End of training

README.md

@@ -1,4 +1,6 @@
 ---
+language:
+- id
 license: mit
 base_model: indolem/indobert-base-uncased
 tags:

all_results.json

@@ -1,21 +1,21 @@
 {
     "accuracy": 0.9203187250996016,
     "epoch": 20.0,
-    "eval_accuracy": 0.8822055137844611,
-    "eval_f1": 0.858259325044405,
-    "eval_loss": 0.30716511607170105,
-    "eval_precision": 0.8573798178418481,
-    "eval_recall": 0.8591562102200401,
-    "eval_runtime": 5.0211,
+    "eval_accuracy": 0.8796992481203008,
+    "eval_f1": 0.8563025210084034,
+    "eval_loss": 0.31680676341056824,
+    "eval_precision": 0.8529936381473334,
+    "eval_recall": 0.8598836152027641,
+    "eval_runtime": 1.8096,
     "eval_samples": 399,
-    "eval_samples_per_second": 79.464,
-    "eval_steps_per_second": 9.958,
-    "f1": 0.9039946451196481,
-    "precision": 0.9032008703618873,
-    "recall": 0.9048002299511353,
-    "train_loss": 0.22629446436147221,
-    "train_runtime": 1956.0503,
+    "eval_samples_per_second": 220.495,
+    "eval_steps_per_second": 27.631,
+    "f1": 0.9045572941422515,
+    "precision": 0.901511292239769,
+    "recall": 0.9077895947111239,
+    "train_loss": 0.2318932650519199,
+    "train_runtime": 643.2373,
     "train_samples": 3645,
-    "train_samples_per_second": 37.269,
-    "train_steps_per_second": 1.247
+    "train_samples_per_second": 113.333,
+    "train_steps_per_second": 3.793
 }

eval_results.json

@@ -1,12 +1,12 @@
 {
     "epoch": 20.0,
-    "eval_accuracy": 0.8822055137844611,
-    "eval_f1": 0.858259325044405,
-    "eval_loss": 0.30716511607170105,
-    "eval_precision": 0.8573798178418481,
-    "eval_recall": 0.8591562102200401,
-    "eval_runtime": 5.0211,
+    "eval_accuracy": 0.8796992481203008,
+    "eval_f1": 0.8563025210084034,
+    "eval_loss": 0.31680676341056824,
+    "eval_precision": 0.8529936381473334,
+    "eval_recall": 0.8598836152027641,
+    "eval_runtime": 1.8096,
     "eval_samples": 399,
-    "eval_samples_per_second": 79.464,
-    "eval_steps_per_second": 9.958
+    "eval_samples_per_second": 220.495,
+    "eval_steps_per_second": 27.631
 }

predict_results.json

@@ -1,6 +1,6 @@
 {
     "accuracy": 0.9203187250996016,
-    "f1": 0.9039946451196481,
-    "precision": 0.9032008703618873,
-    "recall": 0.9048002299511353
+    "f1": 0.9045572941422515,
+    "precision": 0.901511292239769,
+    "recall": 0.9077895947111239
 }

predict_results.txt

@@ -18,7 +18,7 @@ index	prediction
 16	1
 17	1
 18	1
-19	0
+19	1
 20	1
 21	0
 22	1
@@ -39,9 +39,9 @@ index	prediction
 37	1
 38	0
 39	1
-40	0
+40	1
 41	1
-42	1
+42	0
 43	1
 44	1
 45	1
@@ -55,7 +55,7 @@ index	prediction
 53	1
 54	1
 55	1
-56	1
+56	0
 57	1
 58	1
 59	1
@@ -70,17 +70,17 @@ index	prediction
 68	0
 69	1
 70	1
-71	1
+71	0
 72	1
 73	1
 74	1
-75	1
+75	0
 76	1
 77	1
 78	0
 79	1
 80	1
-81	1
+81	0
 82	1
 83	1
 84	1
@@ -99,7 +99,7 @@ index	prediction
 97	1
 98	1
 99	0
-100	0
+100	1
 101	1
 102	1
 103	1
@@ -113,7 +113,7 @@ index	prediction
 111	1
 112	1
 113	1
-114	0
+114	1
 115	1
 116	1
 117	1
@@ -169,7 +169,7 @@ index	prediction
 167	1
 168	1
 169	1
-170	0
+170	1
