Training in progress, step 100

all_results.json

@@ -1,13 +1,13 @@
 {
     "epoch": 2.0,
-    "eval_accuracy": 0.43737166324435317,
-    "eval_loss": 2.7726495265960693,
-    "eval_runtime": 6.4303,
-    "eval_samples_per_second": 75.735,
-    "eval_steps_per_second": 9.486,
+    "eval_accuracy": 0.31025641025641026,
+    "eval_loss": 2.86479115486145,
+    "eval_runtime": 5.2198,
+    "eval_samples_per_second": 74.716,
+    "eval_steps_per_second": 9.387,
     "total_flos": 1.1959742538326016e+17,
-    "train_loss": 3.1177605087162594,
-    "train_runtime": 412.2299,
-    "train_samples_per_second": 7.549,
-    "train_steps_per_second": 3.775
+    "train_loss": 3.140765584832905,
+    "train_runtime": 414.3391,
+    "train_samples_per_second": 7.511,
+    "train_steps_per_second": 3.755
 }

pytorch_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:8c117e8bb1f9074cb1c08fd0fde2ce1f0ad44ce12fc5ef1e8826260f84328396
+oid sha256:e087dc21cee195e52ecc0a8fad4598f302dda3d3f261baa5f8372944a1914521
 size 170850837

train_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 2.0,
     "total_flos": 1.1959742538326016e+17,
-    "train_loss": 3.1177605087162594,
-    "train_runtime": 412.2299,
-    "train_samples_per_second": 7.549,
-    "train_steps_per_second": 3.775
+    "train_loss": 3.140765584832905,
+    "train_runtime": 414.3391,
+    "train_samples_per_second": 7.511,
+    "train_steps_per_second": 3.755
 }

trainer_state.json

@@ -1,5 +1,5 @@
 {
-  "best_metric": 2.783235788345337,
+  "best_metric": 2.86479115486145,
   "best_model_checkpoint": "/content/Train/checkpoint-1400",
   "epoch": 2.0,
   "global_step": 1556,
@@ -10,1084 +10,1084 @@
     {
       "epoch": 0.01,
       "learning_rate": 0.0001994858611825193,
-      "loss": 3.2724,
+      "loss": 3.2909,
       "step": 10
     },
     {
       "epoch": 0.03,
       "learning_rate": 0.00019820051413881748,
-      "loss": 3.2838,
+      "loss": 3.2775,
       "step": 20
     },
     {
       "epoch": 0.04,
       "learning_rate": 0.0001969151670951157,
-      "loss": 3.3858,
+      "loss": 3.4685,
       "step": 30
     },
     {
       "epoch": 0.05,
       "learning_rate": 0.00019562982005141388,
-      "loss": 3.1207,
+      "loss": 3.1491,
       "step": 40
     },
     {
       "epoch": 0.06,
-      "learning_rate": 0.0001943444730077121,
-      "loss": 3.4414,
+      "learning_rate": 0.00019447300771208227,
+      "loss": 3.5003,
       "step": 50
     },
     {
       "epoch": 0.08,
-      "learning_rate": 0.0001930591259640103,
-      "loss": 3.2139,
+      "learning_rate": 0.00019318766066838048,
+      "loss": 3.2438,
       "step": 60
     },
     {
       "epoch": 0.09,
-      "learning_rate": 0.0001917737789203085,
-      "loss": 3.1755,
+      "learning_rate": 0.00019190231362467867,
+      "loss": 3.2587,
       "step": 70
     },
     {
       "epoch": 0.1,
-      "learning_rate": 0.0001904884318766067,
-      "loss": 3.7297,
+      "learning_rate": 0.00019061696658097688,
+      "loss": 3.8744,
       "step": 80
     },
     {
       "epoch": 0.12,
-      "learning_rate": 0.0001892030848329049,
-      "loss": 3.4405,
+      "learning_rate": 0.00018933161953727507,
+      "loss": 3.5397,
       "step": 90
     },
     {
       "epoch": 0.13,
-      "learning_rate": 0.0001879177377892031,
-      "loss": 3.1529,
+      "learning_rate": 0.00018804627249357326,
+      "loss": 3.2433,
       "step": 100
     },
     {
       "epoch": 0.13,
-      "eval_accuracy": 0.05897435897435897,
-      "eval_loss": 3.211608648300171,
-      "eval_runtime": 5.3959,
-      "eval_samples_per_second": 72.278,
-      "eval_steps_per_second": 9.081,
+      "eval_accuracy": 0.046153846153846156,
+      "eval_loss": 3.216766834259033,
+      "eval_runtime": 4.8707,
+      "eval_samples_per_second": 80.071,
+      "eval_steps_per_second": 10.06,
       "step": 100
     },
     {
       "epoch": 0.14,
-      "learning_rate": 0.0001866323907455013,
-      "loss": 3.1837,
+      "learning_rate": 0.00018676092544987147,
+      "loss": 3.1843,
       "step": 110
     },
     {
       "epoch": 0.15,
-      "learning_rate": 0.0001853470437017995,
-      "loss": 3.2418,
+      "learning_rate": 0.00018547557840616966,
+      "loss": 3.1816,
       "step": 120
     },
     {
       "epoch": 0.17,
-      "learning_rate": 0.0001840616966580977,
-      "loss": 3.2269,
+      "learning_rate": 0.00018419023136246788,
+      "loss": 3.247,
       "step": 130
     },
     {
       "epoch": 0.18,
-      "learning_rate": 0.0001827763496143959,
-      "loss": 3.1065,
+      "learning_rate": 0.00018290488431876606,
+      "loss": 3.2118,
