diff --git "a/trainer_state.json" "b/trainer_state.json"
new file mode 100644--- /dev/null
+++ "b/trainer_state.json"
@@ -0,0 +1,7642 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 0.6628735568690273,
+  "global_step": 12000,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.0,
+      "learning_rate": 6.226286401385679e-05,
+      "loss": 1.4854,
+      "step": 10
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 8.100585369797515e-05,
+      "loss": 1.3991,
+      "step": 20
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 9.196979981458785e-05,
+      "loss": 1.2649,
+      "step": 30
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 9.97488433820935e-05,
+      "loss": 1.3398,
+      "step": 40
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0001057827383435952,
+      "loss": 1.1817,
+      "step": 50
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0001107127894987062,
+      "loss": 1.3592,
+      "step": 60
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.00011488108835764138,
+      "loss": 1.1946,
+      "step": 70
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.00011849183306621185,
+      "loss": 1.2433,
+      "step": 80
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.00012167673561531892,
+      "loss": 1.2936,
+      "step": 90
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.00012452572802771359,
+      "loss": 1.1955,
+      "step": 100
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.00012710295515488854,
+      "loss": 1.3258,
+      "step": 110
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.00012945577918282457,
+      "loss": 1.2087,
+      "step": 120
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.00013162016747767712,
+      "loss": 1.2059,
+      "step": 130
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.00013362407804175968,
+      "loss": 1.2616,
+      "step": 140
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.00013548967414432628,
+      "loss": 1.1475,
+      "step": 150
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0001372348227503302,
+      "loss": 1.3272,
+      "step": 160
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.000138874137881629,
+      "loss": 1.2506,
+      "step": 170
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0001404197252994373,
+      "loss": 1.2071,
+      "step": 180
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.00014188172557721247,
+      "loss": 1.3089,
+      "step": 190
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0001432687177118319,
+      "loss": 1.1428,
+      "step": 200
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.00014458802415837242,
+      "loss": 1.2244,
+      "step": 210
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0001458459448390069,
+      "loss": 1.1951,
+      "step": 220
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.00014704793909170365,
+      "loss": 1.215,
+      "step": 230
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.00014819876886694292,
+      "loss": 1.2766,
+      "step": 240
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.00014930261267333362,
+      "loss": 1.1436,
+      "step": 250
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.00015036315716179547,
+      "loss": 1.2817,
+      "step": 260
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.00015138367141604998,
+      "loss": 1.2022,
+      "step": 270
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 0.00015236706772587806,
+      "loss": 1.2149,
+      "step": 280
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 0.0001533159516911759,
+      "loss": 1.2048,
+      "step": 290
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 0.00015423266382844463,
+      "loss": 1.1425,
+      "step": 300
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 0.00015511931435253522,
+      "loss": 1.2738,
+      "step": 310
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 0.0001559778124344486,
+      "loss": 1.1905,
+      "step": 320
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 0.00015680989095561958,
+      "loss": 1.1633,
+      "step": 330
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 0.00015761712756574739,
+      "loss": 1.2421,
+      "step": 340
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 0.00015840096268737977,
+      "loss": 1.144,
+      "step": 350
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 0.00015916271498355564,
+      "loss": 1.2408,
+      "step": 360
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 0.00015990359470573583,
+      "loss": 1.2195,
+      "step": 370
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 0.00016062471526133082,
+      "loss": 1.2085,
+      "step": 380
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 0.0001613271032784082,
+      "loss": 1.243,
+      "step": 390
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 0.0001620117073959503,
+      "loss": 1.1044,
+      "step": 400
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 0.00016267940596854117,
+      "loss": 1.2442,
+      "step": 410
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 0.00016333101384249078,
+      "loss": 1.1882,
+      "step": 420
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 0.00016396728833453322,
+      "loss": 1.1486,
+      "step": 430
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 0.00016458893452312522,
+      "loss": 1.2567,
+      "step": 440
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 0.00016519660994505735,
+      "loss": 1.164,
+      "step": 450
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 0.000165790928775822,
+      "loss": 1.2633,
+      "step": 460
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 0.00016637246556036958,
+      "loss": 1.1913,
+      "step": 470
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 0.00016694175855106128,
+      "loss": 1.1477,
+      "step": 480
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 0.00016749931270142591,
+      "loss": 1.2731,
+      "step": 490
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 0.000168045602357452,
+      "loss": 1.1511,
+      "step": 500
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 0.00016858107368236008,
+      "loss": 1.2301,
+      "step": 510
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 0.00016910614684591385,
+      "loss": 1.1907,
+      "step": 520
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 0.00016962121800518536,
+      "loss": 1.1612,
+      "step": 530
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 0.00017012666110016836,
+      "loss": 1.2319,
+      "step": 540
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 0.00017062282948462692,
+      "loss": 1.1193,
+      "step": 550
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 0.00017111005740999641,
+      "loss": 1.2784,
+      "step": 560
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 0.00017158866137794354,
+      "loss": 1.1673,
+      "step": 570
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 0.00017205894137529426,
+      "loss": 1.1765,
+      "step": 580
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 0.00017252118200339632,
+      "loss": 1.2579,
+      "step": 590
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 0.00017297565351256298,
+      "loss": 1.1393,
+      "step": 600
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 0.00017342261275101354,
+      "loss": 1.2666,
+      "step": 610
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 0.00017386230403665357,
+      "loss": 1.2035,
+      "step": 620
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 0.0001742949599591035,
+      "loss": 1.1859,
+      "step": 630
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 0.00017472080211856694,
+      "loss": 1.2394,
+      "step": 640
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 0.00017514004180741556,
+      "loss": 1.1103,
+      "step": 650
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 0.00017555288063973793,
+      "loss": 1.1979,
+      "step": 660
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 0.00017595951113354868,
+      "loss": 1.1613,
+      "step": 670
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 0.0001763601172498657,
+      "loss": 1.1612,
+      "step": 680
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 0.0001767548748924347,
+      "loss": 1.2676,
+      "step": 690
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 0.00017714395237149812,
+      "loss": 1.1264,
+      "step": 700
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 0.00017752751083466962,
+      "loss": 1.2567,
+      "step": 710
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 0.000177905704667674,
+      "loss": 1.1648,
+      "step": 720
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 0.0001782786818674474,
+      "loss": 1.192,
+      "step": 730
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 0.00017864658438985419,
+      "loss": 1.2161,
+      "step": 740
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 0.00017900954847406472,
+      "loss": 1.1022,
+      "step": 750
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 0.0001793677049454492,
+      "loss": 1.3062,
+      "step": 760
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 0.0001797211794986731,
+      "loss": 1.1351,
+      "step": 770
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 0.00018007009296252654,
+      "loss": 1.209,
+      "step": 780
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 0.00018041456154788414,
+      "loss": 1.2362,
+      "step": 790
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 0.00018075469708006865,
+      "loss": 1.1302,
+      "step": 800
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 0.00018109060721678105,
