 1.346001
+003085/005400, loss: 0.105318, avg_loss: 1.344059
+003090/005400, loss: 0.143131, avg_loss: 1.342206
+003095/005400, loss: 0.170449, avg_loss: 1.340305
+003100/005400, loss: 0.208828, avg_loss: 1.338421
+003105/005400, loss: 0.186506, avg_loss: 1.336552
+003110/005400, loss: 0.138573, avg_loss: 1.334692
+003115/005400, loss: 0.199446, avg_loss: 1.332886
+003120/005400, loss: 0.178179, avg_loss: 1.331061
+003125/005400, loss: 0.158329, avg_loss: 1.329155
+003130/005400, loss: 0.132958, avg_loss: 1.327291
+003135/005400, loss: 0.117738, avg_loss: 1.325437
+003140/005400, loss: 0.187024, avg_loss: 1.323590
+003145/005400, loss: 0.285563, avg_loss: 1.321791
+003150/005400, loss: 0.126655, avg_loss: 1.320009
+003155/005400, loss: 0.246144, avg_loss: 1.318180
+003160/005400, loss: 0.222086, avg_loss: 1.316403
+003165/005400, loss: 0.088263, avg_loss: 1.314602
+003170/005400, loss: 0.159250, avg_loss: 1.312750
+003175/005400, loss: 0.232737, avg_loss: 1.311048
+003180/005400, loss: 0.150258, avg_loss: 1.309249
+003185/005400, loss: 0.149525, avg_loss: 1.307465
+003190/005400, loss: 0.175701, avg_loss: 1.305661
+003195/005400, loss: 0.224868, avg_loss: 1.303942
+003200/005400, loss: 0.151383, avg_loss: 1.302172
+003205/005400, loss: 0.216179, avg_loss: 1.300442
+003210/005400, loss: 0.197382, avg_loss: 1.298647
+003215/005400, loss: 0.174374, avg_loss: 1.296861
+003220/005400, loss: 0.146824, avg_loss: 1.295138
+003225/005400, loss: 0.172476, avg_loss: 1.293391
+003230/005400, loss: 0.180328, avg_loss: 1.291636
+003235/005400, loss: 0.219937, avg_loss: 1.289898
+003240/005400, loss: 0.152960, avg_loss: 1.288163
+***** Running dev evaluation *****
+  Num examples = 1500
+  Instantaneous batch size per device = 32
+epoch 17, step 3240/5400: {'pearson': 0.8504202206275068, 'spearmanr': 0.8473922892792047}
+003245/005400, loss: 0.171524, avg_loss: 1.286479
+003250/005400, loss: 0.116338, avg_loss: 1.284752
+003255/005400, loss: 0.086406, avg_loss: 1.283072
+003260/005400, loss: 0.150628, avg_loss: 1.281353
+003265/005400, loss: 0.139414, avg_loss: 1.279617
+003270/005400, loss: 0.193610, avg_loss: 1.277939
+003275/005400, loss: 0.235554, avg_loss: 1.276249
+003280/005400, loss: 0.166258, avg_loss: 1.274573
+003285/005400, loss: 0.263752, avg_loss: 1.272963
+003290/005400, loss: 0.303736, avg_loss: 1.271314
+003295/005400, loss: 0.119213, avg_loss: 1.269609
+003300/005400, loss: 0.132104, avg_loss: 1.267901
+003305/005400, loss: 0.143845, avg_loss: 1.266212
+003310/005400, loss: 0.115098, avg_loss: 1.264532
+003315/005400, loss: 0.288430, avg_loss: 1.262901
+003320/005400, loss: 0.173986, avg_loss: 1.261220
+003325/005400, loss: 0.120085, avg_loss: 1.259552
+003330/005400, loss: 0.248743, avg_loss: 1.257920
+003335/005400, loss: 0.139627, avg_loss: 1.256220
+003340/005400, loss: 0.147467, avg_loss: 1.254561
+003345/005400, loss: 0.142301, avg_loss: 1.252920
+003350/005400, loss: 0.156088, avg_loss: 1.251271
+003355/005400, loss: 0.151669, avg_loss: 1.249613
+003360/005400, loss: 0.214872, avg_loss: 1.248012
+003365/005400, loss: 0.198525, avg_loss: 1.246435
+003370/005400, loss: 0.088710, avg_loss: 1.244759
+003375/005400, loss: 0.120682, avg_loss: 1.243157
+003380/005400, loss: 0.180583, avg_loss: 1.241588
+003385/005400, loss: 0.228067, avg_loss: 1.240034
+003390/005400, loss: 0.126767, avg_loss: 1.238442
+003395/005400, loss: 0.125910, avg_loss: 1.236902
+003400/005400, loss: 0.139716, avg_loss: 1.235308
+003405/005400, loss: 0.080612, avg_loss: 1.233692
+003410/005400, loss: 0.212925, avg_loss: 1.232123
+003415/005400, loss: 0.131897, avg_loss: 1.230545
+003420/005400, loss: 0.205202, avg_loss: 1.228983
+***** Running dev evaluation *****