       "step": 140
     },
     {
       "epoch": 0.19,
-      "learning_rate": 0.00018149100257069408,
-      "loss": 3.3563,
+      "learning_rate": 0.00018161953727506428,
+      "loss": 3.3541,
       "step": 150
     },
     {
       "epoch": 0.21,
       "learning_rate": 0.0001803341902313625,
-      "loss": 3.2934,
+      "loss": 3.3353,
       "step": 160
     },
     {
       "epoch": 0.22,
       "learning_rate": 0.00017904884318766068,
-      "loss": 3.4139,
+      "loss": 3.4277,
       "step": 170
     },
     {
       "epoch": 0.23,
       "learning_rate": 0.0001777634961439589,
-      "loss": 3.2532,
+      "loss": 3.2286,
       "step": 180
     },
     {
       "epoch": 0.24,
       "learning_rate": 0.00017647814910025708,
-      "loss": 3.2196,
+      "loss": 3.2597,
       "step": 190
     },
     {
       "epoch": 0.26,
       "learning_rate": 0.00017519280205655527,
-      "loss": 3.1644,
+      "loss": 3.2076,
       "step": 200
     },
     {
       "epoch": 0.26,
-      "eval_accuracy": 0.06923076923076923,
-      "eval_loss": 3.209580421447754,
-      "eval_runtime": 4.8409,
-      "eval_samples_per_second": 80.564,
-      "eval_steps_per_second": 10.122,
+      "eval_accuracy": 0.0641025641025641,
+      "eval_loss": 3.2063100337982178,
+      "eval_runtime": 4.8973,
+      "eval_samples_per_second": 79.635,
+      "eval_steps_per_second": 10.005,
       "step": 200
     },
     {
       "epoch": 0.27,
       "learning_rate": 0.00017390745501285349,
-      "loss": 3.3494,
+      "loss": 3.3286,
       "step": 210
     },
     {
       "epoch": 0.28,
       "learning_rate": 0.00017262210796915167,
-      "loss": 3.336,
+      "loss": 3.3623,
       "step": 220
     },
     {
       "epoch": 0.3,
       "learning_rate": 0.0001713367609254499,
-      "loss": 3.2151,
+      "loss": 3.1635,
       "step": 230
     },
     {
       "epoch": 0.31,
       "learning_rate": 0.00017005141388174808,
-      "loss": 3.2835,
+      "loss": 3.3247,
       "step": 240
     },
     {
       "epoch": 0.32,
       "learning_rate": 0.0001687660668380463,
-      "loss": 3.4358,
+      "loss": 3.4129,
       "step": 250
     },
     {
       "epoch": 0.33,
       "learning_rate": 0.00016748071979434448,
-      "loss": 3.303,
+      "loss": 3.2366,
       "step": 260
     },
     {
       "epoch": 0.35,
       "learning_rate": 0.00016619537275064267,
-      "loss": 3.2475,
+      "loss": 3.164,
       "step": 270
     },
     {
       "epoch": 0.36,
       "learning_rate": 0.00016491002570694088,
-      "loss": 3.1522,
+      "loss": 3.1943,
       "step": 280
     },
     {
       "epoch": 0.37,
       "learning_rate": 0.00016362467866323907,
-      "loss": 3.2906,
+      "loss": 3.3545,
       "step": 290
     },
     {
       "epoch": 0.39,
       "learning_rate": 0.00016233933161953728,
-      "loss": 3.1549,
+      "loss": 3.136,
       "step": 300
     },
     {
       "epoch": 0.39,
-      "eval_accuracy": 0.06923076923076923,
-      "eval_loss": 3.1985976696014404,
-      "eval_runtime": 4.6899,
-      "eval_samples_per_second": 83.157,
-      "eval_steps_per_second": 10.448,
+      "eval_accuracy": 0.10512820512820513,
+      "eval_loss": 3.193809986114502,
+      "eval_runtime": 5.0459,
+      "eval_samples_per_second": 77.29,
+      "eval_steps_per_second": 9.711,
       "step": 300
     },
     {
       "epoch": 0.4,
       "learning_rate": 0.0001610539845758355,
-      "loss": 3.4558,
+      "loss": 3.3411,
       "step": 310
     },
     {
       "epoch": 0.41,
       "learning_rate": 0.00015976863753213369,
-      "loss": 3.2304,
+      "loss": 3.1711,
       "step": 320
     },
     {
       "epoch": 0.42,
       "learning_rate": 0.0001584832904884319,
-      "loss": 3.2686,
+      "loss": 3.2203,
       "step": 330
     },
     {
       "epoch": 0.44,
       "learning_rate": 0.0001571979434447301,
-      "loss": 3.3832,
+      "loss": 3.3112,
       "step": 340
     },
     {
       "epoch": 0.45,
       "learning_rate": 0.00015591259640102828,
-      "loss": 3.3443,
+      "loss": 3.3413,
       "step": 350
     },
     {
       "epoch": 0.46,
       "learning_rate": 0.0001546272493573265,
-      "loss": 3.2797,
+      "loss": 3.3321,
       "step": 360
     },
     {
       "epoch": 0.48,
       "learning_rate": 0.00015334190231362468,
-      "loss": 3.2673,
+      "loss": 3.2329,
       "step": 370
     },
     {
       "epoch": 0.49,
       "learning_rate": 0.0001520565552699229,
-      "loss": 3.1919,
+      "loss": 3.1901,
       "step": 380
     },
     {
       "epoch": 0.5,
       "learning_rate": 0.00015077120822622108,