 171	1
 172	1
 173	1
@@ -215,7 +215,7 @@ index	prediction
 213	1
 214	1
 215	1
-216	0
+216	1
 217	1
 218	1
 219	1
@@ -258,8 +258,8 @@ index	prediction
 256	1
 257	1
 258	1
-259	0
-260	1
+259	1
+260	0
 261	1
 262	1
 263	1
@@ -286,11 +286,11 @@ index	prediction
 284	1
 285	1
 286	1
-287	0
+287	1
 288	1
 289	0
 290	1
-291	0
+291	1
 292	1
 293	1
 294	0
@@ -359,7 +359,7 @@ index	prediction
 357	0
 358	0
 359	0
-360	0
+360	1
 361	0
 362	0
 363	0
@@ -372,8 +372,8 @@ index	prediction
 370	0
 371	0
 372	0
-373	0
-374	0
+373	1
+374	1
 375	0
 376	0
 377	0
@@ -383,7 +383,7 @@ index	prediction
 381	0
 382	0
 383	0
-384	1
+384	0
 385	1
 386	0
 387	0
@@ -418,7 +418,7 @@ index	prediction
 416	0
 417	0
 418	0
-419	1
+419	0
 420	0
 421	0
 422	0
@@ -436,7 +436,7 @@ index	prediction
 434	0
 435	0
 436	0
-437	0
+437	1
 438	0
 439	0
 440	0
@@ -485,7 +485,7 @@ index	prediction
 483	0
 484	0
 485	0
-486	1
+486	0
 487	0
 488	0
 489	0
@@ -536,7 +536,7 @@ index	prediction
 534	0
 535	0
 536	0
-537	1
+537	0
 538	0
 539	0
 540	0
@@ -569,7 +569,7 @@ index	prediction
 567	0
 568	0
 569	0
-570	1
+570	0
 571	0
 572	0
 573	0
@@ -585,13 +585,13 @@ index	prediction
 583	0
 584	0
 585	0
-586	0
+586	1
 587	0
 588	0
 589	0
 590	0
 591	0
-592	0
+592	1
 593	0
 594	0
 595	0
@@ -626,7 +626,7 @@ index	prediction
 624	0
 625	0
 626	0
-627	0
+627	1
 628	0
 629	0
 630	0
@@ -656,7 +656,7 @@ index	prediction
 654	0
 655	0
 656	0
-657	1
+657	0
 658	1
 659	0
 660	0
@@ -672,7 +672,7 @@ index	prediction
 670	0
 671	0
 672	0
-673	0
+673	1
 674	0
 675	0
 676	0
@@ -681,7 +681,7 @@ index	prediction
 679	0
 680	0
 681	0
-682	1
+682	0
 683	0
 684	0
 685	0
@@ -716,9 +716,9 @@ index	prediction
 714	0
 715	0
 716	0
-717	0
+717	1
 718	0
-719	0
+719	1
 720	0
 721	0
 722	0
@@ -739,7 +739,7 @@ index	prediction
 737	0
 738	0
 739	0
-740	0
+740	1
 741	0
 742	0
 743	0
@@ -753,7 +753,7 @@ index	prediction
 751	0
 752	0
 753	0
-754	0
+754	1
 755	0
 756	0
 757	0
@@ -773,7 +773,7 @@ index	prediction
 771	0
 772	0
 773	0
-774	1
+774	0
 775	0
 776	0
 777	0
@@ -811,7 +811,7 @@ index	prediction
 809	0
 810	0
 811	0
-812	1
+812	0
 813	0
 814	0
 815	0
@@ -824,7 +824,7 @@ index	prediction
 822	0
 823	0
 824	0
-825	0
+825	1
 826	0
 827	0
 828	0
@@ -851,7 +851,7 @@ index	prediction
 849	0
 850	0
 851	0
-852	1
+852	0
 853	0
 854	0
 855	0

runs/Jun03_15-23-46_a358b85c7679/events.out.tfevents.1717428885.a358b85c7679.170314.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:230b1beb8a7077d9c3359fe928ab04829edae8c1bf07d6d7eb58b5ad4d3c671e
+size 560

train_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 20.0,
-    "train_loss": 0.22629446436147221,
-    "train_runtime": 1956.0503,
+    "train_loss": 0.2318932650519199,
+    "train_runtime": 643.2373,
     "train_samples": 3645,
-    "train_samples_per_second": 37.269,
-    "train_steps_per_second": 1.247
+    "train_samples_per_second": 113.333,
+    "train_steps_per_second": 3.793
 }

trainer_state.json

@@ -10,392 +10,392 @@
   "log_history": [
     {
       "epoch": 1.0,
-      "grad_norm": 3.569611072540283,
+      "grad_norm": 3.639183521270752,