+      "loss": 1.2922,
+      "step": 810
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 0.0001814223956526595,
+      "loss": 1.1519,
+      "step": 820
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 0.00018175016231143995,
+      "loss": 1.1629,
+      "step": 830
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 0.00018207400352660916,
+      "loss": 1.2504,
+      "step": 840
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 0.00018239401221136744,
+      "loss": 1.1551,
+      "step": 850
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 0.0001827102780186516,
+      "loss": 1.2317,
+      "step": 860
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 0.00018302288749190697,
+      "loss": 1.1559,
+      "step": 870
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 0.0001833319242072436,
+      "loss": 1.2097,
+      "step": 880
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 0.00018363746890756038,
+      "loss": 1.228,
+      "step": 890
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 0.0001839395996291757,
+      "loss": 1.1432,
+      "step": 900
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 0.0001842383918214617,
+      "loss": 1.1888,
+      "step": 910
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 0.00018453391845994036,
+      "loss": 1.1958,
+      "step": 920
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 0.00018482625015326626,
+      "loss": 1.148,
+      "step": 930
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 0.00018511545524448793,
+      "loss": 1.242,
+      "step": 940
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 0.0001854015999069509,
+      "loss": 1.1326,
+      "step": 950
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 0.00018568474823517966,
+      "loss": 1.2941,
+      "step": 960
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 0.0001859649623310504,
+      "loss": 1.1983,
+      "step": 970
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 0.00018624230238554427,
+      "loss": 1.156,
+      "step": 980
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 0.00018651682675635065,
+      "loss": 1.2216,
+      "step": 990
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 0.00018678859204157035,
+      "loss": 1.0838,
+      "step": 1000
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 0.00018705765314975205,
+      "loss": 1.2639,
+      "step": 1010
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 0.00018732406336647843,
+      "loss": 1.1877,
+      "step": 1020
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 0.00018758787441770382,
+      "loss": 1.1674,
+      "step": 1030
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 0.0001878491365300322,
+      "loss": 1.1946,
+      "step": 1040
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 0.00018810789848811086,
+      "loss": 1.121,
+      "step": 1050
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 0.00018836420768930372,
+      "loss": 1.2559,
+      "step": 1060
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 0.00018861811019579795,
+      "loss": 1.123,
+      "step": 1070
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 0.0001888696507842867,
+      "loss": 1.1994,
+      "step": 1080
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 0.00018911887299336228,
+      "loss": 1.2116,
+      "step": 1090
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 0.0001893658191687453,
+      "loss": 1.0746,
+      "step": 1100
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 0.0001896105305064669,
+      "loss": 1.2671,
+      "step": 1110
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 0.00018985304709411477,
+      "loss": 1.1409,
+      "step": 1120
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 0.00019009340795024567,
+      "loss": 1.1153,
+      "step": 1130
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 0.0001903316510620619,
+      "loss": 1.2162,
+      "step": 1140
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 0.00019056781342144207,
+      "loss": 1.103,
+      "step": 1150
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 0.0001908019310594126,
+      "loss": 1.2478,
+      "step": 1160
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 0.00019103403907913926,
+      "loss": 1.1723,
+      "step": 1170
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 0.0001912641716875147,
+      "loss": 1.1992,
+      "step": 1180
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 0.0001914923622254136,
+      "loss": 1.2087,
+      "step": 1190
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 0.00019171864319668136,
+      "loss": 1.1551,
+      "step": 1200
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 0.00019194304629592022,
+      "loss": 1.2317,
+      "step": 1210
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 0.00019216560243513192,
+      "loss": 1.1294,
+      "step": 1220
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 0.0001923863417692722,
+      "loss": 1.1429,
+      "step": 1230
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 0.00019260529372077195,
+      "loss": 1.2404,
+      "step": 1240
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 0.00019282248700307206,
+      "loss": 1.1336,
+      "step": 1250
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 0.00019303794964322185,
+      "loss": 1.2671,
+      "step": 1260
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 0.00019325170900358343,
+      "loss": 1.1229,
+      "step": 1270
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 0.0001934637918026853,
+      "loss": 1.1335,
+      "step": 1280
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 0.0001936742241352643,
+      "loss": 1.2139,
+      "step": 1290
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 0.0001938830314915339,
+      "loss": 1.0886,
+      "step": 1300
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 0.000194090238775714,
+      "loss": 1.2325,
+      "step": 1310
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 0.0001942958703238563,
+      "loss": 1.1449,
+      "step": 1320
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 0.00019449994992099703,
+      "loss": 1.1834,
+      "step": 1330
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 0.00019470250081766706,
+      "loss": 1.2746,
+      "step": 1340
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 0.00019490354574578842,
+      "loss": 1.103,
+      "step": 1350
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 0.0001951031069339841,
+      "loss": 1.2797,
+      "step": 1360
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 0.0001953012061223271,
+      "loss": 1.1534,
+      "step": 1370
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 0.00019549786457655305,
+      "loss": 1.1639,
+      "step": 1380
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 0.00019569310310175997,
+      "loss": 1.2115,
+      "step": 1390
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 0.00019588694205561647,
+      "loss": 1.0515,
+      "step": 1400
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 0.00019607940136110065,
+      "loss": 1.2054,
+      "step": 1410
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 0.000196270500518788,
+      "loss": 1.1422,
+      "step": 1420
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 0.00019646025861870886,
+      "loss": 1.1391,
+      "step": 1430
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 0.00019664869435179235,
+      "loss": 1.2412,
+      "step": 1440
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 0.00019683582602091432,
+      "loss": 1.1085,
+      "step": 1450
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 0.00019702167155156576,
+      "loss": 1.2386,
+      "step": 1460
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 0.00019720624850215698,
+      "loss": 1.1997,
+      "step": 1470
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 0.00019738957407397257,
+      "loss": 1.1984,
+      "step": 1480
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 0.00019757166512079108,
+      "loss": 1.2307,
+      "step": 1490
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 0.00019775253815818307,
+      "loss": 1.1343,
+      "step": 1500
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 0.0001979322093725002,
+      "loss": 1.2189,
+      "step": 1510
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 0.00019811069462956756,
+      "loss": 1.1581,
+      "step": 1520
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 0.00019828800948309115,
+      "loss": 1.1699,
+      "step": 1530
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 0.00019846416918279145,
+      "loss": 1.2407,
+      "step": 1540
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 0.00019863918868227366,
+      "loss": 1.0669,
+      "step": 1550
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 0.00019881308264664492,
+      "loss": 1.2039,
+      "step": 1560
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 0.0001989858654598883,
+      "loss": 1.1691,
+      "step": 1570
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 0.0001991575512320025,
+      "loss": 1.1569,
+      "step": 1580
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 0.00019932815380591643,
+      "loss": 1.2092,
+      "step": 1590
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 0.000199497686764187,
+      "loss": 1.128,
+      "step": 1600
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 0.00019966616343548824,
+      "loss": 1.2002,
+      "step": 1610
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 0.0001998335969008994,
+      "loss": 1.0644,
+      "step": 1620
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 0.0002,