-      "loss": 3.1238,
+      "loss": 3.1793,
       "step": 390
     },
     {
       "epoch": 0.51,
       "learning_rate": 0.0001494858611825193,
-      "loss": 3.2998,
+      "loss": 3.2283,
       "step": 400
     },
     {
       "epoch": 0.51,
       "eval_accuracy": 0.1076923076923077,
-      "eval_loss": 3.1967546939849854,
-      "eval_runtime": 4.6763,
-      "eval_samples_per_second": 83.399,
-      "eval_steps_per_second": 10.478,
+      "eval_loss": 3.1784756183624268,
+      "eval_runtime": 4.9684,
+      "eval_samples_per_second": 78.496,
+      "eval_steps_per_second": 9.862,
       "step": 400
     },
     {
       "epoch": 0.53,
       "learning_rate": 0.00014820051413881748,
-      "loss": 3.237,
+      "loss": 3.3054,
       "step": 410
     },
     {
       "epoch": 0.54,
       "learning_rate": 0.00014691516709511567,
-      "loss": 3.1511,
+      "loss": 3.1213,
       "step": 420
     },
     {
       "epoch": 0.55,
       "learning_rate": 0.00014562982005141388,
-      "loss": 3.1972,
+      "loss": 3.1617,
       "step": 430
     },
     {
       "epoch": 0.57,
       "learning_rate": 0.0001443444730077121,
-      "loss": 3.3029,
+      "loss": 3.3134,
       "step": 440
     },
     {
       "epoch": 0.58,
       "learning_rate": 0.0001430591259640103,
-      "loss": 3.2639,
+      "loss": 3.1977,
       "step": 450
     },
     {
       "epoch": 0.59,
       "learning_rate": 0.0001417737789203085,
-      "loss": 3.2602,
+      "loss": 3.4898,
       "step": 460
     },
     {
       "epoch": 0.6,
       "learning_rate": 0.0001404884318766067,
-      "loss": 3.1096,
+      "loss": 3.215,
       "step": 470
     },
     {
       "epoch": 0.62,
       "learning_rate": 0.0001392030848329049,
-      "loss": 3.2692,
+      "loss": 3.1978,
       "step": 480
     },
     {
       "epoch": 0.63,
       "learning_rate": 0.0001379177377892031,
-      "loss": 3.3299,
+      "loss": 3.1939,
       "step": 490
     },
     {
       "epoch": 0.64,
       "learning_rate": 0.0001366323907455013,
-      "loss": 3.1344,
+      "loss": 3.1354,
       "step": 500
     },
     {
       "epoch": 0.64,
-      "eval_accuracy": 0.1717948717948718,
-      "eval_loss": 3.1666767597198486,
-      "eval_runtime": 4.8058,
-      "eval_samples_per_second": 81.152,
-      "eval_steps_per_second": 10.196,
+      "eval_accuracy": 0.1282051282051282,
+      "eval_loss": 3.1671173572540283,
+      "eval_runtime": 4.8377,
+      "eval_samples_per_second": 80.617,
+      "eval_steps_per_second": 10.129,
       "step": 500
     },
     {
       "epoch": 0.66,
       "learning_rate": 0.0001353470437017995,
-      "loss": 3.0552,
+      "loss": 3.0448,
       "step": 510
     },
     {
       "epoch": 0.67,
       "learning_rate": 0.00013406169665809768,
-      "loss": 3.3452,
+      "loss": 3.4826,
       "step": 520
     },
     {
       "epoch": 0.68,
       "learning_rate": 0.0001327763496143959,
-      "loss": 3.3108,
+      "loss": 3.2411,
       "step": 530
     },
     {
       "epoch": 0.69,
       "learning_rate": 0.00013149100257069408,
-      "loss": 3.0806,
+      "loss": 3.0833,
       "step": 540
     },
     {
       "epoch": 0.71,
       "learning_rate": 0.0001302056555269923,
-      "loss": 3.4034,
+      "loss": 3.2685,
       "step": 550
     },
     {
       "epoch": 0.72,
       "learning_rate": 0.00012892030848329049,
-      "loss": 3.0853,
+      "loss": 3.0119,
       "step": 560
     },
     {
       "epoch": 0.73,
       "learning_rate": 0.00012763496143958867,
-      "loss": 3.1718,
+      "loss": 3.1179,
       "step": 570
     },
     {
       "epoch": 0.75,
       "learning_rate": 0.00012634961439588692,
-      "loss": 3.1265,
+      "loss": 3.1146,
       "step": 580
     },
     {
       "epoch": 0.76,
       "learning_rate": 0.0001250642673521851,
-      "loss": 3.183,
+      "loss": 3.2957,
       "step": 590
     },
     {
       "epoch": 0.77,
       "learning_rate": 0.0001237789203084833,
-      "loss": 3.3638,
+      "loss": 3.3212,
       "step": 600
     },
     {
       "epoch": 0.77,
-      "eval_accuracy": 0.17435897435897435,
-      "eval_loss": 3.1383864879608154,
-      "eval_runtime": 4.7872,
-      "eval_samples_per_second": 81.466,
-      "eval_steps_per_second": 10.236,
+      "eval_accuracy": 0.15897435897435896,
+      "eval_loss": 3.1384565830230713,
+      "eval_runtime": 4.7255,
+      "eval_samples_per_second": 82.53,
+      "eval_steps_per_second": 10.369,
       "step": 600
     },
     {
       "epoch": 0.78,
       "learning_rate": 0.0001224935732647815,
-      "loss": 3.1725,
+      "loss": 3.1221,