       "learning_rate": 4.75e-05,
-      "loss": 0.5438,
+      "loss": 0.5413,
       "step": 122
     },
     {
       "epoch": 1.0,
-      "eval_accuracy": 0.7218045112781954,
-      "eval_f1": 0.6545993371027491,
-      "eval_loss": 0.49882158637046814,
-      "eval_precision": 0.6600553802562947,
-      "eval_recall": 0.6506637570467357,
-      "eval_runtime": 5.2428,
-      "eval_samples_per_second": 76.104,
-      "eval_steps_per_second": 9.537,
+      "eval_accuracy": 0.7243107769423559,
+      "eval_f1": 0.6445578231292517,
+      "eval_loss": 0.5006802678108215,
+      "eval_precision": 0.6593400801180687,
+      "eval_recall": 0.6374340789234406,
+      "eval_runtime": 1.8003,
+      "eval_samples_per_second": 221.628,
+      "eval_steps_per_second": 27.773,
       "step": 122
     },
     {
       "epoch": 2.0,
-      "grad_norm": 3.2452878952026367,
+      "grad_norm": 3.4983344078063965,
       "learning_rate": 4.5e-05,
-      "loss": 0.4428,
+      "loss": 0.4584,
       "step": 244
     },
     {
       "epoch": 2.0,
-      "eval_accuracy": 0.8446115288220551,
-      "eval_f1": 0.8161454307628278,
-      "eval_loss": 0.3788329064846039,
-      "eval_precision": 0.8107299270072992,
-      "eval_recall": 0.8225586470267321,
-      "eval_runtime": 5.1661,
-      "eval_samples_per_second": 77.234,
-      "eval_steps_per_second": 9.679,
+      "eval_accuracy": 0.8295739348370927,
+      "eval_f1": 0.7760942760942761,
+      "eval_loss": 0.3855762183666229,
+      "eval_precision": 0.8122789566755084,
+      "eval_recall": 0.7569103473358793,
+      "eval_runtime": 1.8373,
+      "eval_samples_per_second": 217.166,
+      "eval_steps_per_second": 27.214,
       "step": 244
     },
     {
       "epoch": 3.0,
-      "grad_norm": 3.797173500061035,
+      "grad_norm": 5.97755765914917,
       "learning_rate": 4.25e-05,
-      "loss": 0.3441,
+      "loss": 0.3559,
       "step": 366
     },
     {
       "epoch": 3.0,
-      "eval_accuracy": 0.8596491228070176,
-      "eval_f1": 0.8179269882659713,
-      "eval_loss": 0.3289283514022827,
-      "eval_precision": 0.8510239760239761,
-      "eval_recall": 0.7981905801054737,
-      "eval_runtime": 5.1072,
-      "eval_samples_per_second": 78.125,
-      "eval_steps_per_second": 9.79,
+      "eval_accuracy": 0.8571428571428571,
+      "eval_f1": 0.8079089841803424,
+      "eval_loss": 0.3407064378261566,
+      "eval_precision": 0.8638322884012539,
+      "eval_recall": 0.7814148026913984,
+      "eval_runtime": 1.8397,
+      "eval_samples_per_second": 216.877,
+      "eval_steps_per_second": 27.178,
       "step": 366
     },
     {
       "epoch": 4.0,
-      "grad_norm": 1.9981327056884766,
+      "grad_norm": 2.0363476276397705,
       "learning_rate": 4e-05,
-      "loss": 0.2986,
+      "loss": 0.2961,
       "step": 488
     },
     {
       "epoch": 4.0,
-      "eval_accuracy": 0.8796992481203008,
-      "eval_f1": 0.8533986527862829,
-      "eval_loss": 0.28838610649108887,
-      "eval_precision": 0.8572003218020917,
-      "eval_recall": 0.8498817966903074,
-      "eval_runtime": 5.1209,