+      "loss": 1.1275,
+      "step": 1630
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 0.0001999658307864614,
+      "loss": 1.1783,
+      "step": 1640
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 0.00019992786499364073,
+      "loss": 1.1262,
+      "step": 1650
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 0.00019988989920082006,
+      "loss": 1.2625,
+      "step": 1660
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 0.00019985193340799938,
+      "loss": 1.1807,
+      "step": 1670
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 0.00019981396761517873,
+      "loss": 1.1256,
+      "step": 1680
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 0.00019977600182235808,
+      "loss": 1.2082,
+      "step": 1690
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 0.0001997380360295374,
+      "loss": 1.0813,
+      "step": 1700
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 0.00019970007023671673,
+      "loss": 1.1981,
+      "step": 1710
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 0.00019966210444389605,
+      "loss": 1.1224,
+      "step": 1720
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 0.00019962413865107537,
+      "loss": 1.1375,
+      "step": 1730
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 0.00019958617285825472,
+      "loss": 1.1768,
+      "step": 1740
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 0.00019954820706543405,
+      "loss": 1.097,
+      "step": 1750
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 0.00019951024127261337,
+      "loss": 1.2163,
+      "step": 1760
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 0.00019947227547979272,
+      "loss": 1.1339,
+      "step": 1770
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 0.00019943430968697205,
+      "loss": 1.1616,
+      "step": 1780
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 0.0001993963438941514,
+      "loss": 1.2019,
+      "step": 1790
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 0.00019935837810133072,
+      "loss": 1.087,
+      "step": 1800
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 0.00019932041230851004,
+      "loss": 1.2352,
+      "step": 1810
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 0.00019928244651568937,
+      "loss": 1.1305,
+      "step": 1820
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 0.0001992444807228687,
+      "loss": 1.126,
+      "step": 1830
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 0.00019920651493004804,
+      "loss": 1.2041,
+      "step": 1840
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 0.0001991685491372274,
+      "loss": 1.0549,
+      "step": 1850
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 0.00019913058334440671,
+      "loss": 1.2373,
+      "step": 1860
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 0.00019909261755158604,
+      "loss": 1.1491,
+      "step": 1870
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 0.00019905465175876536,
+      "loss": 1.1715,
+      "step": 1880
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 0.00019901668596594468,
+      "loss": 1.2384,
+      "step": 1890
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 0.00019897872017312404,
+      "loss": 1.0732,
+      "step": 1900
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 0.00019894075438030336,
+      "loss": 1.2436,
+      "step": 1910
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 0.00019890278858748268,
+      "loss": 1.1645,
+      "step": 1920
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 0.000198864822794662,
+      "loss": 1.123,
+      "step": 1930
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 0.00019882685700184136,
+      "loss": 1.2387,
+      "step": 1940
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 0.00019878889120902068,
+      "loss": 1.0858,
+      "step": 1950
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 0.00019875092541620003,
+      "loss": 1.2401,
+      "step": 1960
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 0.00019871295962337935,
+      "loss": 1.1272,
+      "step": 1970
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 0.00019867499383055868,
+      "loss": 1.1403,
+      "step": 1980
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 0.000198637028037738,
+      "loss": 1.157,
+      "step": 1990
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 0.00019859906224491732,
+      "loss": 1.0744,
+      "step": 2000
+    },
+    {
+      "epoch": 0.11,
+      "eval_loss": 1.1491358280181885,
+      "eval_runtime": 1.9664,
+      "eval_samples_per_second": 50.854,
+      "eval_steps_per_second": 6.611,
+      "step": 2000
+    },
+    {
+      "epoch": 0.11,
+      "mmlu_eval_accuracy": NaN,
+      "mmlu_eval_accuracy_abstract_algebra": 0.5454545454545454,
+      "mmlu_eval_accuracy_anatomy": 0.21428571428571427,
+      "mmlu_eval_accuracy_astronomy": 0.125,
+      "mmlu_eval_accuracy_business_ethics": 0.45454545454545453,
+      "mmlu_eval_accuracy_clinical_knowledge": 0.4827586206896552,
+      "mmlu_eval_accuracy_college_biology": 0.75,
+      "mmlu_eval_accuracy_college_chemistry": 1.0,
+      "mmlu_eval_accuracy_college_computer_science": 0.6363636363636364,
+      "mmlu_eval_accuracy_college_mathematics": 0.18181818181818182,
+      "mmlu_eval_accuracy_college_medicine": 0.45454545454545453,
+      "mmlu_eval_accuracy_college_physics": 0.45454545454545453,
+      "mmlu_eval_accuracy_computer_security": 0.45454545454545453,
+      "mmlu_eval_accuracy_conceptual_physics": 0.42857142857142855,
+      "mmlu_eval_accuracy_econometrics": NaN,
+      "mmlu_eval_accuracy_electrical_engineering": NaN,
+      "mmlu_eval_accuracy_elementary_mathematics": NaN,
+      "mmlu_eval_accuracy_formal_logic": NaN,
+      "mmlu_eval_accuracy_global_facts": NaN,
+      "mmlu_eval_accuracy_high_school_biology": NaN,
+      "mmlu_eval_accuracy_high_school_chemistry": NaN,
+      "mmlu_eval_accuracy_high_school_computer_science": NaN,
+      "mmlu_eval_accuracy_high_school_european_history": NaN,
+      "mmlu_eval_accuracy_high_school_geography": NaN,
+      "mmlu_eval_accuracy_high_school_government_and_politics": NaN,
+      "mmlu_eval_accuracy_high_school_macroeconomics": NaN,
+      "mmlu_eval_accuracy_high_school_mathematics": NaN,
+      "mmlu_eval_accuracy_high_school_microeconomics": NaN,
+      "mmlu_eval_accuracy_high_school_physics": NaN,
+      "mmlu_eval_accuracy_high_school_psychology": NaN,
+      "mmlu_eval_accuracy_high_school_statistics": NaN,
+      "mmlu_eval_accuracy_high_school_us_history": NaN,
+      "mmlu_eval_accuracy_high_school_world_history": NaN,
+      "mmlu_eval_accuracy_human_aging": NaN,
+      "mmlu_eval_accuracy_human_sexuality": NaN,
+      "mmlu_eval_accuracy_international_law": NaN,
+      "mmlu_eval_accuracy_jurisprudence": NaN,
+      "mmlu_eval_accuracy_logical_fallacies": NaN,
+      "mmlu_eval_accuracy_machine_learning": NaN,
+      "mmlu_eval_accuracy_management": NaN,
+      "mmlu_eval_accuracy_marketing": NaN,
+      "mmlu_eval_accuracy_medical_genetics": NaN,
+      "mmlu_eval_accuracy_miscellaneous": NaN,
+      "mmlu_eval_accuracy_moral_disputes": NaN,
+      "mmlu_eval_accuracy_moral_scenarios": NaN,
+      "mmlu_eval_accuracy_nutrition": NaN,
+      "mmlu_eval_accuracy_philosophy": NaN,
+      "mmlu_eval_accuracy_prehistory": NaN,
+      "mmlu_eval_accuracy_professional_accounting": NaN,
+      "mmlu_eval_accuracy_professional_law": NaN,
+      "mmlu_eval_accuracy_professional_medicine": NaN,
+      "mmlu_eval_accuracy_professional_psychology": NaN,
+      "mmlu_eval_accuracy_public_relations": NaN,
+      "mmlu_eval_accuracy_security_studies": NaN,
+      "mmlu_eval_accuracy_sociology": NaN,
+      "mmlu_eval_accuracy_us_foreign_policy": NaN,
+      "mmlu_eval_accuracy_virology": NaN,
+      "mmlu_eval_accuracy_world_religions": NaN,
+      "mmlu_loss": 1.007582182297483,
+      "step": 2000
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 0.00019856109645209667,
+      "loss": 1.2053,
+      "step": 2010
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 0.000198523130659276,
+      "loss": 1.1405,
+      "step": 2020
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 0.00019848516486645535,
+      "loss": 1.1468,
+      "step": 2030
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 0.00019844719907363467,
+      "loss": 1.2189,
+      "step": 2040
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 0.000198409233280814,
+      "loss": 1.0986,
+      "step": 2050
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 0.00019837126748799332,
+      "loss": 1.1598,
+      "step": 2060
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 0.00019833330169517267,
+      "loss": 1.1875,
+      "step": 2070
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 0.000198295335902352,
+      "loss": 1.1328,
+      "step": 2080
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 0.00019825737010953132,
+      "loss": 1.1765,
+      "step": 2090
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 0.00019821940431671064,
+      "loss": 1.0558,
+      "step": 2100
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 0.00019818143852389,
+      "loss": 1.2161,
+      "step": 2110
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 0.0001981434727310693,
+      "loss": 1.149,
+      "step": 2120
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 0.00019810550693824866,
+      "loss": 1.1905,
+      "step": 2130
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 0.000198067541145428,
+      "loss": 1.2479,
+      "step": 2140
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 0.0001980295753526073,
+      "loss": 1.076,
+      "step": 2150
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 0.00019799160955978663,