       "step": 610
     },
     {
       "epoch": 0.8,
       "learning_rate": 0.0001212082262210797,
-      "loss": 3.5329,
+      "loss": 3.4999,
       "step": 620
     },
     {
       "epoch": 0.81,
       "learning_rate": 0.0001199228791773779,
-      "loss": 3.1399,
+      "loss": 3.2125,
       "step": 630
     },
     {
       "epoch": 0.82,
       "learning_rate": 0.0001186375321336761,
-      "loss": 3.1998,
+      "loss": 3.1819,
       "step": 640
     },
     {
       "epoch": 0.84,
       "learning_rate": 0.0001173521850899743,
-      "loss": 3.1562,
+      "loss": 3.2923,
       "step": 650
     },
     {
       "epoch": 0.85,
       "learning_rate": 0.0001160668380462725,
-      "loss": 3.3428,
+      "loss": 3.3054,
       "step": 660
     },
     {
       "epoch": 0.86,
       "learning_rate": 0.0001147814910025707,
-      "loss": 3.0586,
+      "loss": 3.0682,
       "step": 670
     },
     {
       "epoch": 0.87,
       "learning_rate": 0.0001134961439588689,
-      "loss": 3.0464,
+      "loss": 3.111,
       "step": 680
     },
     {
       "epoch": 0.89,
       "learning_rate": 0.00011221079691516709,
-      "loss": 3.1097,
+      "loss": 3.2225,
       "step": 690
     },
     {
       "epoch": 0.9,
       "learning_rate": 0.00011092544987146529,
-      "loss": 3.1482,
+      "loss": 3.3626,
       "step": 700
     },
     {
       "epoch": 0.9,
-      "eval_accuracy": 0.20256410256410257,
-      "eval_loss": 3.096599578857422,
-      "eval_runtime": 4.7875,
-      "eval_samples_per_second": 81.462,
-      "eval_steps_per_second": 10.235,
+      "eval_accuracy": 0.2128205128205128,
+      "eval_loss": 3.1268680095672607,
+      "eval_runtime": 4.7092,
+      "eval_samples_per_second": 82.816,
+      "eval_steps_per_second": 10.405,
       "step": 700
     },
     {
       "epoch": 0.91,
       "learning_rate": 0.00010964010282776349,
-      "loss": 3.1459,
+      "loss": 3.1624,
       "step": 710
     },
     {
       "epoch": 0.93,
       "learning_rate": 0.00010835475578406172,
-      "loss": 3.1789,
+      "loss": 3.127,
       "step": 720
     },
     {
       "epoch": 0.94,
       "learning_rate": 0.0001070694087403599,
-      "loss": 3.2216,
+      "loss": 3.3503,
       "step": 730
     },
     {
       "epoch": 0.95,
       "learning_rate": 0.0001057840616966581,
-      "loss": 2.8712,
+      "loss": 2.9137,
       "step": 740
     },
     {
       "epoch": 0.96,
-      "learning_rate": 0.00010449871465295631,
-      "loss": 3.1489,
+      "learning_rate": 0.00010462724935732648,
+      "loss": 3.2102,
       "step": 750
     },
     {
       "epoch": 0.98,
-      "learning_rate": 0.00010321336760925451,
-      "loss": 3.438,
+      "learning_rate": 0.00010334190231362468,
+      "loss": 3.3043,
       "step": 760
     },
     {
       "epoch": 0.99,
-      "learning_rate": 0.00010192802056555271,
-      "loss": 3.23,
+      "learning_rate": 0.00010205655526992288,
+      "loss": 3.3108,
       "step": 770
     },
     {
       "epoch": 1.0,
-      "learning_rate": 0.00010064267352185091,
-      "loss": 3.1966,
+      "learning_rate": 0.00010077120822622108,
+      "loss": 3.3166,
       "step": 780
     },
     {
       "epoch": 1.02,
-      "learning_rate": 9.93573264781491e-05,
-      "loss": 3.0066,
+      "learning_rate": 9.948586118251929e-05,
+      "loss": 2.9514,
       "step": 790
     },
     {
       "epoch": 1.03,
-      "learning_rate": 9.80719794344473e-05,
-      "loss": 3.1366,
+      "learning_rate": 9.820051413881749e-05,
+      "loss": 3.1565,
       "step": 800
     },
     {
       "epoch": 1.03,
-      "eval_accuracy": 0.18974358974358974,
-      "eval_loss": 3.0483875274658203,
-      "eval_runtime": 4.6499,
-      "eval_samples_per_second": 83.872,
-      "eval_steps_per_second": 10.538,
+      "eval_accuracy": 0.1564102564102564,
+      "eval_loss": 3.0775740146636963,
+      "eval_runtime": 5.0971,
+      "eval_samples_per_second": 76.514,
+      "eval_steps_per_second": 9.613,
       "step": 800
     },
     {
       "epoch": 1.04,
-      "learning_rate": 9.67866323907455e-05,
-      "loss": 3.1106,
+      "learning_rate": 9.691516709511569e-05,
+      "loss": 3.0278,
       "step": 810
     },
     {
       "epoch": 1.05,
-      "learning_rate": 9.550128534704372e-05,
-      "loss": 2.9235,
+      "learning_rate": 9.562982005141389e-05,
+      "loss": 2.931,
       "step": 820
     },
     {
       "epoch": 1.07,
-      "learning_rate": 9.421593830334192e-05,
-      "loss": 3.0565,
+      "learning_rate": 9.434447300771209e-05,
+      "loss": 3.0111,
       "step": 830
     },
     {
       "epoch": 1.08,
