End of training

all_results.json

@@ -1,18 +1,18 @@
 {
     "epoch": 3.0,
-    "eval_accuracy": 0.9702625751344511,
-    "eval_f1": 0.9683927370544722,
-    "eval_loss": 0.11914637684822083,
-    "eval_precision": 0.9769335142469471,
-    "eval_recall": 0.96,
-    "eval_roc_auc": 0.9697652016857315,
-    "eval_runtime": 9.4949,
-    "eval_samples": 6322,
-    "eval_samples_per_second": 665.828,
-    "eval_steps_per_second": 5.266,
-    "train_loss": 0.11545699092274161,
-    "train_runtime": 488.0473,
-    "train_samples": 25287,
-    "train_samples_per_second": 155.438,
-    "train_steps_per_second": 1.217
+    "eval_accuracy": 0.9273952095808383,
+    "eval_f1": 0.9438332368268674,
+    "eval_loss": 0.23274052143096924,
+    "eval_precision": 0.9690844233055886,
+    "eval_recall": 0.9198645598194131,
+    "eval_roc_auc": 0.9310433910208176,
+    "eval_runtime": 1.9896,
+    "eval_samples": 1336,
+    "eval_samples_per_second": 671.502,
+    "eval_steps_per_second": 5.529,
+    "train_loss": 0.19368775685628256,
+    "train_runtime": 134.498,
+    "train_samples": 5342,
+    "train_samples_per_second": 119.154,
+    "train_steps_per_second": 0.937
 }

eval_results.json

@@ -1,13 +1,13 @@
 {
     "epoch": 3.0,
-    "eval_accuracy": 0.9702625751344511,
-    "eval_f1": 0.9683927370544722,
-    "eval_loss": 0.11914637684822083,
-    "eval_precision": 0.9769335142469471,
-    "eval_recall": 0.96,
-    "eval_roc_auc": 0.9697652016857315,
-    "eval_runtime": 9.4949,
-    "eval_samples": 6322,
-    "eval_samples_per_second": 665.828,
-    "eval_steps_per_second": 5.266
+    "eval_accuracy": 0.9273952095808383,
+    "eval_f1": 0.9438332368268674,
+    "eval_loss": 0.23274052143096924,
+    "eval_precision": 0.9690844233055886,
+    "eval_recall": 0.9198645598194131,
+    "eval_roc_auc": 0.9310433910208176,
+    "eval_runtime": 1.9896,
+    "eval_samples": 1336,
+    "eval_samples_per_second": 671.502,
+    "eval_steps_per_second": 5.529
 }

train_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 3.0,
-    "train_loss": 0.11545699092274161,
-    "train_runtime": 488.0473,
-    "train_samples": 25287,
-    "train_samples_per_second": 155.438,
-    "train_steps_per_second": 1.217
+    "train_loss": 0.19368775685628256,
+    "train_runtime": 134.498,
+    "train_samples": 5342,
+    "train_samples_per_second": 119.154,
+    "train_steps_per_second": 0.937
 }