+  Num examples = 1500
+  Instantaneous batch size per device = 32
+epoch 18, step 3420/5400: {'pearson': 0.8498450703665391, 'spearmanr': 0.8479951774929629}
+003425/005400, loss: 0.122507, avg_loss: 1.227378
+003430/005400, loss: 0.250203, avg_loss: 1.225830
+003435/005400, loss: 0.173522, avg_loss: 1.224317
+003440/005400, loss: 0.087732, avg_loss: 1.222770
+003445/005400, loss: 0.154733, avg_loss: 1.221229
+003450/005400, loss: 0.217941, avg_loss: 1.219678
+003455/005400, loss: 0.137303, avg_loss: 1.218170
+003460/005400, loss: 0.112234, avg_loss: 1.216591
+003465/005400, loss: 0.150905, avg_loss: 1.215047
+003470/005400, loss: 0.158825, avg_loss: 1.213517
+003475/005400, loss: 0.173023, avg_loss: 1.212032
+003480/005400, loss: 0.178021, avg_loss: 1.210536
+003485/005400, loss: 0.247019, avg_loss: 1.209080
+003490/005400, loss: 0.072551, avg_loss: 1.207569
+003495/005400, loss: 0.162839, avg_loss: 1.206022
+003500/005400, loss: 0.189042, avg_loss: 1.204516
+003505/005400, loss: 0.173782, avg_loss: 1.203007
+003510/005400, loss: 0.138777, avg_loss: 1.201515
+003515/005400, loss: 0.177656, avg_loss: 1.200013
+003520/005400, loss: 0.103750, avg_loss: 1.198508
+003525/005400, loss: 0.169574, avg_loss: 1.197020
+003530/005400, loss: 0.119396, avg_loss: 1.195545
+003535/005400, loss: 0.264826, avg_loss: 1.194100
+003540/005400, loss: 0.098011, avg_loss: 1.192637
+003545/005400, loss: 0.088810, avg_loss: 1.191125
+003550/005400, loss: 0.107876, avg_loss: 1.189654
+003555/005400, loss: 0.157520, avg_loss: 1.188192
+003560/005400, loss: 0.176217, avg_loss: 1.186812
+003565/005400, loss: 0.111337, avg_loss: 1.185342
+003570/005400, loss: 0.166201, avg_loss: 1.183889
+003575/005400, loss: 0.171814, avg_loss: 1.182409
+003580/005400, loss: 0.112979, avg_loss: 1.181004
+003585/005400, loss: 0.119157, avg_loss: 1.179598
+003590/005400, loss: 0.114437, avg_loss: 1.178189
+003595/005400, loss: 0.155447, avg_loss: 1.176771
+003600/005400, loss: 0.157078, avg_loss: 1.175317
+***** Running dev evaluation *****
+  Num examples = 1500
+  Instantaneous batch size per device = 32
+epoch 19, step 3600/5400: {'pearson': 0.8482436057295935, 'spearmanr': 0.8472426908693901}
+003605/005400, loss: 0.154441, avg_loss: 1.173877
+003610/005400, loss: 0.100947, avg_loss: 1.172478
+003615/005400, loss: 0.125365, avg_loss: 1.171029
+003620/005400, loss: 0.106434, avg_loss: 1.169605
+003625/005400, loss: 0.130245, avg_loss: 1.168211
+003630/005400, loss: 0.134600, avg_loss: 1.166787
+003635/005400, loss: 0.266648, avg_loss: 1.165400
+003640/005400, loss: 0.144939, avg_loss: 1.164021
+003645/005400, loss: 0.106222, avg_loss: 1.162577
+003650/005400, loss: 0.117357, avg_loss: 1.161193
+003655/005400, loss: 0.202359, avg_loss: 1.159805
+003660/005400, loss: 0.166776, avg_loss: 1.158439
+003665/005400, loss: 0.107025, avg_loss: 1.157045
+003670/005400, loss: 0.143284, avg_loss: 1.155661
+003675/005400, loss: 0.198224, avg_loss: 1.154297
+003680/005400, loss: 0.280506, avg_loss: 1.152964
+003685/005400, loss: 0.130698, avg_loss: 1.151564
+003690/005400, loss: 0.129304, avg_loss: 1.150198
+003695/005400, loss: 0.137243, avg_loss: 1.148803
+003700/005400, loss: 0.097097, avg_loss: 1.147449
+003705/005400, loss: 0.144787, avg_loss: 1.146119
+003710/005400, loss: 0.127824, avg_loss: 1.144796
+003715/005400, loss: 0.176846, avg_loss: 1.143457
+003720/005400, loss: 0.100565, avg_loss: 1.142128
+003725/005400, loss: 0.080043, avg_loss: 1.140760
+003730/005400, loss: 0.125706, avg_loss: 1.139474
+003735/005400, loss: 0.117341, avg_loss: 1.138159
+003740/005400, loss: 0.158067, avg_loss: 1.136843
+003745/005400, loss: 0.151995, avg_loss: 1.135553