-      "learning_rate": 9.29305912596401e-05,
-      "loss": 3.0503,
+      "learning_rate": 9.305912596401029e-05,
+      "loss": 3.0158,
       "step": 840
     },
     {
       "epoch": 1.09,
-      "learning_rate": 9.16452442159383e-05,
-      "loss": 3.2726,
+      "learning_rate": 9.177377892030848e-05,
+      "loss": 3.2079,
       "step": 850
     },
     {
       "epoch": 1.11,
-      "learning_rate": 9.03598971722365e-05,
-      "loss": 3.0256,
+      "learning_rate": 9.048843187660668e-05,
+      "loss": 3.0987,
       "step": 860
     },
     {
       "epoch": 1.12,
-      "learning_rate": 8.907455012853471e-05,
-      "loss": 3.0892,
+      "learning_rate": 8.92030848329049e-05,
+      "loss": 3.0854,
       "step": 870
     },
     {
       "epoch": 1.13,
-      "learning_rate": 8.778920308483291e-05,
-      "loss": 3.1357,
+      "learning_rate": 8.79177377892031e-05,
+      "loss": 3.2495,
       "step": 880
     },
     {
       "epoch": 1.14,
-      "learning_rate": 8.650385604113111e-05,
-      "loss": 3.1402,
+      "learning_rate": 8.66323907455013e-05,
+      "loss": 3.018,
       "step": 890
     },
     {
       "epoch": 1.16,
-      "learning_rate": 8.521850899742931e-05,
-      "loss": 3.0206,
+      "learning_rate": 8.534704370179948e-05,
+      "loss": 3.0873,
       "step": 900
     },
     {
       "epoch": 1.16,
-      "eval_accuracy": 0.30256410256410254,
-      "eval_loss": 3.016371250152588,
-      "eval_runtime": 4.6485,
-      "eval_samples_per_second": 83.899,
-      "eval_steps_per_second": 10.541,
+      "eval_accuracy": 0.21794871794871795,
+      "eval_loss": 3.0445914268493652,
+      "eval_runtime": 4.9901,
+      "eval_samples_per_second": 78.155,
+      "eval_steps_per_second": 9.819,
       "step": 900
     },
     {
       "epoch": 1.17,
-      "learning_rate": 8.393316195372751e-05,
-      "loss": 3.0057,
+      "learning_rate": 8.406169665809769e-05,
+      "loss": 3.0805,
       "step": 910
     },
     {
       "epoch": 1.18,
-      "learning_rate": 8.264781491002571e-05,
-      "loss": 3.1709,
+      "learning_rate": 8.277634961439589e-05,
+      "loss": 3.2969,
       "step": 920
     },
     {
       "epoch": 1.2,
-      "learning_rate": 8.136246786632391e-05,
-      "loss": 3.2243,
+      "learning_rate": 8.149100257069409e-05,
+      "loss": 3.1201,
       "step": 930
     },
     {
       "epoch": 1.21,
-      "learning_rate": 8.007712082262212e-05,
-      "loss": 3.19,
+      "learning_rate": 8.02056555269923e-05,
+      "loss": 3.2055,
       "step": 940
     },
     {
       "epoch": 1.22,
-      "learning_rate": 7.87917737789203e-05,
-      "loss": 3.2785,
+      "learning_rate": 7.892030848329049e-05,
+      "loss": 3.1691,
       "step": 950
     },
     {
       "epoch": 1.23,
-      "learning_rate": 7.750642673521852e-05,
-      "loss": 2.8424,
+      "learning_rate": 7.763496143958869e-05,
+      "loss": 2.9756,
       "step": 960
     },
     {
       "epoch": 1.25,
-      "learning_rate": 7.622107969151672e-05,
-      "loss": 3.2575,
+      "learning_rate": 7.634961439588689e-05,
+      "loss": 3.1635,
       "step": 970
     },
     {
       "epoch": 1.26,
-      "learning_rate": 7.493573264781492e-05,
-      "loss": 2.9012,
+      "learning_rate": 7.50642673521851e-05,
+      "loss": 2.9979,
       "step": 980
     },
     {
       "epoch": 1.27,
-      "learning_rate": 7.365038560411311e-05,
-      "loss": 3.1516,
+      "learning_rate": 7.37789203084833e-05,
+      "loss": 3.1009,
       "step": 990
     },
     {
       "epoch": 1.29,
-      "learning_rate": 7.236503856041131e-05,
-      "loss": 2.921,
+      "learning_rate": 7.24935732647815e-05,
+      "loss": 2.9991,
       "step": 1000
     },
     {
       "epoch": 1.29,
-      "eval_accuracy": 0.3230769230769231,
-      "eval_loss": 2.984600305557251,
-      "eval_runtime": 4.9539,
-      "eval_samples_per_second": 78.726,
-      "eval_steps_per_second": 9.891,
+      "eval_accuracy": 0.2743589743589744,
+      "eval_loss": 2.994030475616455,
+      "eval_runtime": 5.0834,
+      "eval_samples_per_second": 76.721,
+      "eval_steps_per_second": 9.639,
       "step": 1000
     },
     {
       "epoch": 1.3,
-      "learning_rate": 7.107969151670951e-05,
-      "loss": 2.9828,
+      "learning_rate": 7.12082262210797e-05,
+      "loss": 3.0882,
       "step": 1010
     },
     {
       "epoch": 1.31,
-      "learning_rate": 6.979434447300771e-05,
-      "loss": 3.0413,
+      "learning_rate": 6.99228791773779e-05,
+      "loss": 3.0527,
       "step": 1020
     },
     {
       "epoch": 1.32,
-      "learning_rate": 6.850899742930593e-05,