+      "loss": 1.2602,
+      "step": 2160
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 0.00019795364376696596,
+      "loss": 1.1454,
+      "step": 2170
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 0.0001979156779741453,
+      "loss": 1.1334,
+      "step": 2180
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 0.00019787771218132463,
+      "loss": 1.1795,
+      "step": 2190
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 0.00019783974638850398,
+      "loss": 1.1137,
+      "step": 2200
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 0.0001978017805956833,
+      "loss": 1.2005,
+      "step": 2210
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 0.00019776381480286263,
+      "loss": 1.1376,
+      "step": 2220
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 0.00019772584901004195,
+      "loss": 1.166,
+      "step": 2230
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 0.0001976878832172213,
+      "loss": 1.192,
+      "step": 2240
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 0.00019764991742440063,
+      "loss": 1.0876,
+      "step": 2250
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 0.00019761195163157995,
+      "loss": 1.2518,
+      "step": 2260
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 0.00019757398583875927,
+      "loss": 1.196,
+      "step": 2270
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 0.0001975360200459386,
+      "loss": 1.2188,
+      "step": 2280
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 0.00019749805425311795,
+      "loss": 1.2075,
+      "step": 2290
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 0.0001974600884602973,
+      "loss": 1.0276,
+      "step": 2300
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 0.00019742212266747662,
+      "loss": 1.227,
+      "step": 2310
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 0.00019738415687465594,
+      "loss": 1.1681,
+      "step": 2320
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 0.00019734619108183527,
+      "loss": 1.1402,
+      "step": 2330
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 0.0001973082252890146,
+      "loss": 1.1977,
+      "step": 2340
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 0.00019727025949619394,
+      "loss": 1.0541,
+      "step": 2350
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 0.00019723229370337327,
+      "loss": 1.2152,
+      "step": 2360
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 0.00019719432791055262,
+      "loss": 1.1588,
+      "step": 2370
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 0.00019715636211773194,
+      "loss": 1.0637,
+      "step": 2380
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 0.00019711839632491126,
+      "loss": 1.2027,
+      "step": 2390
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 0.0001970804305320906,
+      "loss": 1.1089,
+      "step": 2400
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 0.00019704246473926994,
+      "loss": 1.2193,
+      "step": 2410
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 0.00019700449894644926,
+      "loss": 1.1809,
+      "step": 2420
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 0.00019696653315362858,
+      "loss": 1.1154,
+      "step": 2430
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 0.0001969285673608079,
+      "loss": 1.1509,
+      "step": 2440
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 0.00019689060156798723,
+      "loss": 1.1225,
+      "step": 2450
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 0.00019685263577516658,
+      "loss": 1.246,
+      "step": 2460
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 0.00019681466998234593,
+      "loss": 1.1565,
+      "step": 2470
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 0.00019677670418952526,
+      "loss": 1.1319,
+      "step": 2480
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 0.00019673873839670458,
+      "loss": 1.1855,
+      "step": 2490
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 0.0001967007726038839,
+      "loss": 1.142,
+      "step": 2500
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 0.00019666280681106323,
+      "loss": 1.2383,
+      "step": 2510
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 0.00019662484101824258,
+      "loss": 1.1175,
+      "step": 2520
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 0.0001965868752254219,
+      "loss": 1.1014,
+      "step": 2530
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 0.00019654890943260122,
+      "loss": 1.1959,
+      "step": 2540
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 0.00019651094363978057,
+      "loss": 1.0617,
+      "step": 2550
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 0.0001964729778469599,
+      "loss": 1.2567,
+      "step": 2560
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 0.00019643501205413922,
+      "loss": 1.1222,
+      "step": 2570
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 0.00019639704626131857,
+      "loss": 1.1147,
+      "step": 2580
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 0.0001963590804684979,
+      "loss": 1.2296,
+      "step": 2590
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 0.00019632111467567722,
+      "loss": 1.094,
+      "step": 2600
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 0.00019628314888285654,
+      "loss": 1.1776,
+      "step": 2610
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 0.00019624518309003586,
+      "loss": 1.1303,
+      "step": 2620
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 0.00019620721729721524,
+      "loss": 1.137,
+      "step": 2630
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 0.00019616925150439457,
+      "loss": 1.2335,
+      "step": 2640
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 0.0001961312857115739,
+      "loss": 1.0721,
+      "step": 2650
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 0.0001960933199187532,
+      "loss": 1.2737,
+      "step": 2660
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 0.00019605535412593254,
+      "loss": 1.0844,
+      "step": 2670
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 0.0001960173883331119,
+      "loss": 1.1289,
+      "step": 2680
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 0.0001959794225402912,
+      "loss": 1.1722,
+      "step": 2690
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 0.00019594145674747053,
+      "loss": 1.0891,
+      "step": 2700
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 0.00019590349095464986,
+      "loss": 1.2333,
+      "step": 2710
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 0.0001958655251618292,
+      "loss": 1.1545,
+      "step": 2720
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 0.00019582755936900853,
+      "loss": 1.1221,
+      "step": 2730
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 0.00019578959357618788,
+      "loss": 1.1968,
+      "step": 2740
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 0.0001957516277833672,
+      "loss": 1.159,
+      "step": 2750
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 0.00019571366199054653,
+      "loss": 1.2564,
+      "step": 2760
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 0.00019567569619772585,
+      "loss": 1.1335,
+      "step": 2770
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 0.00019563773040490518,
+      "loss": 1.1433,
+      "step": 2780
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 0.00019559976461208453,
+      "loss": 1.1722,
+      "step": 2790
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 0.00019556179881926385,
+      "loss": 1.0977,
+      "step": 2800
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 0.0001955238330264432,
+      "loss": 1.2189,
+      "step": 2810
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 0.00019548586723362252,
+      "loss": 1.1602,
+      "step": 2820
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 0.00019544790144080185,
+      "loss": 1.147,
+      "step": 2830
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 0.00019540993564798117,
+      "loss": 1.2212,
+      "step": 2840
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 0.00019537196985516052,
+      "loss": 1.1267,
+      "step": 2850
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 0.00019533400406233984,
+      "loss": 1.2403,
+      "step": 2860
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 0.00019529603826951917,
+      "loss": 1.13,
+      "step": 2870
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 0.0001952580724766985,
+      "loss": 1.1179,
+      "step": 2880
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 0.00019522010668387784,
+      "loss": 1.1641,
+      "step": 2890
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 0.00019518214089105716,
+      "loss": 1.0859,
+      "step": 2900
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 0.00019514417509823652,
+      "loss": 1.267,
+      "step": 2910
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 0.00019510620930541584,
+      "loss": 1.1309,
+      "step": 2920
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 0.00019506824351259516,
+      "loss": 1.0973,
+      "step": 2930
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 0.00019503027771977449,
+      "loss": 1.1753,
+      "step": 2940
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 0.0001949923119269538,
+      "loss": 1.0664,
+      "step": 2950
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 0.00019495434613413316,
+      "loss": 1.2195,
+      "step": 2960
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 0.00019491638034131248,
+      "loss": 1.1501,
+      "step": 2970