-      "eval_samples_per_second": 77.916,
-      "eval_steps_per_second": 9.764,
+      "eval_accuracy": 0.8696741854636592,
+      "eval_f1": 0.8419946387230413,
+      "eval_loss": 0.3088829517364502,
+      "eval_precision": 0.8437691365584814,
+      "eval_recall": 0.8402891434806329,
+      "eval_runtime": 1.8466,
+      "eval_samples_per_second": 216.069,
+      "eval_steps_per_second": 27.076,
       "step": 488
     },
     {
       "epoch": 5.0,
-      "grad_norm": 1.763756513595581,
+      "grad_norm": 2.3318538665771484,
       "learning_rate": 3.7500000000000003e-05,
-      "loss": 0.2667,
+      "loss": 0.276,
       "step": 610
     },
     {
       "epoch": 5.0,
-      "eval_accuracy": 0.8771929824561403,
-      "eval_f1": 0.850729517396184,
-      "eval_loss": 0.26981133222579956,
-      "eval_precision": 0.8535087719298247,
-      "eval_recall": 0.8481087470449173,
-      "eval_runtime": 5.1232,
-      "eval_samples_per_second": 77.881,
-      "eval_steps_per_second": 9.759,
+      "eval_accuracy": 0.8621553884711779,
+      "eval_f1": 0.8365204824303285,
+      "eval_loss": 0.29173794388771057,
+      "eval_precision": 0.8313636363636363,
+      "eval_recall": 0.8424713584288053,
+      "eval_runtime": 1.8414,
+      "eval_samples_per_second": 216.677,
+      "eval_steps_per_second": 27.153,
       "step": 610
     },
     {
       "epoch": 6.0,
-      "grad_norm": 2.7370102405548096,
+      "grad_norm": 2.3067033290863037,
       "learning_rate": 3.5e-05,
-      "loss": 0.2524,
+      "loss": 0.2555,
       "step": 732
     },
     {
       "epoch": 6.0,
-      "eval_accuracy": 0.8847117794486216,
-      "eval_f1": 0.8609292598654301,
-      "eval_loss": 0.27233511209487915,
-      "eval_precision": 0.8609292598654301,
-      "eval_recall": 0.8609292598654301,
-      "eval_runtime": 5.1586,
-      "eval_samples_per_second": 77.347,
-      "eval_steps_per_second": 9.693,
+      "eval_accuracy": 0.8696741854636592,
+      "eval_f1": 0.842789598108747,
+      "eval_loss": 0.29054704308509827,
+      "eval_precision": 0.842789598108747,
+      "eval_recall": 0.842789598108747,
+      "eval_runtime": 1.847,
+      "eval_samples_per_second": 216.025,
+      "eval_steps_per_second": 27.071,
       "step": 732
     },
     {
       "epoch": 7.0,
-      "grad_norm": 1.2413272857666016,
+      "grad_norm": 1.5437530279159546,
       "learning_rate": 3.2500000000000004e-05,
-      "loss": 0.2343,
+      "loss": 0.2427,
       "step": 854
     },
     {
       "epoch": 7.0,
-      "eval_accuracy": 0.8646616541353384,
-      "eval_f1": 0.8265664241097301,
-      "eval_loss": 0.3179858326911926,
-      "eval_precision": 0.8532894736842105,
-      "eval_recall": 0.8092380432805966,
-      "eval_runtime": 5.1498,
-      "eval_samples_per_second": 77.479,
-      "eval_steps_per_second": 9.709,
+      "eval_accuracy": 0.8771929824561403,
+      "eval_f1": 0.84402249790578,
+      "eval_loss": 0.30313801765441895,
+      "eval_precision": 0.8669909824394875,
+      "eval_recall": 0.8281051100200036,
+      "eval_runtime": 1.8479,