-      "loss": 2.9759,
+      "learning_rate": 6.86375321336761e-05,
+      "loss": 2.9187,
       "step": 1030
     },
     {
       "epoch": 1.34,
-      "learning_rate": 6.722365038560411e-05,
-      "loss": 2.9434,
+      "learning_rate": 6.73521850899743e-05,
+      "loss": 3.011,
       "step": 1040
     },
     {
       "epoch": 1.35,
-      "learning_rate": 6.593830334190231e-05,
-      "loss": 3.0893,
+      "learning_rate": 6.606683804627249e-05,
+      "loss": 3.138,
       "step": 1050
     },
     {
       "epoch": 1.36,
-      "learning_rate": 6.465295629820052e-05,
-      "loss": 3.0797,
+      "learning_rate": 6.478149100257069e-05,
+      "loss": 3.0516,
       "step": 1060
     },
     {
       "epoch": 1.38,
-      "learning_rate": 6.336760925449872e-05,
-      "loss": 2.9291,
+      "learning_rate": 6.349614395886889e-05,
+      "loss": 2.9856,
       "step": 1070
     },
     {
       "epoch": 1.39,
-      "learning_rate": 6.208226221079692e-05,
-      "loss": 3.0357,
+      "learning_rate": 6.22107969151671e-05,
+      "loss": 2.9641,
       "step": 1080
     },
     {
       "epoch": 1.4,
-      "learning_rate": 6.079691516709511e-05,
-      "loss": 3.0109,
+      "learning_rate": 6.09254498714653e-05,
+      "loss": 2.9252,
       "step": 1090
     },
     {
       "epoch": 1.41,
-      "learning_rate": 5.951156812339333e-05,
-      "loss": 3.0027,
+      "learning_rate": 5.96401028277635e-05,
+      "loss": 2.9279,
       "step": 1100
     },
     {
       "epoch": 1.41,
-      "eval_accuracy": 0.33589743589743587,
-      "eval_loss": 2.933824062347412,
-      "eval_runtime": 4.8337,
-      "eval_samples_per_second": 80.684,
-      "eval_steps_per_second": 10.137,
+      "eval_accuracy": 0.2717948717948718,
+      "eval_loss": 2.9565303325653076,
+      "eval_runtime": 4.7654,
+      "eval_samples_per_second": 81.839,
+      "eval_steps_per_second": 10.282,
       "step": 1100
     },
     {
       "epoch": 1.43,
-      "learning_rate": 5.822622107969152e-05,
-      "loss": 3.2462,
+      "learning_rate": 5.83547557840617e-05,
+      "loss": 3.1913,
       "step": 1110
     },
     {
       "epoch": 1.44,
-      "learning_rate": 5.694087403598972e-05,
-      "loss": 2.9376,
+      "learning_rate": 5.7069408740359896e-05,
+      "loss": 2.8937,
       "step": 1120
     },
     {
       "epoch": 1.45,
-      "learning_rate": 5.5655526992287924e-05,
-      "loss": 2.9242,
+      "learning_rate": 5.57840616966581e-05,
+      "loss": 3.0555,
       "step": 1130
     },
     {
       "epoch": 1.47,
-      "learning_rate": 5.437017994858612e-05,
-      "loss": 2.862,
+      "learning_rate": 5.44987146529563e-05,
+      "loss": 3.0048,
       "step": 1140
     },
     {
       "epoch": 1.48,
-      "learning_rate": 5.308483290488432e-05,
-      "loss": 3.1005,
+      "learning_rate": 5.3213367609254506e-05,
+      "loss": 3.1386,
       "step": 1150
     },
     {
       "epoch": 1.49,
-      "learning_rate": 5.1799485861182514e-05,
-      "loss": 3.1811,
+      "learning_rate": 5.192802056555271e-05,
+      "loss": 3.1599,
       "step": 1160
     },
     {
       "epoch": 1.5,
-      "learning_rate": 5.051413881748073e-05,
-      "loss": 3.1254,
+      "learning_rate": 5.06426735218509e-05,
+      "loss": 3.1027,
       "step": 1170
     },
     {
       "epoch": 1.52,
-      "learning_rate": 4.922879177377892e-05,
-      "loss": 2.797,
+      "learning_rate": 4.93573264781491e-05,
+      "loss": 2.7227,
       "step": 1180
     },
     {
       "epoch": 1.53,
-      "learning_rate": 4.7943444730077124e-05,
-      "loss": 3.1899,
+      "learning_rate": 4.80719794344473e-05,
+      "loss": 3.2903,
       "step": 1190
     },
     {
       "epoch": 1.54,
-      "learning_rate": 4.6658097686375325e-05,
-      "loss": 2.9047,
+      "learning_rate": 4.6786632390745505e-05,
+      "loss": 2.8635,
       "step": 1200
     },
     {
       "epoch": 1.54,
-      "eval_accuracy": 0.34615384615384615,
-      "eval_loss": 2.8916842937469482,
-      "eval_runtime": 4.6342,
-      "eval_samples_per_second": 84.157,
-      "eval_steps_per_second": 10.574,
+      "eval_accuracy": 0.27692307692307694,
+      "eval_loss": 2.9233222007751465,
+      "eval_runtime": 4.7217,
+      "eval_samples_per_second": 82.598,
+      "eval_steps_per_second": 10.378,
       "step": 1200
     },
     {
       "epoch": 1.56,
-      "learning_rate": 4.537275064267352e-05,
-      "loss": 3.1712,
+      "learning_rate": 4.5501285347043706e-05,
+      "loss": 3.4916,
       "step": 1210
     },
     {
       "epoch": 1.57,
-      "learning_rate": 4.408740359897173e-05,