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 0.00019487841454849183,
+      "loss": 1.1266,
+      "step": 2980
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 0.00019484044875567116,
+      "loss": 1.1611,
+      "step": 2990
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 0.00019480248296285048,
+      "loss": 1.0625,
+      "step": 3000
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 0.0001947645171700298,
+      "loss": 1.2264,
+      "step": 3010
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 0.00019472655137720915,
+      "loss": 1.1014,
+      "step": 3020
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 0.00019468858558438848,
+      "loss": 1.1016,
+      "step": 3030
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 0.0001946506197915678,
+      "loss": 1.2034,
+      "step": 3040
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 0.00019461265399874712,
+      "loss": 1.0768,
+      "step": 3050
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 0.00019457468820592645,
+      "loss": 1.1682,
+      "step": 3060
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 0.0001945367224131058,
+      "loss": 1.1381,
+      "step": 3070
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 0.00019449875662028515,
+      "loss": 1.1447,
+      "step": 3080
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 0.00019446079082746447,
+      "loss": 1.2159,
+      "step": 3090
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 0.0001944228250346438,
+      "loss": 1.0857,
+      "step": 3100
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 0.00019438485924182312,
+      "loss": 1.2315,
+      "step": 3110
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 0.00019434689344900244,
+      "loss": 1.1214,
+      "step": 3120
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 0.0001943089276561818,
+      "loss": 1.1003,
+      "step": 3130
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 0.00019427096186336112,
+      "loss": 1.2121,
+      "step": 3140
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 0.00019423299607054047,
+      "loss": 1.0662,
+      "step": 3150
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 0.0001941950302777198,
+      "loss": 1.2078,
+      "step": 3160
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 0.00019415706448489911,
+      "loss": 1.1218,
+      "step": 3170
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 0.00019411909869207844,
+      "loss": 1.136,
+      "step": 3180
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 0.0001940811328992578,
+      "loss": 1.1991,
+      "step": 3190
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 0.0001940431671064371,
+      "loss": 1.0649,
+      "step": 3200
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 0.00019400520131361644,
+      "loss": 1.2392,
+      "step": 3210
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 0.00019396723552079576,
+      "loss": 1.118,
+      "step": 3220
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 0.00019392926972797508,
+      "loss": 1.1365,
+      "step": 3230
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 0.00019389130393515443,
+      "loss": 1.1871,
+      "step": 3240
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 0.00019385333814233378,
+      "loss": 1.0481,
+      "step": 3250
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 0.0001938153723495131,
+      "loss": 1.2319,
+      "step": 3260
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 0.00019377740655669243,
+      "loss": 1.1525,
+      "step": 3270
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 0.00019373944076387175,
+      "loss": 1.1213,
+      "step": 3280
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 0.00019370147497105108,
+      "loss": 1.1762,
+      "step": 3290
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 0.00019366350917823043,
+      "loss": 1.0419,
+      "step": 3300
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 0.00019362554338540975,
+      "loss": 1.2288,
+      "step": 3310
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 0.00019358757759258907,
+      "loss": 1.1777,
+      "step": 3320
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 0.00019354961179976843,
+      "loss": 1.1398,
+      "step": 3330
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 0.00019351164600694775,
+      "loss": 1.1758,
+      "step": 3340
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 0.00019347368021412707,
+      "loss": 1.0846,
+      "step": 3350
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 0.00019343571442130642,
+      "loss": 1.2201,
+      "step": 3360
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 0.00019339774862848575,
+      "loss": 1.1009,
+      "step": 3370
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 0.00019335978283566507,
+      "loss": 1.1371,
+      "step": 3380
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 0.0001933218170428444,
+      "loss": 1.204,
+      "step": 3390
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 0.00019328385125002372,
+      "loss": 1.0918,
+      "step": 3400
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 0.00019324588545720307,
+      "loss": 1.19,
+      "step": 3410
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 0.00019320791966438242,
+      "loss": 1.1501,
+      "step": 3420
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 0.00019316995387156174,
+      "loss": 1.1232,
+      "step": 3430
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 0.00019313198807874106,
+      "loss": 1.2088,
+      "step": 3440
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 0.0001930940222859204,
+      "loss": 1.0601,
+      "step": 3450
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 0.0001930560564930997,
+      "loss": 1.2256,
+      "step": 3460
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 0.00019301809070027906,
+      "loss": 1.0896,
+      "step": 3470
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 0.00019298012490745839,
+      "loss": 1.1774,
+      "step": 3480
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 0.0001929421591146377,
+      "loss": 1.1691,
+      "step": 3490
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 0.00019290419332181706,
+      "loss": 1.052,
+      "step": 3500
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 0.00019286622752899638,
+      "loss": 1.2578,
+      "step": 3510
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 0.0001928282617361757,
+      "loss": 1.14,
+      "step": 3520
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 0.00019279029594335506,
+      "loss": 1.1038,
+      "step": 3530
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 0.00019275233015053438,
+      "loss": 1.2139,
+      "step": 3540
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 0.0001927143643577137,
+      "loss": 1.0836,
+      "step": 3550
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 0.00019267639856489303,
+      "loss": 1.2205,
+      "step": 3560
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 0.00019263843277207235,
+      "loss": 1.1611,
+      "step": 3570
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 0.0001926004669792517,
+      "loss": 1.1044,
+      "step": 3580
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 0.00019256250118643105,
+      "loss": 1.176,
+      "step": 3590
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 0.00019252453539361037,
+      "loss": 1.088,
+      "step": 3600
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 0.0001924865696007897,
+      "loss": 1.2273,
+      "step": 3610
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 0.00019244860380796902,
+      "loss": 1.1066,
+      "step": 3620
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 0.00019241063801514837,
+      "loss": 1.1157,
+      "step": 3630
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 0.0001923726722223277,
+      "loss": 1.1903,
+      "step": 3640
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 0.00019233470642950702,
+      "loss": 1.0617,
+      "step": 3650
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 0.00019229674063668634,
+      "loss": 1.2024,
+      "step": 3660
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 0.0001922587748438657,
+      "loss": 1.1173,
+      "step": 3670
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 0.00019222080905104502,
+      "loss": 1.1281,
+      "step": 3680
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 0.00019218284325822437,
+      "loss": 1.1616,
+      "step": 3690
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 0.0001921448774654037,
+      "loss": 1.1023,
+      "step": 3700
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 0.00019210691167258301,
+      "loss": 1.2164,
+      "step": 3710
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 0.00019206894587976234,
+      "loss": 1.1053,
+      "step": 3720
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 0.00019203098008694166,
+      "loss": 1.172,
+      "step": 3730
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 0.000191993014294121,
+      "loss": 1.1863,
+      "step": 3740
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 0.00019195504850130033,
+      "loss": 1.0662,
+      "step": 3750
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 0.00019191708270847969,
+      "loss": 1.2229,
+      "step": 3760
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 0.000191879116915659,
+      "loss": 1.1553,
+      "step": 3770
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 0.00019184115112283833,
+      "loss": 1.1146,
+      "step": 3780
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 0.00019180318533001766,
+      "loss": 1.1671,