+      "eval_samples_per_second": 215.922,
+      "eval_steps_per_second": 27.058,
       "step": 854
     },
     {
       "epoch": 8.0,
-      "grad_norm": 3.469871997833252,
+      "grad_norm": 3.033709764480591,
       "learning_rate": 3e-05,
-      "loss": 0.2212,
+      "loss": 0.2219,
       "step": 976
     },
     {
       "epoch": 8.0,
-      "eval_accuracy": 0.8822055137844611,
-      "eval_f1": 0.852937255424767,
-      "eval_loss": 0.29489362239837646,
-      "eval_precision": 0.8674217731421121,
-      "eval_recall": 0.8416530278232406,
-      "eval_runtime": 5.155,
-      "eval_samples_per_second": 77.4,
-      "eval_steps_per_second": 9.699,
+      "eval_accuracy": 0.8771929824561403,
+      "eval_f1": 0.8522278069611882,
+      "eval_loss": 0.2907596528530121,
+      "eval_precision": 0.8513631702756499,
+      "eval_recall": 0.8531096563011457,
+      "eval_runtime": 1.8468,
+      "eval_samples_per_second": 216.053,
+      "eval_steps_per_second": 27.074,
       "step": 976
     },
     {
       "epoch": 9.0,
-      "grad_norm": 4.085997581481934,
+      "grad_norm": 2.670888900756836,
       "learning_rate": 2.7500000000000004e-05,
-      "loss": 0.2142,
+      "loss": 0.2158,
       "step": 1098
     },
     {
       "epoch": 9.0,
       "eval_accuracy": 0.8847117794486216,
-      "eval_f1": 0.8564658408408408,
-      "eval_loss": 0.2828481197357178,
-      "eval_precision": 0.8697278911564625,
-      "eval_recall": 0.8459265320967448,
-      "eval_runtime": 5.1873,
-      "eval_samples_per_second": 76.919,
-      "eval_steps_per_second": 9.639,
+      "eval_f1": 0.8539996181748759,
+      "eval_loss": 0.3083769381046295,
+      "eval_precision": 0.8759595959595959,
+      "eval_recall": 0.8384251682124022,
+      "eval_runtime": 1.856,
+      "eval_samples_per_second": 214.979,
+      "eval_steps_per_second": 26.94,
       "step": 1098
     },
     {
       "epoch": 10.0,
-      "grad_norm": 3.2397539615631104,
+      "grad_norm": 3.6058239936828613,
       "learning_rate": 2.5e-05,
-      "loss": 0.1958,
+      "loss": 0.2,
       "step": 1220
     },
     {
       "epoch": 10.0,
-      "eval_accuracy": 0.8696741854636592,
-      "eval_f1": 0.8457993935430168,
-      "eval_loss": 0.28871509432792664,
-      "eval_precision": 0.8398540145985401,
-      "eval_recall": 0.8527914166212038,
-      "eval_runtime": 5.1373,
-      "eval_samples_per_second": 77.667,
-      "eval_steps_per_second": 9.733,
+      "eval_accuracy": 0.87468671679198,
+      "eval_f1": 0.8517301860990547,
+      "eval_loss": 0.29381993412971497,
+      "eval_precision": 0.8456788321167883,
+      "eval_recall": 0.8588379705400981,
+      "eval_runtime": 1.8468,
+      "eval_samples_per_second": 216.045,
+      "eval_steps_per_second": 27.073,
       "step": 1220
     },
     {
       "epoch": 11.0,
-      "grad_norm": 2.4002835750579834,
+      "grad_norm": 3.181007146835327,
       "learning_rate": 2.25e-05,
-      "loss": 0.1855,
+      "loss": 0.1885,
       "step": 1342
     },
     {
       "epoch": 11.0,
-      "eval_accuracy": 0.8822055137844611,