+003750/005400, loss: 0.277281, avg_loss: 1.134297
+003755/005400, loss: 0.133230, avg_loss: 1.132962
+003760/005400, loss: 0.186799, avg_loss: 1.131718
+003765/005400, loss: 0.205163, avg_loss: 1.130425
+003770/005400, loss: 0.157280, avg_loss: 1.129118
+003775/005400, loss: 0.250720, avg_loss: 1.127838
+003780/005400, loss: 0.138770, avg_loss: 1.126563
+***** Running dev evaluation *****
+  Num examples = 1500
+  Instantaneous batch size per device = 32
+epoch 20, step 3780/5400: {'pearson': 0.8516633883376111, 'spearmanr': 0.848796837026541}
+003785/005400, loss: 0.280053, avg_loss: 1.125306
+003790/005400, loss: 0.119360, avg_loss: 1.124036
+003795/005400, loss: 0.150453, avg_loss: 1.122750
+003800/005400, loss: 0.150021, avg_loss: 1.121459
+003805/005400, loss: 0.077378, avg_loss: 1.120170
+003810/005400, loss: 0.148403, avg_loss: 1.118906
+003815/005400, loss: 0.178699, avg_loss: 1.117645
+003820/005400, loss: 0.149582, avg_loss: 1.116330
+003825/005400, loss: 0.128546, avg_loss: 1.115013
+003830/005400, loss: 0.268229, avg_loss: 1.113776
+003835/005400, loss: 0.195517, avg_loss: 1.112531
+003840/005400, loss: 0.208493, avg_loss: 1.111268
+003845/005400, loss: 0.193140, avg_loss: 1.110029
+003850/005400, loss: 0.088294, avg_loss: 1.108804
+003855/005400, loss: 0.149382, avg_loss: 1.107547
+003860/005400, loss: 0.198664, avg_loss: 1.106283
+003865/005400, loss: 0.126898, avg_loss: 1.104976
+003870/005400, loss: 0.129632, avg_loss: 1.103741
+003875/005400, loss: 0.123535, avg_loss: 1.102518
+003880/005400, loss: 0.165960, avg_loss: 1.101254
+003885/005400, loss: 0.138942, avg_loss: 1.100020
+003890/005400, loss: 0.128230, avg_loss: 1.098769
+003895/005400, loss: 0.104971, avg_loss: 1.097568
+003900/005400, loss: 0.085618, avg_loss: 1.096347
+003905/005400, loss: 0.126211, avg_loss: 1.095101
+003910/005400, loss: 0.172208, avg_loss: 1.093876
+003915/005400, loss: 0.134293, avg_loss: 1.092629
+003920/005400, loss: 0.130413, avg_loss: 1.091402
+003925/005400, loss: 0.126139, avg_loss: 1.090250
+003930/005400, loss: 0.133957, avg_loss: 1.089023
+003935/005400, loss: 0.235973, avg_loss: 1.087812
+003940/005400, loss: 0.145638, avg_loss: 1.086616
+003945/005400, loss: 0.101992, avg_loss: 1.085413
+003950/005400, loss: 0.126402, avg_loss: 1.084220
+003955/005400, loss: 0.117492, avg_loss: 1.083047
+003960/005400, loss: 0.130239, avg_loss: 1.081839
+***** Running dev evaluation *****
+  Num examples = 1500
+  Instantaneous batch size per device = 32
+epoch 21, step 3960/5400: {'pearson': 0.847780716202824, 'spearmanr': 0.8469865580881132}
+003965/005400, loss: 0.064479, avg_loss: 1.080611
+003970/005400, loss: 0.171006, avg_loss: 1.079406
+003975/005400, loss: 0.085861, avg_loss: 1.078165
+003980/005400, loss: 0.095522, avg_loss: 1.076947
+003985/005400, loss: 0.130852, avg_loss: 1.075774
+003990/005400, loss: 0.134866, avg_loss: 1.074596
+003995/005400, loss: 0.074542, avg_loss: 1.073402
+004000/005400, loss: 0.116856, avg_loss: 1.072228
+004005/005400, loss: 0.105077, avg_loss: 1.071026
+004010/005400, loss: 0.125664, avg_loss: 1.069845
+004015/005400, loss: 0.103024, avg_loss: 1.068720
+004020/005400, loss: 0.128571, avg_loss: 1.067519
+004025/005400, loss: 0.112454, avg_loss: 1.066343
+004030/005400, loss: 0.150399, avg_loss: 1.065202
+004035/005400, loss: 0.073474, avg_loss: 1.064020
+004040/005400, loss: 0.209692, avg_loss: 1.062895
+004045/005400, loss: 0.126410, avg_loss: 1.061723
+004050/005400, loss: 0.168801, avg_loss: 1.060549
+004055/005400, loss: 0.159003, avg_loss: 1.059423
+004060/005400, loss: 0.153256, avg_loss: 1.058304
+004065/005400, loss: 0.139507, avg_loss: 1.057153
+004070/005400, loss: 0.196264, avg_loss: 1.056009