trainer_state.json

@@ -2,195 +2,43 @@
   "best_metric": null,
   "best_model_checkpoint": null,
   "epoch": 3.0,
-  "global_step": 594,
+  "global_step": 126,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
     {
-      "epoch": 0.33,
-      "learning_rate": 8.905723905723906e-05,
-      "loss": 0.3716,
+      "epoch": 1.57,
+      "learning_rate": 4.9206349206349204e-05,
+      "loss": 0.2872,
       "step": 66
     },
     {
-      "epoch": 0.33,
-      "eval_accuracy": 0.9296108826320785,
-      "eval_f1": 0.9248945147679325,
-      "eval_loss": 0.20176392793655396,
-      "eval_precision": 0.9367521367521368,
-      "eval_recall": 0.9133333333333333,
-      "eval_roc_auc": 0.9288219947822597,
-      "eval_runtime": 9.531,
-      "eval_samples_per_second": 663.307,
-      "eval_steps_per_second": 5.246,
+      "epoch": 1.57,
+      "eval_accuracy": 0.9176646706586826,
+      "eval_f1": 0.9374999999999999,
+      "eval_loss": 0.20166487991809845,
+      "eval_precision": 0.9439359267734554,
+      "eval_recall": 0.9311512415349887,
+      "eval_roc_auc": 0.9111311763230499,
+      "eval_runtime": 1.9771,
+      "eval_samples_per_second": 675.722,
+      "eval_steps_per_second": 5.564,
       "step": 66
     },
-    {
-      "epoch": 0.67,
-      "learning_rate": 7.794612794612795e-05,
-      "loss": 0.1745,
-      "step": 132
-    },
-    {
-      "epoch": 0.67,
-      "eval_accuracy": 0.9533375514077823,
-      "eval_f1": 0.9497530233350365,
-      "eval_loss": 0.14681555330753326,
-      "eval_precision": 0.9710902124695228,
-      "eval_recall": 0.9293333333333333,
-      "eval_roc_auc": 0.9521741922536624,
-      "eval_runtime": 9.5186,
-      "eval_samples_per_second": 664.175,
-      "eval_steps_per_second": 5.253,
-      "step": 132
-    },
-    {
-      "epoch": 1.0,
-      "learning_rate": 6.683501683501684e-05,
-      "loss": 0.1346,
-      "step": 198
-    },
-    {
-      "epoch": 1.0,
-      "eval_accuracy": 0.9656754191711484,
-      "eval_f1": 0.9633631605605266,
-      "eval_loss": 0.10907050222158432,
-      "eval_precision": 0.976052001368457,
-      "eval_recall": 0.951,
-      "eval_roc_auc": 0.9649641782059001,
-      "eval_runtime": 9.5208,
-      "eval_samples_per_second": 664.018,
-      "eval_steps_per_second": 5.252,
-      "step": 198
-    },
-    {
-      "epoch": 1.33,
-      "learning_rate": 5.572390572390572e-05,
-      "loss": 0.0917,
-      "step": 264
-    },
-    {
-      "epoch": 1.33,
-      "eval_accuracy": 0.9647263524201202,
-      "eval_f1": 0.9621969825394135,
-      "eval_loss": 0.12939223647117615,
-      "eval_precision": 0.9789582614694722,
-      "eval_recall": 0.946,
-      "eval_roc_auc": 0.9638187838651415,
-      "eval_runtime": 9.5288,
-      "eval_samples_per_second": 663.461,
-      "eval_steps_per_second": 5.247,
-      "step": 264
-    },
-    {
-      "epoch": 1.67,
-      "learning_rate": 4.4612794612794616e-05,
-      "loss": 0.0877,
-      "step": 330
-    },
-    {
-      "epoch": 1.67,
-      "eval_accuracy": 0.9667826637140146,
-      "eval_f1": 0.9651162790697674,
-      "eval_loss": 0.10901537537574768,
-      "eval_precision": 0.9619205298013245,
-      "eval_recall": 0.9683333333333334,
-      "eval_roc_auc": 0.9668578165763597,
-      "eval_runtime": 9.5197,
-      "eval_samples_per_second": 664.094,
-      "eval_steps_per_second": 5.252,
-      "step": 330
-    },
-    {
-      "epoch": 2.0,
-      "learning_rate": 3.35016835016835e-05,
-      "loss": 0.0731,
-      "step": 396
-    },
-    {
-      "epoch": 2.0,
-      "eval_accuracy": 0.9688389750079089,
-      "eval_f1": 0.9669074416260709,
-      "eval_loss": 0.10423740744590759,
-      "eval_precision": 0.9746020995597697,
-      "eval_recall": 0.9593333333333334,
-      "eval_roc_auc": 0.9683782861729882,
-      "eval_runtime": 9.5669,
-      "eval_samples_per_second": 660.818,
-      "eval_steps_per_second": 5.226,
-      "step": 396
-    },
-    {
-      "epoch": 2.33,
-      "learning_rate": 2.2390572390572393e-05,
-      "loss": 0.0342,
-      "step": 462
-    },
-    {
-      "epoch": 2.33,
-      "eval_accuracy": 0.969155330591585,
-      "eval_f1": 0.9674620390455532,
-      "eval_loss": 0.12906306982040405,
-      "eval_precision": 0.9685933845639826,
-      "eval_recall": 0.9663333333333334,
-      "eval_roc_auc": 0.9690185631145897,
-      "eval_runtime": 9.5007,
-      "eval_samples_per_second": 665.425,
-      "eval_steps_per_second": 5.263,
-      "step": 462
-    },
-    {
-      "epoch": 2.67,
-      "learning_rate": 1.127946127946128e-05,
-      "loss": 0.0375,
-      "step": 528
-    },
-    {
-      "epoch": 2.67,
-      "eval_accuracy": 0.9705789307181272,
-      "eval_f1": 0.9687919463087248,
-      "eval_loss": 0.12022514641284943,
-      "eval_precision": 0.9753378378378378,
-      "eval_recall": 0.9623333333333334,
-      "eval_roc_auc": 0.9701793096528197,
-      "eval_runtime": 9.5237,
-      "eval_samples_per_second": 663.816,
-      "eval_steps_per_second": 5.25,
-      "step": 528
-    },
-    {
-      "epoch": 3.0,
-      "learning_rate": 1.6835016835016834e-07,
-      "loss": 0.0342,
-      "step": 594
-    },
-    {
-      "epoch": 3.0,
-      "eval_accuracy": 0.9702625751344511,
-      "eval_f1": 0.9683927370544722,
-      "eval_loss": 0.11914637684822083,
-      "eval_precision": 0.9769335142469471,
-      "eval_recall": 0.96,
-      "eval_roc_auc": 0.9697652016857315,
-      "eval_runtime": 9.5437,
-      "eval_samples_per_second": 662.428,
-      "eval_steps_per_second": 5.239,
-      "step": 594
-    },
     {
       "epoch": 3.0,
-      "step": 594,
-      "total_flos": 9979933885332480.0,
-      "train_loss": 0.11545699092274161,
-      "train_runtime": 488.0473,
-      "train_samples_per_second": 155.438,
-      "train_steps_per_second": 1.217
+      "step": 126,
+      "total_flos": 2108308886599680.0,
+      "train_loss": 0.19368775685628256,
+      "train_runtime": 134.498,
+      "train_samples_per_second": 119.154,
+      "train_steps_per_second": 0.937
     }
   ],
-  "max_steps": 594,
+  "max_steps": 126,
   "num_train_epochs": 3,
-  "total_flos": 9979933885332480.0,
+  "total_flos": 2108308886599680.0,
   "trial_name": null,
   "trial_params": null
 }