-      "eval_f1": 0.8602993213495533,
-      "eval_loss": 0.2867955267429352,
-      "eval_precision": 0.8547653958944281,
-      "eval_recall": 0.8666575741043827,
-      "eval_runtime": 5.1317,
-      "eval_samples_per_second": 77.752,
-      "eval_steps_per_second": 9.743,
+      "eval_accuracy": 0.8771929824561403,
+      "eval_f1": 0.8514869535493182,
+      "eval_loss": 0.2976568639278412,
+      "eval_precision": 0.8523821128305106,
+      "eval_recall": 0.8506092016730314,
+      "eval_runtime": 1.8451,
+      "eval_samples_per_second": 216.253,
+      "eval_steps_per_second": 27.099,
       "step": 1342
     },
     {
       "epoch": 12.0,
-      "grad_norm": 1.633034348487854,
+      "grad_norm": 3.1322133541107178,
       "learning_rate": 2e-05,
-      "loss": 0.1742,
+      "loss": 0.183,
       "step": 1464
     },
     {
       "epoch": 12.0,
-      "eval_accuracy": 0.87468671679198,
-      "eval_f1": 0.8448388501742161,
-      "eval_loss": 0.29811105132102966,
-      "eval_precision": 0.8551721930610677,
-      "eval_recall": 0.8363338788870704,
-      "eval_runtime": 5.1494,
-      "eval_samples_per_second": 77.484,
-      "eval_steps_per_second": 9.71,
+      "eval_accuracy": 0.8847117794486216,
+      "eval_f1": 0.8556621579112929,
+      "eval_loss": 0.30698344111442566,
+      "eval_precision": 0.871654421411703,
+      "eval_recall": 0.8434260774686306,
+      "eval_runtime": 1.8533,
+      "eval_samples_per_second": 215.295,
+      "eval_steps_per_second": 26.979,
       "step": 1464
     },
     {
       "epoch": 13.0,
-      "grad_norm": 2.338294506072998,
+      "grad_norm": 1.6321003437042236,
       "learning_rate": 1.75e-05,
-      "loss": 0.1601,
+      "loss": 0.1752,
       "step": 1586
     },
     {
       "epoch": 13.0,
       "eval_accuracy": 0.8796992481203008,
-      "eval_f1": 0.8556004584112431,
-      "eval_loss": 0.29304954409599304,
-      "eval_precision": 0.8538865546218487,
-      "eval_recall": 0.85738316057465,
-      "eval_runtime": 5.0982,
-      "eval_samples_per_second": 78.263,
-      "eval_steps_per_second": 9.807,
+      "eval_f1": 0.8569892473118279,
+      "eval_loss": 0.29585033655166626,
+      "eval_precision": 0.8522004241781549,
+      "eval_recall": 0.8623840698308783,
+      "eval_runtime": 1.846,
+      "eval_samples_per_second": 216.146,
+      "eval_steps_per_second": 27.086,
       "step": 1586
     },
     {
       "epoch": 14.0,
-      "grad_norm": 1.3382197618484497,
+      "grad_norm": 3.546229600906372,
       "learning_rate": 1.5e-05,
-      "loss": 0.1602,
+      "loss": 0.1558,
       "step": 1708
     },
     {
       "epoch": 14.0,
-      "eval_accuracy": 0.8796992481203008,
-      "eval_f1": 0.8589607635206786,
-      "eval_loss": 0.29793980717658997,
-      "eval_precision": 0.8503875968992248,
-      "eval_recall": 0.8698854337152209,
-      "eval_runtime": 5.1508,
-      "eval_samples_per_second": 77.464,
-      "eval_steps_per_second": 9.707,
+      "eval_accuracy": 0.87468671679198,
+      "eval_f1": 0.8530841286673736,