+004075/005400, loss: 0.153815, avg_loss: 1.054860
+004080/005400, loss: 0.182606, avg_loss: 1.053734
+004085/005400, loss: 0.093651, avg_loss: 1.052585
+004090/005400, loss: 0.138306, avg_loss: 1.051489
+004095/005400, loss: 0.125193, avg_loss: 1.050385
+004100/005400, loss: 0.086516, avg_loss: 1.049260
+004105/005400, loss: 0.120107, avg_loss: 1.048157
+004110/005400, loss: 0.246864, avg_loss: 1.047057
+004115/005400, loss: 0.120596, avg_loss: 1.045902
+004120/005400, loss: 0.121840, avg_loss: 1.044833
+004125/005400, loss: 0.141377, avg_loss: 1.043755
+004130/005400, loss: 0.130236, avg_loss: 1.042661
+004135/005400, loss: 0.077593, avg_loss: 1.041535
+004140/005400, loss: 0.096709, avg_loss: 1.040430
+***** Running dev evaluation *****
+  Num examples = 1500
+  Instantaneous batch size per device = 32
+epoch 22, step 4140/5400: {'pearson': 0.8516815294443599, 'spearmanr': 0.8481674736867748}
+004145/005400, loss: 0.107884, avg_loss: 1.039325
+004150/005400, loss: 0.163758, avg_loss: 1.038213
+004155/005400, loss: 0.107559, avg_loss: 1.037071
+004160/005400, loss: 0.221394, avg_loss: 1.035994
+004165/005400, loss: 0.095282, avg_loss: 1.034890
+004170/005400, loss: 0.115735, avg_loss: 1.033791
+004175/005400, loss: 0.120850, avg_loss: 1.032702
+004180/005400, loss: 0.148173, avg_loss: 1.031638
+004185/005400, loss: 0.150222, avg_loss: 1.030549
+004190/005400, loss: 0.167705, avg_loss: 1.029474
+004195/005400, loss: 0.080327, avg_loss: 1.028393
+004200/005400, loss: 0.163523, avg_loss: 1.027316
+004205/005400, loss: 0.091747, avg_loss: 1.026259
+004210/005400, loss: 0.146581, avg_loss: 1.025186
+004215/005400, loss: 0.138113, avg_loss: 1.024116
+004220/005400, loss: 0.126675, avg_loss: 1.023037
+004225/005400, loss: 0.174576, avg_loss: 1.021997
+004230/005400, loss: 0.192664, avg_loss: 1.020943
+004235/005400, loss: 0.075478, avg_loss: 1.019836
+004240/005400, loss: 0.152823, avg_loss: 1.018803
+004245/005400, loss: 0.116004, avg_loss: 1.017760
+004250/005400, loss: 0.151843, avg_loss: 1.016687
+004255/005400, loss: 0.198972, avg_loss: 1.015644
+004260/005400, loss: 0.158850, avg_loss: 1.014584
+004265/005400, loss: 0.140898, avg_loss: 1.013570
+004270/005400, loss: 0.102441, avg_loss: 1.012552
+004275/005400, loss: 0.116065, avg_loss: 1.011494
+004280/005400, loss: 0.093895, avg_loss: 1.010467
+004285/005400, loss: 0.091400, avg_loss: 1.009428
+004290/005400, loss: 0.135847, avg_loss: 1.008452
+004295/005400, loss: 0.131350, avg_loss: 1.007404
+004300/005400, loss: 0.086305, avg_loss: 1.006382
+004305/005400, loss: 0.149123, avg_loss: 1.005382
+004310/005400, loss: 0.077175, avg_loss: 1.004378
+004315/005400, loss: 0.130131, avg_loss: 1.003319
+004320/005400, loss: 0.081299, avg_loss: 1.002324
+***** Running dev evaluation *****
+  Num examples = 1500
+  Instantaneous batch size per device = 32
+epoch 23, step 4320/5400: {'pearson': 0.8516975249826064, 'spearmanr': 0.848650349577711}
+004325/005400, loss: 0.162849, avg_loss: 1.001306
+004330/005400, loss: 0.106408, avg_loss: 1.000275
+004335/005400, loss: 0.112816, avg_loss: 0.999257
+004340/005400, loss: 0.117222, avg_loss: 0.998237
+004345/005400, loss: 0.163939, avg_loss: 0.997231
+004350/005400, loss: 0.132185, avg_loss: 0.996227
+004355/005400, loss: 0.120796, avg_loss: 0.995204
+004360/005400, loss: 0.110429, avg_loss: 0.994174
+004365/005400, loss: 0.176232, avg_loss: 0.993202
+004370/005400, loss: 0.108074, avg_loss: 0.992212
+004375/005400, loss: 0.172169, avg_loss: 0.991219
+004380/005400, loss: 0.115517, avg_loss: 0.990262
+004385/005400, loss: 0.121775, avg_loss: 0.989272
+004390/005400, loss: 0.126275, avg_loss: 0.988271
+004395/005400, loss: 0.107515, avg_loss: 0.987280