-      "loss": 2.9201,
+      "learning_rate": 4.42159383033419e-05,
+      "loss": 2.9104,
       "step": 1220
     },
     {
       "epoch": 1.58,
-      "learning_rate": 4.280205655526993e-05,
-      "loss": 2.9952,
+      "learning_rate": 4.293059125964011e-05,
+      "loss": 3.0576,
       "step": 1230
     },
     {
       "epoch": 1.59,
-      "learning_rate": 4.151670951156812e-05,
-      "loss": 2.8225,
+      "learning_rate": 4.16452442159383e-05,
+      "loss": 2.8137,
       "step": 1240
     },
     {
       "epoch": 1.61,
-      "learning_rate": 4.0231362467866324e-05,
-      "loss": 2.9507,
+      "learning_rate": 4.0359897172236504e-05,
+      "loss": 3.1792,
       "step": 1250
     },
     {
       "epoch": 1.62,
-      "learning_rate": 3.8946015424164526e-05,
-      "loss": 2.96,
+      "learning_rate": 3.9074550128534705e-05,
+      "loss": 3.0377,
       "step": 1260
     },
     {
       "epoch": 1.63,
-      "learning_rate": 3.766066838046273e-05,
-      "loss": 3.0775,
+      "learning_rate": 3.7789203084832907e-05,
+      "loss": 3.1019,
       "step": 1270
     },
     {
       "epoch": 1.65,
-      "learning_rate": 3.637532133676093e-05,
-      "loss": 2.9242,
+      "learning_rate": 3.650385604113111e-05,
+      "loss": 2.9101,
       "step": 1280
     },
     {
       "epoch": 1.66,
-      "learning_rate": 3.508997429305913e-05,
-      "loss": 2.8814,
+      "learning_rate": 3.521850899742931e-05,
+      "loss": 2.9055,
       "step": 1290
     },
     {
       "epoch": 1.67,
-      "learning_rate": 3.380462724935733e-05,
-      "loss": 2.8579,
+      "learning_rate": 3.393316195372751e-05,
+      "loss": 2.9502,
       "step": 1300
     },
     {
       "epoch": 1.67,
-      "eval_accuracy": 0.4025641025641026,
-      "eval_loss": 2.8616135120391846,
-      "eval_runtime": 4.7031,
-      "eval_samples_per_second": 82.924,
-      "eval_steps_per_second": 10.419,
+      "eval_accuracy": 0.29743589743589743,
+      "eval_loss": 2.92795467376709,
+      "eval_runtime": 4.8649,
+      "eval_samples_per_second": 80.166,
+      "eval_steps_per_second": 10.072,
       "step": 1300
     },
     {
       "epoch": 1.68,
-      "learning_rate": 3.251928020565553e-05,
-      "loss": 2.6972,
+      "learning_rate": 3.264781491002571e-05,
+      "loss": 2.837,
       "step": 1310
     },
     {
       "epoch": 1.7,
-      "learning_rate": 3.1233933161953726e-05,
-      "loss": 2.8446,
+      "learning_rate": 3.1362467866323906e-05,
+      "loss": 3.0021,
       "step": 1320
     },
     {
       "epoch": 1.71,
-      "learning_rate": 2.994858611825193e-05,
-      "loss": 2.915,
+      "learning_rate": 3.0077120822622107e-05,
+      "loss": 2.974,
       "step": 1330
     },
     {
       "epoch": 1.72,
-      "learning_rate": 2.866323907455013e-05,
-      "loss": 3.1146,
+      "learning_rate": 2.879177377892031e-05,
+      "loss": 3.1642,
       "step": 1340
     },
     {
       "epoch": 1.74,
-      "learning_rate": 2.737789203084833e-05,
-      "loss": 2.9168,
+      "learning_rate": 2.750642673521851e-05,
+      "loss": 2.8847,
       "step": 1350
     },
     {
       "epoch": 1.75,
-      "learning_rate": 2.6092544987146534e-05,
-      "loss": 2.8762,
+      "learning_rate": 2.622107969151671e-05,
+      "loss": 2.963,
       "step": 1360
     },
     {
       "epoch": 1.76,
-      "learning_rate": 2.480719794344473e-05,
-      "loss": 2.9499,
+      "learning_rate": 2.493573264781491e-05,
+      "loss": 2.9772,
       "step": 1370
     },
     {
       "epoch": 1.77,
-      "learning_rate": 2.3521850899742933e-05,
-      "loss": 2.907,
+      "learning_rate": 2.3650385604113112e-05,
+      "loss": 2.9531,
       "step": 1380
     },
     {
       "epoch": 1.79,
-      "learning_rate": 2.2236503856041134e-05,
-      "loss": 3.0447,
+      "learning_rate": 2.236503856041131e-05,
+      "loss": 3.0706,
       "step": 1390
     },
     {
       "epoch": 1.8,
-      "learning_rate": 2.095115681233933e-05,
-      "loss": 2.988,
+      "learning_rate": 2.107969151670951e-05,
+      "loss": 3.084,
       "step": 1400
     },
     {
       "epoch": 1.8,
-      "eval_accuracy": 0.4076923076923077,
-      "eval_loss": 2.783235788345337,
-      "eval_runtime": 4.64,
-      "eval_samples_per_second": 84.052,
-      "eval_steps_per_second": 10.56,
+      "eval_accuracy": 0.31025641025641026,
+      "eval_loss": 2.86479115486145,
+      "eval_runtime": 4.9511,
+      "eval_samples_per_second": 78.77,
+      "eval_steps_per_second": 9.897,
       "step": 1400
     },
     {
       "epoch": 1.81,
-      "learning_rate": 1.9665809768637533e-05,
