Training in progress, epoch 1

all_results.json

@@ -1,15 +1,15 @@
 {
-    "epoch": 5.0,
-    "eval_accuracy": 0.2919999957084656,
-    "eval_loss": 1.569107174873352,
-    "eval_runtime": 0.7522,
+    "epoch": 3.0,
+    "eval_accuracy": 0.2160000056028366,
+    "eval_loss": 1.5779266357421875,
+    "eval_runtime": 0.8459,
     "eval_samples": 500,
-    "eval_samples_per_second": 664.689,
-    "eval_steps_per_second": 42.54,
-    "total_flos": 10492735676640.0,
-    "train_loss": 1.5705763679504394,
-    "train_runtime": 78.9966,
+    "eval_samples_per_second": 591.1,
+    "eval_steps_per_second": 37.83,
+    "total_flos": 16105346090400.0,
+    "train_loss": 1.5832408892313639,
+    "train_runtime": 76.8946,
     "train_samples": 4000,
-    "train_samples_per_second": 253.176,
-    "train_steps_per_second": 15.823
+    "train_samples_per_second": 156.058,
+    "train_steps_per_second": 9.754
 }

eval_results.json

@@ -1,9 +1,9 @@
 {
-    "epoch": 5.0,
-    "eval_accuracy": 0.2919999957084656,
-    "eval_loss": 1.569107174873352,
-    "eval_runtime": 0.7522,
+    "epoch": 3.0,
+    "eval_accuracy": 0.2160000056028366,
+    "eval_loss": 1.5779266357421875,
+    "eval_runtime": 0.8459,
     "eval_samples": 500,
-    "eval_samples_per_second": 664.689,
-    "eval_steps_per_second": 42.54
+    "eval_samples_per_second": 591.1,
+    "eval_steps_per_second": 37.83
 }

pytorch_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:a9a88d0d3a073a3f519bc18050c71d3fd14e044502a026012f7c1dd7c2cc0c6e
+oid sha256:7dfedd6c25cc6b194b200602880102dd46490c4ffb2738f2ae4339ced585ecd1
 size 12755881

train_results.json

@@ -1,9 +1,9 @@
 {
-    "epoch": 5.0,
-    "total_flos": 10492735676640.0,
-    "train_loss": 1.5705763679504394,
-    "train_runtime": 78.9966,
+    "epoch": 3.0,
+    "total_flos": 16105346090400.0,
+    "train_loss": 1.5832408892313639,
+    "train_runtime": 76.8946,
     "train_samples": 4000,
-    "train_samples_per_second": 253.176,
-    "train_steps_per_second": 15.823
+    "train_samples_per_second": 156.058,
+    "train_steps_per_second": 9.754
 }