+      "step": 3790
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 0.000191765219537197,
+      "loss": 1.139,
+      "step": 3800
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 0.00019172725374437633,
+      "loss": 1.213,
+      "step": 3810
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 0.00019168928795155565,
+      "loss": 1.103,
+      "step": 3820
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 0.00019165132215873498,
+      "loss": 1.144,
+      "step": 3830
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 0.0001916133563659143,
+      "loss": 1.1856,
+      "step": 3840
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 0.00019157539057309365,
+      "loss": 1.0548,
+      "step": 3850
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 0.000191537424780273,
+      "loss": 1.2376,
+      "step": 3860
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 0.00019149945898745232,
+      "loss": 1.1467,
+      "step": 3870
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 0.00019146149319463165,
+      "loss": 1.1517,
+      "step": 3880
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 0.00019142352740181097,
+      "loss": 1.1989,
+      "step": 3890
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 0.0001913855616089903,
+      "loss": 1.0422,
+      "step": 3900
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 0.00019134759581616965,
+      "loss": 1.2195,
+      "step": 3910
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 0.00019130963002334897,
+      "loss": 1.1608,
+      "step": 3920
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 0.00019127166423052832,
+      "loss": 1.1162,
+      "step": 3930
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 0.00019123369843770764,
+      "loss": 1.1381,
+      "step": 3940
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 0.00019119573264488697,
+      "loss": 1.063,
+      "step": 3950
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 0.0001911577668520663,
+      "loss": 1.2331,
+      "step": 3960
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 0.00019111980105924564,
+      "loss": 1.1366,
+      "step": 3970
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 0.00019108183526642496,
+      "loss": 1.1444,
+      "step": 3980
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 0.0001910438694736043,
+      "loss": 1.1837,
+      "step": 3990
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 0.0001910059036807836,
+      "loss": 1.0373,
+      "step": 4000
+    },
+    {
+      "epoch": 0.22,
+      "eval_loss": 1.110398769378662,
+      "eval_runtime": 1.9777,
+      "eval_samples_per_second": 50.563,
+      "eval_steps_per_second": 6.573,
+      "step": 4000
+    },
+    {
+      "epoch": 0.22,
+      "mmlu_eval_accuracy": NaN,
+      "mmlu_eval_accuracy_abstract_algebra": 0.5454545454545454,
+      "mmlu_eval_accuracy_anatomy": 0.2857142857142857,
+      "mmlu_eval_accuracy_astronomy": 0.125,
+      "mmlu_eval_accuracy_business_ethics": 0.5454545454545454,
+      "mmlu_eval_accuracy_clinical_knowledge": 0.41379310344827586,
+      "mmlu_eval_accuracy_college_biology": 0.75,
+      "mmlu_eval_accuracy_college_chemistry": 1.0,
+      "mmlu_eval_accuracy_college_computer_science": 0.6363636363636364,
+      "mmlu_eval_accuracy_college_mathematics": 0.18181818181818182,
+      "mmlu_eval_accuracy_college_medicine": 0.45454545454545453,
+      "mmlu_eval_accuracy_college_physics": 0.5454545454545454,
+      "mmlu_eval_accuracy_computer_security": 0.5454545454545454,
+      "mmlu_eval_accuracy_conceptual_physics": 0.42857142857142855,
+      "mmlu_eval_accuracy_econometrics": NaN,
+      "mmlu_eval_accuracy_electrical_engineering": NaN,
+      "mmlu_eval_accuracy_elementary_mathematics": NaN,
+      "mmlu_eval_accuracy_formal_logic": NaN,
+      "mmlu_eval_accuracy_global_facts": NaN,
+      "mmlu_eval_accuracy_high_school_biology": NaN,
+      "mmlu_eval_accuracy_high_school_chemistry": NaN,
+      "mmlu_eval_accuracy_high_school_computer_science": NaN,
+      "mmlu_eval_accuracy_high_school_european_history": NaN,
+      "mmlu_eval_accuracy_high_school_geography": NaN,
+      "mmlu_eval_accuracy_high_school_government_and_politics": NaN,
+      "mmlu_eval_accuracy_high_school_macroeconomics": NaN,
+      "mmlu_eval_accuracy_high_school_mathematics": NaN,
+      "mmlu_eval_accuracy_high_school_microeconomics": NaN,
+      "mmlu_eval_accuracy_high_school_physics": NaN,
+      "mmlu_eval_accuracy_high_school_psychology": NaN,
+      "mmlu_eval_accuracy_high_school_statistics": NaN,
+      "mmlu_eval_accuracy_high_school_us_history": NaN,
+      "mmlu_eval_accuracy_high_school_world_history": NaN,
+      "mmlu_eval_accuracy_human_aging": NaN,
+      "mmlu_eval_accuracy_human_sexuality": NaN,
+      "mmlu_eval_accuracy_international_law": NaN,
+      "mmlu_eval_accuracy_jurisprudence": NaN,
+      "mmlu_eval_accuracy_logical_fallacies": NaN,
+      "mmlu_eval_accuracy_machine_learning": NaN,
+      "mmlu_eval_accuracy_management": NaN,
+      "mmlu_eval_accuracy_marketing": NaN,
+      "mmlu_eval_accuracy_medical_genetics": NaN,
+      "mmlu_eval_accuracy_miscellaneous": NaN,
+      "mmlu_eval_accuracy_moral_disputes": NaN,
+      "mmlu_eval_accuracy_moral_scenarios": NaN,
+      "mmlu_eval_accuracy_nutrition": NaN,
+      "mmlu_eval_accuracy_philosophy": NaN,
+      "mmlu_eval_accuracy_prehistory": NaN,
+      "mmlu_eval_accuracy_professional_accounting": NaN,
+      "mmlu_eval_accuracy_professional_law": NaN,
+      "mmlu_eval_accuracy_professional_medicine": NaN,
+      "mmlu_eval_accuracy_professional_psychology": NaN,
+      "mmlu_eval_accuracy_public_relations": NaN,
+      "mmlu_eval_accuracy_security_studies": NaN,
+      "mmlu_eval_accuracy_sociology": NaN,
+      "mmlu_eval_accuracy_us_foreign_policy": NaN,
+      "mmlu_eval_accuracy_virology": NaN,
+      "mmlu_eval_accuracy_world_religions": NaN,
+      "mmlu_loss": 1.2074859381342928,
+      "step": 4000
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 0.00019096793788796293,
+      "loss": 1.2292,
+      "step": 4010
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 0.00019092997209514228,
+      "loss": 1.085,
+      "step": 4020
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 0.00019089200630232164,
+      "loss": 1.1231,
+      "step": 4030
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 0.00019085404050950096,
+      "loss": 1.2231,
+      "step": 4040
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 0.00019081607471668028,
+      "loss": 1.0526,
+      "step": 4050
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 0.0001907781089238596,
+      "loss": 1.2899,
+      "step": 4060
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 0.00019074014313103893,
+      "loss": 1.1367,
+      "step": 4070
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 0.00019070217733821828,
+      "loss": 1.1275,
+      "step": 4080
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 0.0001906642115453976,
+      "loss": 1.2279,
+      "step": 4090
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 0.00019062624575257693,
+      "loss": 1.0699,
+      "step": 4100
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 0.00019058827995975628,
+      "loss": 1.1987,
+      "step": 4110
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 0.0001905503141669356,
+      "loss": 1.1191,
+      "step": 4120
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 0.00019051234837411492,
+      "loss": 1.1265,
+      "step": 4130
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 0.00019047438258129427,
+      "loss": 1.1401,
+      "step": 4140
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 0.0001904364167884736,
+      "loss": 1.0537,
+      "step": 4150
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 0.00019039845099565292,
+      "loss": 1.1899,
+      "step": 4160
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 0.00019036048520283224,
+      "loss": 1.1534,
+      "step": 4170
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 0.00019032251941001157,
+      "loss": 1.1586,
+      "step": 4180
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 0.00019028455361719092,
+      "loss": 1.1836,
+      "step": 4190
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 0.00019024658782437027,
+      "loss": 1.0386,
+      "step": 4200
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 0.0001902086220315496,
+      "loss": 1.2267,
+      "step": 4210
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 0.00019017065623872892,
+      "loss": 1.1185,
+      "step": 4220
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 0.00019013269044590824,
+      "loss": 1.1282,
+      "step": 4230
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 0.00019009472465308756,
+      "loss": 1.1845,
+      "step": 4240
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 0.0001900567588602669,
+      "loss": 1.0246,
+      "step": 4250
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 0.00019001879306744624,
+      "loss": 1.2289,
+      "step": 4260
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 0.00018998082727462556,
+      "loss": 1.1126,
+      "step": 4270
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 0.0001899428614818049,
+      "loss": 1.1857,
+      "step": 4280
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 0.00018990489568898423,
+      "loss": 1.1578,
+      "step": 4290
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 0.00018986692989616356,
+      "loss": 1.0807,
+      "step": 4300
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 0.0001898289641033429,
+      "loss": 1.1825,
+      "step": 4310
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 0.00018979099831052223,