+004400/005400, loss: 0.086793, avg_loss: 0.986292
+004405/005400, loss: 0.124593, avg_loss: 0.985330
+004410/005400, loss: 0.132320, avg_loss: 0.984378
+004415/005400, loss: 0.167460, avg_loss: 0.983421
+004420/005400, loss: 0.143033, avg_loss: 0.982441
+004425/005400, loss: 0.123850, avg_loss: 0.981486
+004430/005400, loss: 0.095938, avg_loss: 0.980500
+004435/005400, loss: 0.135725, avg_loss: 0.979532
+004440/005400, loss: 0.081924, avg_loss: 0.978545
+004445/005400, loss: 0.122266, avg_loss: 0.977577
+004450/005400, loss: 0.101690, avg_loss: 0.976582
+004455/005400, loss: 0.159788, avg_loss: 0.975642
+004460/005400, loss: 0.173950, avg_loss: 0.974689
+004465/005400, loss: 0.140697, avg_loss: 0.973738
+004470/005400, loss: 0.118950, avg_loss: 0.972768
+004475/005400, loss: 0.090251, avg_loss: 0.971827
+004480/005400, loss: 0.142421, avg_loss: 0.970929
+004485/005400, loss: 0.093378, avg_loss: 0.969989
+004490/005400, loss: 0.132023, avg_loss: 0.969045
+004495/005400, loss: 0.177814, avg_loss: 0.968117
+004500/005400, loss: 0.098508, avg_loss: 0.967177
+***** Running dev evaluation *****
+  Num examples = 1500
+  Instantaneous batch size per device = 32
+epoch 24, step 4500/5400: {'pearson': 0.8521285961729241, 'spearmanr': 0.8483649029133034}
+004505/005400, loss: 0.110328, avg_loss: 0.966241
+004510/005400, loss: 0.102069, avg_loss: 0.965301
+004515/005400, loss: 0.162850, avg_loss: 0.964387
+004520/005400, loss: 0.151388, avg_loss: 0.963471
+004525/005400, loss: 0.087982, avg_loss: 0.962517
+004530/005400, loss: 0.107709, avg_loss: 0.961596
+004535/005400, loss: 0.093900, avg_loss: 0.960650
+004540/005400, loss: 0.087082, avg_loss: 0.959718
+004545/005400, loss: 0.125253, avg_loss: 0.958788
+004550/005400, loss: 0.091174, avg_loss: 0.957888
+004555/005400, loss: 0.064617, avg_loss: 0.956960
+004560/005400, loss: 0.079881, avg_loss: 0.956014
+004565/005400, loss: 0.103206, avg_loss: 0.955083
+004570/005400, loss: 0.143321, avg_loss: 0.954173
+004575/005400, loss: 0.149977, avg_loss: 0.953271
+004580/005400, loss: 0.092622, avg_loss: 0.952352
+004585/005400, loss: 0.126467, avg_loss: 0.951438
+004590/005400, loss: 0.094483, avg_loss: 0.950528
+004595/005400, loss: 0.107402, avg_loss: 0.949638
+004600/005400, loss: 0.082824, avg_loss: 0.948713
+004605/005400, loss: 0.195657, avg_loss: 0.947822
+004610/005400, loss: 0.124441, avg_loss: 0.946935
+004615/005400, loss: 0.121700, avg_loss: 0.946019
+004620/005400, loss: 0.148236, avg_loss: 0.945115
+004625/005400, loss: 0.140154, avg_loss: 0.944207
+004630/005400, loss: 0.165997, avg_loss: 0.943339
+004635/005400, loss: 0.098995, avg_loss: 0.942421
+004640/005400, loss: 0.120260, avg_loss: 0.941555
+004645/005400, loss: 0.125061, avg_loss: 0.940660
+004650/005400, loss: 0.112413, avg_loss: 0.939759
+004655/005400, loss: 0.104798, avg_loss: 0.938884
+004660/005400, loss: 0.105972, avg_loss: 0.937982
+004665/005400, loss: 0.137016, avg_loss: 0.937103
+004670/005400, loss: 0.086489, avg_loss: 0.936211
+004675/005400, loss: 0.130223, avg_loss: 0.935320
+004680/005400, loss: 0.067240, avg_loss: 0.934422
+***** Running dev evaluation *****
+  Num examples = 1500
+  Instantaneous batch size per device = 32
+epoch 25, step 4680/5400: {'pearson': 0.8527350758782244, 'spearmanr': 0.8507346588341773}
+004685/005400, loss: 0.153279, avg_loss: 0.933544
+004690/005400, loss: 0.084691, avg_loss: 0.932667
+004695/005400, loss: 0.127423, avg_loss: 0.931791
+004700/005400, loss: 0.123418, avg_loss: 0.930916
+004705/005400, loss: 0.102127, avg_loss: 0.930041
+004710/005400, loss: 0.094543, avg_loss: 0.929162
+004715/005400, loss: 0.112821, avg_loss: 0.928298
+004720/005400, loss: 0.094509, avg_loss: 0.927409