trainer_state.json

@@ -1,442 +1,274 @@
 {
-  "best_metric": 1.569107174873352,
-  "best_model_checkpoint": "chinese_roberta_L-2_H-128-finetuned-MC-hyper/checkpoint-1250",
-  "epoch": 5.0,
-  "global_step": 1250,
+  "best_metric": 1.5779266357421875,
+  "best_model_checkpoint": "chinese_roberta_L-2_H-128-finetuned-MC-hyper/checkpoint-750",
+  "epoch": 3.0,
+  "global_step": 750,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
     {
       "epoch": 0.08,
-      "learning_rate": 4.92e-05,
-      "loss": 1.6109,
+      "learning_rate": 4.866666666666667e-05,
+      "loss": 1.6118,
       "step": 20
     },
     {
       "epoch": 0.16,
-      "learning_rate": 4.8400000000000004e-05,
-      "loss": 1.6093,
+      "learning_rate": 4.7333333333333336e-05,
+      "loss": 1.6098,
       "step": 40
     },
     {
       "epoch": 0.24,
-      "learning_rate": 4.76e-05,
-      "loss": 1.6059,
+      "learning_rate": 4.600000000000001e-05,
+      "loss": 1.606,
       "step": 60
     },
     {
       "epoch": 0.32,
-      "learning_rate": 4.6800000000000006e-05,
-      "loss": 1.6083,
+      "learning_rate": 4.466666666666667e-05,
+      "loss": 1.6078,
       "step": 80
     },
     {
       "epoch": 0.4,
-      "learning_rate": 4.600000000000001e-05,
-      "loss": 1.6063,
+      "learning_rate": 4.3333333333333334e-05,
+      "loss": 1.6084,
       "step": 100
     },
     {
       "epoch": 0.48,
-      "learning_rate": 4.52e-05,
-      "loss": 1.606,
+      "learning_rate": 4.2e-05,
+      "loss": 1.6069,
       "step": 120
     },
     {
       "epoch": 0.56,
-      "learning_rate": 4.44e-05,
-      "loss": 1.6085,
+      "learning_rate": 4.066666666666667e-05,
+      "loss": 1.6077,
       "step": 140
     },
     {
       "epoch": 0.64,
-      "learning_rate": 4.36e-05,
-      "loss": 1.609,
+      "learning_rate": 3.933333333333333e-05,
+      "loss": 1.6025,
       "step": 160
     },
     {
       "epoch": 0.72,
-      "learning_rate": 4.2800000000000004e-05,
-      "loss": 1.6032,
+      "learning_rate": 3.8e-05,
+      "loss": 1.6015,
       "step": 180
     },
     {
       "epoch": 0.8,
-      "learning_rate": 4.2e-05,
-      "loss": 1.609,
+      "learning_rate": 3.6666666666666666e-05,
+      "loss": 1.5974,
       "step": 200
     },
     {
       "epoch": 0.88,
-      "learning_rate": 4.12e-05,
-      "loss": 1.6022,
+      "learning_rate": 3.5333333333333336e-05,
+      "loss": 1.6043,
       "step": 220
     },
     {
       "epoch": 0.96,
-      "learning_rate": 4.0400000000000006e-05,
-      "loss": 1.5993,
+      "learning_rate": 3.4000000000000007e-05,
+      "loss": 1.5896,
       "step": 240
     },
     {
       "epoch": 1.0,
-      "eval_accuracy": 0.27799999713897705,
-      "eval_loss": 1.596663236618042,
-      "eval_runtime": 0.7533,
-      "eval_samples_per_second": 663.789,
-      "eval_steps_per_second": 42.483,
+      "eval_accuracy": 0.22200000286102295,
+      "eval_loss": 1.5897876024246216,
+      "eval_runtime": 0.8298,
+      "eval_samples_per_second": 602.546,
+      "eval_steps_per_second": 38.563,
       "step": 250
     },
     {
       "epoch": 1.04,
-      "learning_rate": 3.960000000000001e-05,
-      "loss": 1.6047,
+      "learning_rate": 3.266666666666667e-05,
+      "loss": 1.5971,
       "step": 260
     },
     {
       "epoch": 1.12,
-      "learning_rate": 3.88e-05,
-      "loss": 1.6005,
+      "learning_rate": 3.1333333333333334e-05,
+      "loss": 1.5927,
       "step": 280
     },
     {
       "epoch": 1.2,
-      "learning_rate": 3.8e-05,
-      "loss": 1.5939,
+      "learning_rate": 3e-05,
+      "loss": 1.5924,
       "step": 300
     },
     {
       "epoch": 1.28,
-      "learning_rate": 3.72e-05,
-      "loss": 1.5916,
+      "learning_rate": 2.8666666666666668e-05,
+      "loss": 1.5821,
       "step": 320
     },
     {
       "epoch": 1.36,
-      "learning_rate": 3.6400000000000004e-05,
-      "loss": 1.5908,
+      "learning_rate": 2.733333333333333e-05,
+      "loss": 1.5791,
       "step": 340
     },
     {
       "epoch": 1.44,
-      "learning_rate": 3.56e-05,
-      "loss": 1.59,
+      "learning_rate": 2.6000000000000002e-05,
+      "loss": 1.5902,
       "step": 360
     },
     {
       "epoch": 1.52,
-      "learning_rate": 3.48e-05,
-      "loss": 1.5996,
+      "learning_rate": 2.466666666666667e-05,