+      "eval_loss": 0.3040328025817871,
+      "eval_precision": 0.8446597760551249,
+      "eval_recall": 0.8638388797963266,
+      "eval_runtime": 1.8508,
+      "eval_samples_per_second": 215.586,
+      "eval_steps_per_second": 27.016,
       "step": 1708
     },
     {
       "epoch": 15.0,
-      "grad_norm": 0.5800030827522278,
+      "grad_norm": 0.5455009937286377,
       "learning_rate": 1.25e-05,
-      "loss": 0.1497,
+      "loss": 0.1538,
       "step": 1830
     },
     {
       "epoch": 15.0,
-      "eval_accuracy": 0.8872180451127819,
-      "eval_f1": 0.8662440310793597,
-      "eval_loss": 0.29690659046173096,
-      "eval_precision": 0.8606158357771261,
-      "eval_recall": 0.872704128023277,
-      "eval_runtime": 5.1461,
-      "eval_samples_per_second": 77.534,
-      "eval_steps_per_second": 9.716,
+      "eval_accuracy": 0.8721804511278195,
+      "eval_f1": 0.8484099018899409,
+      "eval_loss": 0.30823931097984314,
+      "eval_precision": 0.8430645161290322,
+      "eval_recall": 0.8545644662665939,
+      "eval_runtime": 1.848,
+      "eval_samples_per_second": 215.904,
+      "eval_steps_per_second": 27.056,
       "step": 1830
     },
     {
       "epoch": 16.0,
-      "grad_norm": 6.755856037139893,
+      "grad_norm": 2.4319658279418945,
       "learning_rate": 1e-05,
-      "loss": 0.1447,
+      "loss": 0.152,
       "step": 1952
     },
     {
       "epoch": 16.0,
-      "eval_accuracy": 0.8847117794486216,
-      "eval_f1": 0.8616171059774413,
-      "eval_loss": 0.29627636075019836,
-      "eval_precision": 0.859873949579832,
-      "eval_recall": 0.8634297144935443,
-      "eval_runtime": 5.1639,
-      "eval_samples_per_second": 77.267,
-      "eval_steps_per_second": 9.683,
+      "eval_accuracy": 0.8771929824561403,
+      "eval_f1": 0.8483536940081443,
+      "eval_loss": 0.30997762084007263,
+      "eval_precision": 0.8575792287132493,
+      "eval_recall": 0.8406073831605747,
+      "eval_runtime": 1.8455,
+      "eval_samples_per_second": 216.202,
+      "eval_steps_per_second": 27.093,
       "step": 1952
     },
     {
       "epoch": 17.0,
-      "grad_norm": 2.1538662910461426,
+      "grad_norm": 3.420119285583496,
       "learning_rate": 7.5e-06,
-      "loss": 0.1394,
+      "loss": 0.1436,
       "step": 2074
     },
     {
       "epoch": 17.0,
-      "eval_accuracy": 0.8822055137844611,
-      "eval_f1": 0.8589543987905864,
-      "eval_loss": 0.30176377296447754,
-      "eval_precision": 0.8564068100358423,
-      "eval_recall": 0.8616566648481543,
-      "eval_runtime": 5.17,
-      "eval_samples_per_second": 77.176,
-      "eval_steps_per_second": 9.671,
+      "eval_accuracy": 0.87468671679198,
+      "eval_f1": 0.8510304659498208,
+      "eval_loss": 0.31050172448158264,
+      "eval_precision": 0.8463237893248498,
+      "eval_recall": 0.8563375159119839,
+      "eval_runtime": 1.8478,
+      "eval_samples_per_second": 215.931,
+      "eval_steps_per_second": 27.059,
       "step": 2074
     },
     {
       "epoch": 18.0,
-      "grad_norm": 2.54630708694458,
+      "grad_norm": 0.7008257508277893,