+004725/005400, loss: 0.148458, avg_loss: 0.926568
+004730/005400, loss: 0.084626, avg_loss: 0.925698
+004735/005400, loss: 0.074723, avg_loss: 0.924825
+004740/005400, loss: 0.101060, avg_loss: 0.923955
+004745/005400, loss: 0.202314, avg_loss: 0.923109
+004750/005400, loss: 0.055448, avg_loss: 0.922249
+004755/005400, loss: 0.179162, avg_loss: 0.921398
+004760/005400, loss: 0.109892, avg_loss: 0.920548
+004765/005400, loss: 0.128364, avg_loss: 0.919729
+004770/005400, loss: 0.064578, avg_loss: 0.918864
+004775/005400, loss: 0.119190, avg_loss: 0.918013
+004780/005400, loss: 0.106971, avg_loss: 0.917168
+004785/005400, loss: 0.103674, avg_loss: 0.916293
+004790/005400, loss: 0.086906, avg_loss: 0.915437
+004795/005400, loss: 0.078381, avg_loss: 0.914585
+004800/005400, loss: 0.075235, avg_loss: 0.913733
+004805/005400, loss: 0.089770, avg_loss: 0.912904
+004810/005400, loss: 0.104229, avg_loss: 0.912068
+004815/005400, loss: 0.100126, avg_loss: 0.911218
+004820/005400, loss: 0.118553, avg_loss: 0.910386
+004825/005400, loss: 0.143384, avg_loss: 0.909569
+004830/005400, loss: 0.102282, avg_loss: 0.908771
+004835/005400, loss: 0.148549, avg_loss: 0.907961
+004840/005400, loss: 0.182026, avg_loss: 0.907128
+004845/005400, loss: 0.190450, avg_loss: 0.906318
+004850/005400, loss: 0.206217, avg_loss: 0.905508
+004855/005400, loss: 0.064875, avg_loss: 0.904663
+004860/005400, loss: 0.099118, avg_loss: 0.903843
+***** Running dev evaluation *****
+  Num examples = 1500
+  Instantaneous batch size per device = 32
+epoch 26, step 4860/5400: {'pearson': 0.8501907223027365, 'spearmanr': 0.8489084429386828}
+004865/005400, loss: 0.155720, avg_loss: 0.903033
+004870/005400, loss: 0.123857, avg_loss: 0.902210
+004875/005400, loss: 0.106955, avg_loss: 0.901405
+004880/005400, loss: 0.141843, avg_loss: 0.900608
+004885/005400, loss: 0.101737, avg_loss: 0.899809
+004890/005400, loss: 0.159319, avg_loss: 0.899006
+004895/005400, loss: 0.095495, avg_loss: 0.898173
+004900/005400, loss: 0.134695, avg_loss: 0.897373
+004905/005400, loss: 0.062803, avg_loss: 0.896564
+004910/005400, loss: 0.132602, avg_loss: 0.895749
+004915/005400, loss: 0.117661, avg_loss: 0.894927
+004920/005400, loss: 0.134668, avg_loss: 0.894128
+004925/005400, loss: 0.089291, avg_loss: 0.893325
+004930/005400, loss: 0.116079, avg_loss: 0.892549
+004935/005400, loss: 0.092115, avg_loss: 0.891750
+004940/005400, loss: 0.132650, avg_loss: 0.890975
+004945/005400, loss: 0.062088, avg_loss: 0.890193
+004950/005400, loss: 0.062359, avg_loss: 0.889396
+004955/005400, loss: 0.086961, avg_loss: 0.888640
+004960/005400, loss: 0.155230, avg_loss: 0.887873
+004965/005400, loss: 0.110812, avg_loss: 0.887072
+004970/005400, loss: 0.068260, avg_loss: 0.886263
+004975/005400, loss: 0.156115, avg_loss: 0.885500
+004980/005400, loss: 0.124095, avg_loss: 0.884712
+004985/005400, loss: 0.126226, avg_loss: 0.883916
+004990/005400, loss: 0.083915, avg_loss: 0.883123
+004995/005400, loss: 0.083612, avg_loss: 0.882345
+005000/005400, loss: 0.129824, avg_loss: 0.881565
+005005/005400, loss: 0.131232, avg_loss: 0.880788
+005010/005400, loss: 0.122785, avg_loss: 0.879998
+005015/005400, loss: 0.103774, avg_loss: 0.879201
+005020/005400, loss: 0.090597, avg_loss: 0.878425
+005025/005400, loss: 0.084932, avg_loss: 0.877641
+005030/005400, loss: 0.107362, avg_loss: 0.876879
+005035/005400, loss: 0.127814, avg_loss: 0.876147
+005040/005400, loss: 0.170438, avg_loss: 0.875395
+***** Running dev evaluation *****
+  Num examples = 1500
+  Instantaneous batch size per device = 32
+epoch 27, step 5040/5400: {'pearson': 0.8554061134436448, 'spearmanr': 0.8524378109427393}
+005045/005400, loss: 0.076105, avg_loss: 0.874619