+      "loss": 1.5781,
       "step": 380
     },
     {
       "epoch": 1.6,
-      "learning_rate": 3.4000000000000007e-05,
-      "loss": 1.586,
+      "learning_rate": 2.3333333333333336e-05,
+      "loss": 1.5712,
       "step": 400
     },
     {
       "epoch": 1.68,
-      "learning_rate": 3.32e-05,
-      "loss": 1.5712,
+      "learning_rate": 2.2000000000000003e-05,
+      "loss": 1.5792,
       "step": 420
     },
     {
       "epoch": 1.76,
-      "learning_rate": 3.24e-05,
-      "loss": 1.5726,
+      "learning_rate": 2.0666666666666666e-05,
+      "loss": 1.5764,
       "step": 440
     },
     {
       "epoch": 1.84,
-      "learning_rate": 3.16e-05,
-      "loss": 1.5727,
+      "learning_rate": 1.9333333333333333e-05,
+      "loss": 1.5908,
       "step": 460
     },
     {
       "epoch": 1.92,
-      "learning_rate": 3.08e-05,
-      "loss": 1.5822,
+      "learning_rate": 1.8e-05,
+      "loss": 1.558,
       "step": 480
     },
     {
       "epoch": 2.0,
-      "learning_rate": 3e-05,
-      "loss": 1.5731,
+      "learning_rate": 1.6666666666666667e-05,
+      "loss": 1.5831,
       "step": 500
     },
     {
       "epoch": 2.0,
-      "eval_accuracy": 0.3019999861717224,
-      "eval_loss": 1.5877653360366821,
-      "eval_runtime": 0.647,
-      "eval_samples_per_second": 772.82,
-      "eval_steps_per_second": 49.46,
+      "eval_accuracy": 0.21199999749660492,
+      "eval_loss": 1.5786162614822388,
+      "eval_runtime": 0.8638,
+      "eval_samples_per_second": 578.811,
+      "eval_steps_per_second": 37.044,
       "step": 500
     },
     {
       "epoch": 2.08,
-      "learning_rate": 2.9199999999999998e-05,
-      "loss": 1.5914,
+      "learning_rate": 1.5333333333333334e-05,
+      "loss": 1.5675,
       "step": 520
     },
     {
       "epoch": 2.16,
-      "learning_rate": 2.84e-05,
-      "loss": 1.5627,
+      "learning_rate": 1.4000000000000001e-05,
+      "loss": 1.5696,
       "step": 540
     },
     {
       "epoch": 2.24,
-      "learning_rate": 2.7600000000000003e-05,
-      "loss": 1.5672,
+      "learning_rate": 1.2666666666666668e-05,
+      "loss": 1.5679,
       "step": 560
     },
     {
       "epoch": 2.32,
-      "learning_rate": 2.6800000000000004e-05,
-      "loss": 1.564,
+      "learning_rate": 1.1333333333333334e-05,
+      "loss": 1.5625,
       "step": 580
     },
     {
       "epoch": 2.4,
-      "learning_rate": 2.6000000000000002e-05,
-      "loss": 1.577,
+      "learning_rate": 1e-05,
+      "loss": 1.5726,
       "step": 600
     },
     {
       "epoch": 2.48,
-      "learning_rate": 2.5200000000000003e-05,
-      "loss": 1.5712,
+      "learning_rate": 8.666666666666668e-06,
+      "loss": 1.5627,
       "step": 620
     },
     {
       "epoch": 2.56,
-      "learning_rate": 2.44e-05,
-      "loss": 1.5725,
+      "learning_rate": 7.333333333333334e-06,
+      "loss": 1.5596,
       "step": 640
     },
     {
       "epoch": 2.64,
-      "learning_rate": 2.36e-05,
-      "loss": 1.5532,
+      "learning_rate": 6e-06,
+      "loss": 1.5648,
       "step": 660
     },
     {
       "epoch": 2.72,
-      "learning_rate": 2.2800000000000002e-05,
-      "loss": 1.5703,
+      "learning_rate": 4.666666666666667e-06,
+      "loss": 1.5623,
       "step": 680
     },
     {
       "epoch": 2.8,
-      "learning_rate": 2.2000000000000003e-05,
-      "loss": 1.5803,
+      "learning_rate": 3.3333333333333333e-06,
+      "loss": 1.5766,
       "step": 700
     },
     {
       "epoch": 2.88,
-      "learning_rate": 2.12e-05,
-      "loss": 1.5728,
+      "learning_rate": 2.0000000000000003e-06,
+      "loss": 1.5596,
       "step": 720
     },
     {
       "epoch": 2.96,
-      "learning_rate": 2.04e-05,
-      "loss": 1.5434,
+      "learning_rate": 6.666666666666667e-07,
+      "loss": 1.54,
       "step": 740
     },
     {
       "epoch": 3.0,
-      "eval_accuracy": 0.29600000381469727,
-      "eval_loss": 1.5791035890579224,
-      "eval_runtime": 0.7356,
-      "eval_samples_per_second": 679.717,
-      "eval_steps_per_second": 43.502,
+      "eval_accuracy": 0.2160000056028366,
+      "eval_loss": 1.5779266357421875,
+      "eval_runtime": 0.9285,
+      "eval_samples_per_second": 538.513,
+      "eval_steps_per_second": 34.465,
       "step": 750
     },
     {
-      "epoch": 3.04,