+      "loss": 1.1619,
+      "step": 4320
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 0.00018975303251770156,
+      "loss": 1.0785,
+      "step": 4330
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 0.00018971506672488088,
+      "loss": 1.1788,
+      "step": 4340
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 0.0001896771009320602,
+      "loss": 1.0419,
+      "step": 4350
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 0.00018963913513923955,
+      "loss": 1.1988,
+      "step": 4360
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 0.0001896011693464189,
+      "loss": 1.0989,
+      "step": 4370
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 0.00018956320355359823,
+      "loss": 1.103,
+      "step": 4380
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 0.00018952523776077755,
+      "loss": 1.2271,
+      "step": 4390
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 0.00018948727196795687,
+      "loss": 1.0848,
+      "step": 4400
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 0.0001894493061751362,
+      "loss": 1.244,
+      "step": 4410
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 0.00018941134038231555,
+      "loss": 1.1194,
+      "step": 4420
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 0.00018937337458949487,
+      "loss": 1.1216,
+      "step": 4430
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 0.0001893354087966742,
+      "loss": 1.2267,
+      "step": 4440
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 0.00018929744300385354,
+      "loss": 1.0305,
+      "step": 4450
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 0.00018925947721103287,
+      "loss": 1.2269,
+      "step": 4460
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 0.0001892215114182122,
+      "loss": 1.0973,
+      "step": 4470
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 0.00018918354562539154,
+      "loss": 1.1478,
+      "step": 4480
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 0.00018914557983257087,
+      "loss": 1.1812,
+      "step": 4490
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 0.0001891076140397502,
+      "loss": 1.071,
+      "step": 4500
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 0.0001890696482469295,
+      "loss": 1.2323,
+      "step": 4510
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 0.00018903168245410884,
+      "loss": 1.1268,
+      "step": 4520
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 0.0001889937166612882,
+      "loss": 1.1018,
+      "step": 4530
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 0.00018895575086846754,
+      "loss": 1.2016,
+      "step": 4540
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 0.00018891778507564686,
+      "loss": 1.083,
+      "step": 4550
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 0.00018887981928282618,
+      "loss": 1.2251,
+      "step": 4560
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 0.0001888418534900055,
+      "loss": 1.1421,
+      "step": 4570
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 0.00018880388769718486,
+      "loss": 1.1455,
+      "step": 4580
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 0.00018876592190436418,
+      "loss": 1.1418,
+      "step": 4590
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 0.0001887279561115435,
+      "loss": 1.0783,
+      "step": 4600
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 0.00018868999031872283,
+      "loss": 1.187,
+      "step": 4610
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 0.00018865202452590215,
+      "loss": 1.1095,
+      "step": 4620
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 0.0001886140587330815,
+      "loss": 1.1307,
+      "step": 4630
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 0.00018857609294026085,
+      "loss": 1.1626,
+      "step": 4640
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 0.00018853812714744018,
+      "loss": 1.0711,
+      "step": 4650
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 0.0001885001613546195,
+      "loss": 1.1894,
+      "step": 4660
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 0.00018846219556179882,
+      "loss": 1.1372,
+      "step": 4670
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 0.00018842422976897815,
+      "loss": 1.1343,
+      "step": 4680
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 0.0001883862639761575,
+      "loss": 1.1851,
+      "step": 4690
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 0.00018834829818333682,
+      "loss": 1.0585,
+      "step": 4700
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 0.00018831033239051617,
+      "loss": 1.1874,
+      "step": 4710
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 0.0001882723665976955,
+      "loss": 1.1464,
+      "step": 4720
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 0.00018823440080487482,
+      "loss": 1.1317,
+      "step": 4730
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 0.00018819643501205414,
+      "loss": 1.2306,
+      "step": 4740
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 0.0001881584692192335,
+      "loss": 1.0564,
+      "step": 4750
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 0.00018812050342641282,
+      "loss": 1.209,
+      "step": 4760
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 0.00018808253763359214,
+      "loss": 1.1539,
+      "step": 4770
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 0.00018804457184077146,
+      "loss": 1.1303,
+      "step": 4780
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 0.00018800660604795079,
+      "loss": 1.2107,
+      "step": 4790
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 0.00018796864025513014,
+      "loss": 1.0294,
+      "step": 4800
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 0.0001879306744623095,
+      "loss": 1.234,
+      "step": 4810
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 0.0001878927086694888,
+      "loss": 1.1464,
+      "step": 4820
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 0.00018785474287666813,
+      "loss": 1.107,
+      "step": 4830
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 0.00018781677708384746,
+      "loss": 1.1833,
+      "step": 4840
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 0.00018777881129102678,
+      "loss": 1.0265,
+      "step": 4850
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 0.00018774084549820613,
+      "loss": 1.2075,
+      "step": 4860
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 0.00018770287970538545,
+      "loss": 1.1351,
+      "step": 4870
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 0.00018766491391256478,
+      "loss": 1.1321,
+      "step": 4880
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 0.00018762694811974413,
+      "loss": 1.1653,
+      "step": 4890
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 0.00018758898232692345,
+      "loss": 1.0572,
+      "step": 4900
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 0.00018755101653410278,
+      "loss": 1.2171,
+      "step": 4910
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 0.00018751305074128213,
+      "loss": 1.098,
+      "step": 4920
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 0.00018747508494846145,
+      "loss": 1.1458,
+      "step": 4930
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 0.00018743711915564077,
+      "loss": 1.1418,
+      "step": 4940
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 0.0001873991533628201,
+      "loss": 1.0089,
+      "step": 4950
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 0.00018736118756999942,
+      "loss": 1.1623,
+      "step": 4960
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 0.00018732322177717877,
+      "loss": 1.1495,
+      "step": 4970
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 0.00018728525598435812,
+      "loss": 1.1035,
+      "step": 4980
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 0.00018724729019153744,
+      "loss": 1.161,
+      "step": 4990
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 0.00018720932439871677,
+      "loss": 1.0411,
+      "step": 5000
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 0.0001871713586058961,
+      "loss": 1.2324,
+      "step": 5010
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 0.00018713339281307541,
+      "loss": 1.1385,
+      "step": 5020
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 0.00018709542702025477,
+      "loss": 1.1043,
+      "step": 5030
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 0.0001870574612274341,
+      "loss": 1.181,
+      "step": 5040
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 0.0001870194954346134,
+      "loss": 1.0469,
+      "step": 5050
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 0.00018698152964179276,
+      "loss": 1.2059,
+      "step": 5060
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 0.00018694356384897209,
+      "loss": 1.1162,
+      "step": 5070
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 0.0001869055980561514,
+      "loss": 1.1082,
+      "step": 5080
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 0.00018686763226333076,
+      "loss": 1.1686,
+      "step": 5090
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 0.00018682966647051008,
+      "loss": 1.0448,
+      "step": 5100
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 0.0001867917006776894,
+      "loss": 1.1741,
+      "step": 5110
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 0.00018675373488486873,
+      "loss": 1.0716,
+      "step": 5120
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 0.00018671576909204805,
+      "loss": 1.1176,
+      "step": 5130
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 0.0001866778032992274,