+005050/005400, loss: 0.103730, avg_loss: 0.873847
+005055/005400, loss: 0.064459, avg_loss: 0.873081
+005060/005400, loss: 0.112303, avg_loss: 0.872322
+005065/005400, loss: 0.071940, avg_loss: 0.871558
+005070/005400, loss: 0.088598, avg_loss: 0.870804
+005075/005400, loss: 0.090358, avg_loss: 0.870038
+005080/005400, loss: 0.068572, avg_loss: 0.869276
+005085/005400, loss: 0.068552, avg_loss: 0.868498
+005090/005400, loss: 0.140135, avg_loss: 0.867759
+005095/005400, loss: 0.076569, avg_loss: 0.866984
+005100/005400, loss: 0.098298, avg_loss: 0.866226
+005105/005400, loss: 0.108042, avg_loss: 0.865460
+005110/005400, loss: 0.072785, avg_loss: 0.864705
+005115/005400, loss: 0.155046, avg_loss: 0.863964
+005120/005400, loss: 0.225429, avg_loss: 0.863240
+005125/005400, loss: 0.089123, avg_loss: 0.862489
+005130/005400, loss: 0.062908, avg_loss: 0.861729
+005135/005400, loss: 0.050400, avg_loss: 0.860973
+005140/005400, loss: 0.051159, avg_loss: 0.860225
+005145/005400, loss: 0.092017, avg_loss: 0.859488
+005150/005400, loss: 0.119527, avg_loss: 0.858754
+005155/005400, loss: 0.089991, avg_loss: 0.858025
+005160/005400, loss: 0.093003, avg_loss: 0.857288
+005165/005400, loss: 0.096349, avg_loss: 0.856544
+005170/005400, loss: 0.070699, avg_loss: 0.855824
+005175/005400, loss: 0.061141, avg_loss: 0.855089
+005180/005400, loss: 0.117543, avg_loss: 0.854368
+005185/005400, loss: 0.070801, avg_loss: 0.853643
+005190/005400, loss: 0.110298, avg_loss: 0.852924
+005195/005400, loss: 0.114286, avg_loss: 0.852209
+005200/005400, loss: 0.129566, avg_loss: 0.851483
+005205/005400, loss: 0.101796, avg_loss: 0.850774
+005210/005400, loss: 0.125870, avg_loss: 0.850059
+005215/005400, loss: 0.049415, avg_loss: 0.849329
+005220/005400, loss: 0.107189, avg_loss: 0.848612
+***** Running dev evaluation *****
+  Num examples = 1500
+  Instantaneous batch size per device = 32
+epoch 28, step 5220/5400: {'pearson': 0.8540466796613693, 'spearmanr': 0.850937622804088}
+005225/005400, loss: 0.080942, avg_loss: 0.847872
+005230/005400, loss: 0.102161, avg_loss: 0.847163
+005235/005400, loss: 0.082529, avg_loss: 0.846440
+005240/005400, loss: 0.105809, avg_loss: 0.845731
+005245/005400, loss: 0.117093, avg_loss: 0.845006
+005250/005400, loss: 0.106933, avg_loss: 0.844305
+005255/005400, loss: 0.074675, avg_loss: 0.843584
+005260/005400, loss: 0.102407, avg_loss: 0.842881
+005265/005400, loss: 0.148522, avg_loss: 0.842175
+005270/005400, loss: 0.087407, avg_loss: 0.841448
+005275/005400, loss: 0.098112, avg_loss: 0.840739
+005280/005400, loss: 0.092396, avg_loss: 0.840018
+005285/005400, loss: 0.062919, avg_loss: 0.839300
+005290/005400, loss: 0.132550, avg_loss: 0.838632
+005295/005400, loss: 0.145091, avg_loss: 0.837944
+005300/005400, loss: 0.118631, avg_loss: 0.837256
+005305/005400, loss: 0.056487, avg_loss: 0.836545
+005310/005400, loss: 0.103461, avg_loss: 0.835856
+005315/005400, loss: 0.112280, avg_loss: 0.835183
+005320/005400, loss: 0.037065, avg_loss: 0.834495
+005325/005400, loss: 0.102541, avg_loss: 0.833812
+005330/005400, loss: 0.052560, avg_loss: 0.833121
+005335/005400, loss: 0.118150, avg_loss: 0.832437
+005340/005400, loss: 0.093599, avg_loss: 0.831748
+005345/005400, loss: 0.057692, avg_loss: 0.831051
+005350/005400, loss: 0.083881, avg_loss: 0.830350
+005355/005400, loss: 0.092801, avg_loss: 0.829662
+005360/005400, loss: 0.109509, avg_loss: 0.828983
+005365/005400, loss: 0.126566, avg_loss: 0.828295
+005370/005400, loss: 0.090441, avg_loss: 0.827624
+005375/005400, loss: 0.098362, avg_loss: 0.826956
+005380/005400, loss: 0.086417, avg_loss: 0.826275
+005385/005400, loss: 0.090084, avg_loss: 0.825580