-      "learning_rate": 1.9600000000000002e-05,
-      "loss": 1.5569,
-      "step": 760
-    },
-    {
-      "epoch": 3.12,
-      "learning_rate": 1.88e-05,
-      "loss": 1.5382,
-      "step": 780
-    },
-    {
-      "epoch": 3.2,
-      "learning_rate": 1.8e-05,
-      "loss": 1.5478,
-      "step": 800
-    },
-    {
-      "epoch": 3.28,
-      "learning_rate": 1.7199999999999998e-05,
-      "loss": 1.5665,
-      "step": 820
-    },
-    {
-      "epoch": 3.36,
-      "learning_rate": 1.6400000000000002e-05,
-      "loss": 1.5494,
-      "step": 840
-    },
-    {
-      "epoch": 3.44,
-      "learning_rate": 1.56e-05,
-      "loss": 1.5489,
-      "step": 860
-    },
-    {
-      "epoch": 3.52,
-      "learning_rate": 1.48e-05,
-      "loss": 1.5483,
-      "step": 880
-    },
-    {
-      "epoch": 3.6,
-      "learning_rate": 1.4000000000000001e-05,
-      "loss": 1.5388,
-      "step": 900
-    },
-    {
-      "epoch": 3.68,
-      "learning_rate": 1.32e-05,
-      "loss": 1.5563,
-      "step": 920
-    },
-    {
-      "epoch": 3.76,
-      "learning_rate": 1.24e-05,
-      "loss": 1.5616,
-      "step": 940
-    },
-    {
-      "epoch": 3.84,
-      "learning_rate": 1.16e-05,
-      "loss": 1.5512,
-      "step": 960
-    },
-    {
-      "epoch": 3.92,
-      "learning_rate": 1.08e-05,
-      "loss": 1.5288,
-      "step": 980
-    },
-    {
-      "epoch": 4.0,
-      "learning_rate": 1e-05,
-      "loss": 1.5473,
-      "step": 1000
-    },
-    {
-      "epoch": 4.0,
-      "eval_accuracy": 0.2939999997615814,
-      "eval_loss": 1.5701889991760254,
-      "eval_runtime": 0.7589,
-      "eval_samples_per_second": 658.833,
-      "eval_steps_per_second": 42.165,
-      "step": 1000
-    },
-    {
-      "epoch": 4.08,
-      "learning_rate": 9.2e-06,
-      "loss": 1.5642,
-      "step": 1020
-    },
-    {
-      "epoch": 4.16,
-      "learning_rate": 8.400000000000001e-06,
-      "loss": 1.5334,
-      "step": 1040
-    },
-    {
-      "epoch": 4.24,
-      "learning_rate": 7.6e-06,
-      "loss": 1.526,
-      "step": 1060
-    },
-    {
-      "epoch": 4.32,
-      "learning_rate": 6.800000000000001e-06,
-      "loss": 1.5361,
-      "step": 1080
-    },
-    {
-      "epoch": 4.4,
-      "learning_rate": 6e-06,
-      "loss": 1.573,
-      "step": 1100
-    },
-    {
-      "epoch": 4.48,
-      "learning_rate": 5.2e-06,
-      "loss": 1.5543,
-      "step": 1120
-    },
-    {
-      "epoch": 4.56,
-      "learning_rate": 4.4e-06,
-      "loss": 1.5193,
-      "step": 1140
-    },
-    {
-      "epoch": 4.64,
-      "learning_rate": 3.6e-06,
-      "loss": 1.5401,
-      "step": 1160
-    },
-    {
-      "epoch": 4.72,
-      "learning_rate": 2.8000000000000003e-06,
-      "loss": 1.5373,
-      "step": 1180
-    },
-    {
-      "epoch": 4.8,
-      "learning_rate": 2.0000000000000003e-06,
-      "loss": 1.5279,
-      "step": 1200
-    },
-    {
-      "epoch": 4.88,
-      "learning_rate": 1.2000000000000002e-06,
-      "loss": 1.5503,
-      "step": 1220
-    },
-    {
-      "epoch": 4.96,
-      "learning_rate": 4.0000000000000003e-07,
-      "loss": 1.5422,
-      "step": 1240
-    },
-    {
-      "epoch": 5.0,
-      "eval_accuracy": 0.2919999957084656,
-      "eval_loss": 1.569107174873352,
-      "eval_runtime": 0.7454,
-      "eval_samples_per_second": 670.819,
-      "eval_steps_per_second": 42.932,
-      "step": 1250
-    },
-    {
-      "epoch": 5.0,
-      "step": 1250,
-      "total_flos": 10492735676640.0,
-      "train_loss": 1.5705763679504394,
-      "train_runtime": 78.9966,
-      "train_samples_per_second": 253.176,
-      "train_steps_per_second": 15.823
+      "epoch": 3.0,
+      "step": 750,
+      "total_flos": 16105346090400.0,
+      "train_loss": 1.5832408892313639,
+      "train_runtime": 76.8946,
+      "train_samples_per_second": 156.058,
+      "train_steps_per_second": 9.754
     }
   ],
-  "max_steps": 1250,
-  "num_train_epochs": 5,
-  "total_flos": 10492735676640.0,
+  "max_steps": 750,
+  "num_train_epochs": 3,
+  "total_flos": 16105346090400.0,
   "trial_name": null,
   "trial_params": null
 }

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:f1eb0324ba0a73ee29fb2559c05fad4e6c64a97f34e7bb4de884b1c0c9cc415f
+oid sha256:cdb89f3c4237c498a27550264d267734eb7e511c05c2d3aa52333e0f2f6ca78c
 size 2799