-      "loss": 2.8267,
+      "learning_rate": 1.9794344473007716e-05,
+      "loss": 2.8422,
       "step": 1410
     },
     {
       "epoch": 1.83,
-      "learning_rate": 1.8380462724935734e-05,
-      "loss": 2.7718,
+      "learning_rate": 1.8508997429305914e-05,
+      "loss": 2.8483,
       "step": 1420
     },
     {
       "epoch": 1.84,
-      "learning_rate": 1.7095115681233935e-05,
-      "loss": 2.8327,
+      "learning_rate": 1.7223650385604115e-05,
+      "loss": 2.8629,
       "step": 1430
     },
     {
       "epoch": 1.85,
-      "learning_rate": 1.5809768637532136e-05,
-      "loss": 2.9942,
+      "learning_rate": 1.5938303341902313e-05,
+      "loss": 3.0296,
       "step": 1440
     },
     {
       "epoch": 1.86,
-      "learning_rate": 1.4524421593830334e-05,
-      "loss": 3.0647,
+      "learning_rate": 1.4652956298200515e-05,
+      "loss": 3.2828,
       "step": 1450
     },
     {
       "epoch": 1.88,
-      "learning_rate": 1.3239074550128535e-05,
-      "loss": 2.8267,
+      "learning_rate": 1.3367609254498715e-05,
+      "loss": 2.8906,
       "step": 1460
     },
     {
       "epoch": 1.89,
-      "learning_rate": 1.1953727506426736e-05,
-      "loss": 2.8312,
+      "learning_rate": 1.2082262210796916e-05,
+      "loss": 2.8501,
       "step": 1470
     },
     {
       "epoch": 1.9,
-      "learning_rate": 1.0668380462724936e-05,
-      "loss": 3.1317,
+      "learning_rate": 1.0796915167095115e-05,
+      "loss": 3.1941,
       "step": 1480
     },
     {
       "epoch": 1.92,
-      "learning_rate": 9.383033419023137e-06,
-      "loss": 3.1244,
+      "learning_rate": 9.511568123393317e-06,
+      "loss": 3.2227,
       "step": 1490
     },
     {
       "epoch": 1.93,
-      "learning_rate": 8.097686375321336e-06,
-      "loss": 2.8553,
+      "learning_rate": 8.226221079691516e-06,
+      "loss": 2.8916,
       "step": 1500
     },
     {
       "epoch": 1.93,
-      "eval_accuracy": 0.3871794871794872,
-      "eval_loss": 2.821709632873535,
-      "eval_runtime": 4.6901,
-      "eval_samples_per_second": 83.154,
-      "eval_steps_per_second": 10.448,
+      "eval_accuracy": 0.3,
+      "eval_loss": 2.87473464012146,
+      "eval_runtime": 5.5554,
+      "eval_samples_per_second": 70.202,
+      "eval_steps_per_second": 8.82,
       "step": 1500
     },
     {
       "epoch": 1.94,
-      "learning_rate": 6.812339331619537e-06,
-      "loss": 2.6944,
+      "learning_rate": 6.940874035989718e-06,
+      "loss": 2.696,
       "step": 1510
     },
     {
       "epoch": 1.95,
-      "learning_rate": 5.526992287917738e-06,
-      "loss": 3.1573,
+      "learning_rate": 5.6555269922879175e-06,
+      "loss": 3.1394,
       "step": 1520
     },
     {
       "epoch": 1.97,
-      "learning_rate": 4.241645244215939e-06,
-      "loss": 2.8412,
+      "learning_rate": 4.370179948586119e-06,
+      "loss": 2.8452,
       "step": 1530
     },
     {
       "epoch": 1.98,
-      "learning_rate": 2.956298200514139e-06,
-      "loss": 2.6937,
+      "learning_rate": 3.084832904884319e-06,
+      "loss": 2.7683,
       "step": 1540
     },
     {
       "epoch": 1.99,
-      "learning_rate": 1.6709511568123394e-06,
-      "loss": 2.65,
+      "learning_rate": 1.7994858611825194e-06,
+      "loss": 2.681,
       "step": 1550
     },
     {
       "epoch": 2.0,
       "step": 1556,
       "total_flos": 1.1959742538326016e+17,
-      "train_loss": 3.1177605087162594,
-      "train_runtime": 412.2299,
-      "train_samples_per_second": 7.549,
-      "train_steps_per_second": 3.775
+      "train_loss": 3.140765584832905,
+      "train_runtime": 414.3391,
+      "train_samples_per_second": 7.511,
+      "train_steps_per_second": 3.755
     },
     {
       "epoch": 2.0,
-      "eval_accuracy": 0.4076923076923077,
-      "eval_loss": 2.783235788345337,
-      "eval_runtime": 5.3396,
-      "eval_samples_per_second": 73.039,
-      "eval_steps_per_second": 9.177,
+      "eval_accuracy": 0.31025641025641026,
+      "eval_loss": 2.86479115486145,
+      "eval_runtime": 5.2198,
+      "eval_samples_per_second": 74.716,
+      "eval_steps_per_second": 9.387,
       "step": 1556
     }
   ],

validation_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 2.0,
-    "eval_accuracy": 0.4076923076923077,
-    "eval_loss": 2.783235788345337,
-    "eval_runtime": 5.3396,
-    "eval_samples_per_second": 73.039,
-    "eval_steps_per_second": 9.177
+    "eval_accuracy": 0.31025641025641026,
+    "eval_loss": 2.86479115486145,
+    "eval_runtime": 5.2198,
+    "eval_samples_per_second": 74.716,
+    "eval_steps_per_second": 9.387
 }