+005390/005400, loss: 0.089639, avg_loss: 0.824919
+005395/005400, loss: 0.112607, avg_loss: 0.824232
+005400/005400, loss: 0.079185, avg_loss: 0.823571
+***** Running dev evaluation *****
+  Num examples = 1500
+  Instantaneous batch size per device = 32
+epoch 29, step 5400/5400: {'pearson': 0.8466150052031443, 'spearmanr': 0.845214209063919}
+***** Running train evaluation *****
+  Num examples = 5749
+  Instantaneous batch size per device = 32
+Train Dataset Result: {'pearson': 0.9873720770623174, 'spearmanr': 0.987580466183614}
+***** Running dev evaluation *****
+  Num examples = 1500
+  Instantaneous batch size per device = 32
+Dev Dataset Result: {'pearson': 0.8466150052031443, 'spearmanr': 0.845214209063919}
+Training time 0:04:24

pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:dd2392d68ba15911e69c429109e06a5b13b3cae51981bd4cb24f0a2677f8603a
+size 34298509

result.txt

@@ -0,0 +1,30 @@
+{'pearson': 0.21495300918671972, 'spearmanr': 0.18778433070729544}
+{'pearson': 0.1693196142024497, 'spearmanr': 0.1520939753827761}
+{'pearson': 0.5585231671416229, 'spearmanr': 0.5968823171253705}
+{'pearson': 0.7538161883822286, 'spearmanr': 0.7339178388810693}
+{'pearson': 0.8037387020413668, 'spearmanr': 0.8107612065966875}
+{'pearson': 0.8115941618503355, 'spearmanr': 0.8282434089896973}
+{'pearson': 0.8244669741341696, 'spearmanr': 0.8347289521968146}
+{'pearson': 0.8356315632016451, 'spearmanr': 0.8428067774651329}
+{'pearson': 0.840875635131036, 'spearmanr': 0.8391187190190564}
+{'pearson': 0.8342714757320445, 'spearmanr': 0.8376185602281018}
+{'pearson': 0.8390370712384592, 'spearmanr': 0.8380421225427299}
+{'pearson': 0.8434982902424131, 'spearmanr': 0.8445651086908786}
+{'pearson': 0.8415414818553372, 'spearmanr': 0.8425621296013649}
+{'pearson': 0.8425599117367437, 'spearmanr': 0.8414850205786223}
+{'pearson': 0.8428262938537643, 'spearmanr': 0.8418967117492774}
+{'pearson': 0.8465462185651544, 'spearmanr': 0.8451574856196069}
+{'pearson': 0.8475945534372652, 'spearmanr': 0.8462737598699491}
+{'pearson': 0.8504202206275068, 'spearmanr': 0.8473922892792047}
+{'pearson': 0.8498450703665391, 'spearmanr': 0.8479951774929629}
+{'pearson': 0.8482436057295935, 'spearmanr': 0.8472426908693901}
+{'pearson': 0.8516633883376111, 'spearmanr': 0.848796837026541}
+{'pearson': 0.847780716202824, 'spearmanr': 0.8469865580881132}
+{'pearson': 0.8516815294443599, 'spearmanr': 0.8481674736867748}
+{'pearson': 0.8516975249826064, 'spearmanr': 0.848650349577711}
+{'pearson': 0.8521285961729241, 'spearmanr': 0.8483649029133034}
+{'pearson': 0.8527350758782244, 'spearmanr': 0.8507346588341773}
+{'pearson': 0.8501907223027365, 'spearmanr': 0.8489084429386828}
+{'pearson': 0.8554061134436448, 'spearmanr': 0.8524378109427393}
+{'pearson': 0.8540466796613693, 'spearmanr': 0.850937622804088}
+{'pearson': 0.8466150052031443, 'spearmanr': 0.845214209063919}

special_tokens_map.json

@@ -0,0 +1 @@
+{"unk_token": "[UNK]", "sep_token": "[SEP]", "pad_token": "[PAD]", "cls_token": "[CLS]", "mask_token": "[MASK]"}

tokenizer_config.json

@@ -0,0 +1 @@
+{"do_lower_case": true, "unk_token": "[UNK]", "sep_token": "[SEP]", "pad_token": "[PAD]", "cls_token": "[CLS]", "mask_token": "[MASK]", "tokenize_chinese_chars": true, "strip_accents": null, "do_basic_tokenize": true, "model_max_length": 512, "name_or_path": "/home.local/jianwei/workspace/archive/SparseOptimizer/output/Layer_7_12_Hid_160_768_Head_10_12_IMRatio_3.5", "never_split": null, "special_tokens_map_file": "/home.local/jianwei/.cache/huggingface/transformers/b680d52711d2451bbd6c6b1700365d6d731977c1357ae86bd7227f61145d3be2.dd8bd9bfd3664b530ea4e645105f557769387b3da9f79bdb55ed556bdd80611d", "tokenizer_class": "BertTokenizer"}