       "learning_rate": 5e-06,
-      "loss": 0.1333,
+      "loss": 0.1426,
       "step": 2196
     },
     {
       "epoch": 18.0,
-      "eval_accuracy": 0.8822055137844611,
-      "eval_f1": 0.858259325044405,
-      "eval_loss": 0.30650317668914795,
-      "eval_precision": 0.8573798178418481,
-      "eval_recall": 0.8591562102200401,
-      "eval_runtime": 5.128,
-      "eval_samples_per_second": 77.807,
-      "eval_steps_per_second": 9.75,
+      "eval_accuracy": 0.87468671679198,
+      "eval_f1": 0.8503151260504201,
+      "eval_loss": 0.3118613362312317,
+      "eval_precision": 0.8470628455912955,
+      "eval_recall": 0.8538370612838698,
+      "eval_runtime": 1.8481,
+      "eval_samples_per_second": 215.903,
+      "eval_steps_per_second": 27.056,
       "step": 2196
     },
     {
       "epoch": 19.0,
-      "grad_norm": 0.7696042656898499,
+      "grad_norm": 2.8680572509765625,
       "learning_rate": 2.5e-06,
-      "loss": 0.1406,
+      "loss": 0.1398,
       "step": 2318
     },
     {
       "epoch": 19.0,
-      "eval_accuracy": 0.8822055137844611,
-      "eval_f1": 0.858259325044405,
-      "eval_loss": 0.30623340606689453,
-      "eval_precision": 0.8573798178418481,
-      "eval_recall": 0.8591562102200401,
-      "eval_runtime": 5.1122,
-      "eval_samples_per_second": 78.049,
-      "eval_steps_per_second": 9.781,
+      "eval_accuracy": 0.8796992481203008,
+      "eval_f1": 0.8569892473118279,
+      "eval_loss": 0.316354900598526,
+      "eval_precision": 0.8522004241781549,
+      "eval_recall": 0.8623840698308783,
+      "eval_runtime": 1.8465,
+      "eval_samples_per_second": 216.08,
+      "eval_steps_per_second": 27.078,
       "step": 2318
     },
     {
       "epoch": 20.0,
-      "grad_norm": 3.025254487991333,
+      "grad_norm": 3.442784547805786,
       "learning_rate": 0.0,
-      "loss": 0.1243,
+      "loss": 0.14,
       "step": 2440
     },
     {
       "epoch": 20.0,
-      "eval_accuracy": 0.8822055137844611,
-      "eval_f1": 0.858259325044405,
-      "eval_loss": 0.30716511607170105,
-      "eval_precision": 0.8573798178418481,
-      "eval_recall": 0.8591562102200401,
-      "eval_runtime": 5.0928,
-      "eval_samples_per_second": 78.347,
-      "eval_steps_per_second": 9.818,
+      "eval_accuracy": 0.8796992481203008,
+      "eval_f1": 0.8563025210084034,
+      "eval_loss": 0.31680676341056824,
+      "eval_precision": 0.8529936381473334,
+      "eval_recall": 0.8598836152027641,
+      "eval_runtime": 1.8511,
+      "eval_samples_per_second": 215.549,
+      "eval_steps_per_second": 27.011,
       "step": 2440
     },
     {
       "epoch": 20.0,
       "step": 2440,
       "total_flos": 8460375995160000.0,
-      "train_loss": 0.22629446436147221,
-      "train_runtime": 1956.0503,
-      "train_samples_per_second": 37.269,
-      "train_steps_per_second": 1.247
+      "train_loss": 0.2318932650519199,
+      "train_runtime": 643.2373,
+      "train_samples_per_second": 113.333,
+      "train_steps_per_second": 3.793
     }
   ],
   "logging_steps": 500,