+      "loss": 1.1733,
+      "step": 5140
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 0.00018663983750640675,
+      "loss": 1.0743,
+      "step": 5150
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 0.00018660187171358608,
+      "loss": 1.2193,
+      "step": 5160
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 0.0001865639059207654,
+      "loss": 1.1102,
+      "step": 5170
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 0.00018652594012794473,
+      "loss": 1.121,
+      "step": 5180
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 0.00018648797433512405,
+      "loss": 1.1776,
+      "step": 5190
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 0.0001864500085423034,
+      "loss": 1.0163,
+      "step": 5200
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 0.00018641204274948272,
+      "loss": 1.2541,
+      "step": 5210
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 0.00018637407695666205,
+      "loss": 1.0853,
+      "step": 5220
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 0.0001863361111638414,
+      "loss": 1.1733,
+      "step": 5230
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 0.00018629814537102072,
+      "loss": 1.2336,
+      "step": 5240
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 0.00018626017957820004,
+      "loss": 1.0562,
+      "step": 5250
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 0.0001862222137853794,
+      "loss": 1.2348,
+      "step": 5260
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 0.00018618424799255872,
+      "loss": 1.1489,
+      "step": 5270
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 0.00018614628219973804,
+      "loss": 1.15,
+      "step": 5280
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 0.00018610831640691736,
+      "loss": 1.1572,
+      "step": 5290
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 0.0001860703506140967,
+      "loss": 1.0913,
+      "step": 5300
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 0.00018603238482127604,
+      "loss": 1.1969,
+      "step": 5310
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 0.0001859944190284554,
+      "loss": 1.0944,
+      "step": 5320
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 0.0001859564532356347,
+      "loss": 1.1065,
+      "step": 5330
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 0.00018591848744281404,
+      "loss": 1.1457,
+      "step": 5340
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 0.00018588052164999336,
+      "loss": 1.0607,
+      "step": 5350
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 0.00018584255585717268,
+      "loss": 1.2277,
+      "step": 5360
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 0.00018580459006435203,
+      "loss": 1.1081,
+      "step": 5370
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 0.00018576662427153136,
+      "loss": 1.1414,
+      "step": 5380
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 0.00018572865847871068,
+      "loss": 1.2183,
+      "step": 5390
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 0.00018569069268589003,
+      "loss": 1.0255,
+      "step": 5400
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 0.00018565272689306935,
+      "loss": 1.1896,
+      "step": 5410
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 0.00018561476110024868,
+      "loss": 1.127,
+      "step": 5420
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 0.00018557679530742803,
+      "loss": 1.1278,
+      "step": 5430
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 0.00018553882951460735,
+      "loss": 1.2204,
+      "step": 5440
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 0.00018550086372178667,
+      "loss": 1.0458,
+      "step": 5450
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 0.000185462897928966,
+      "loss": 1.2106,
+      "step": 5460
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 0.00018542493213614532,
+      "loss": 1.145,
+      "step": 5470
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 0.00018538696634332467,
+      "loss": 1.1016,
+      "step": 5480
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 0.00018534900055050402,
+      "loss": 1.1442,
+      "step": 5490
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 0.00018531103475768335,
+      "loss": 1.0923,
+      "step": 5500
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 0.00018527306896486267,
+      "loss": 1.1807,
+      "step": 5510
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 0.000185235103172042,
+      "loss": 1.1165,
+      "step": 5520
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 0.00018519713737922134,
+      "loss": 1.148,
+      "step": 5530
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 0.00018515917158640067,
+      "loss": 1.193,
+      "step": 5540
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 0.00018512120579358,
+      "loss": 1.0173,
+      "step": 5550
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 0.00018508324000075931,
+      "loss": 1.2175,
+      "step": 5560
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 0.00018504527420793864,
+      "loss": 1.1131,
+      "step": 5570
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 0.000185007308415118,
+      "loss": 1.1276,
+      "step": 5580
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 0.00018496934262229734,
+      "loss": 1.1736,
+      "step": 5590
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 0.00018493137682947666,
+      "loss": 1.0398,
+      "step": 5600
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 0.00018489341103665599,
+      "loss": 1.1924,
+      "step": 5610
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 0.0001848554452438353,
+      "loss": 1.1046,
+      "step": 5620
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 0.00018481747945101463,
+      "loss": 1.1178,
+      "step": 5630
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 0.00018477951365819398,
+      "loss": 1.1669,
+      "step": 5640
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 0.0001847415478653733,
+      "loss": 1.0463,
+      "step": 5650
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 0.00018470358207255266,
+      "loss": 1.2307,
+      "step": 5660
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 0.00018466561627973198,
+      "loss": 1.0907,
+      "step": 5670
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 0.0001846276504869113,
+      "loss": 1.1094,
+      "step": 5680
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 0.00018458968469409063,
+      "loss": 1.1678,
+      "step": 5690
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 0.00018455171890126998,
+      "loss": 1.0547,
+      "step": 5700
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 0.0001845137531084493,
+      "loss": 1.2321,
+      "step": 5710
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 0.00018447578731562862,
+      "loss": 1.1277,
+      "step": 5720
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 0.00018443782152280795,
+      "loss": 1.1338,
+      "step": 5730
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 0.00018439985572998727,
+      "loss": 1.2256,
+      "step": 5740
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 0.00018436188993716662,
+      "loss": 1.0613,
+      "step": 5750
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 0.00018432392414434597,
+      "loss": 1.2168,
+      "step": 5760
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 0.0001842859583515253,
+      "loss": 1.0944,
+      "step": 5770
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 0.00018424799255870462,
+      "loss": 1.1116,
+      "step": 5780
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 0.00018421002676588394,
+      "loss": 1.2336,
+      "step": 5790
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 0.00018417206097306327,
+      "loss": 1.0786,
+      "step": 5800
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 0.00018413409518024262,
+      "loss": 1.2013,
+      "step": 5810
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 0.00018409612938742194,
+      "loss": 1.0692,
+      "step": 5820
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 0.00018405816359460126,
+      "loss": 1.1132,
+      "step": 5830
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 0.00018402019780178061,
+      "loss": 1.1931,
+      "step": 5840
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 0.00018398223200895994,
+      "loss": 1.0281,
+      "step": 5850
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 0.00018394426621613926,
+      "loss": 1.2425,
+      "step": 5860
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 0.0001839063004233186,
+      "loss": 1.149,
+      "step": 5870
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 0.00018386833463049793,
+      "loss": 1.1379,
+      "step": 5880
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 0.00018383036883767726,
+      "loss": 1.1725,
+      "step": 5890
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 0.00018379240304485658,
+      "loss": 1.0951,
+      "step": 5900
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 0.0001837544372520359,
+      "loss": 1.195,
+      "step": 5910
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 0.00018371647145921526,
+      "loss": 1.1485,
+      "step": 5920
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 0.0001836785056663946,
+      "loss": 1.1079,
+      "step": 5930
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 0.00018364053987357393,
+      "loss": 1.1965,
+      "step": 5940
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 0.00018360257408075325,
+      "loss": 1.0699,
+      "step": 5950
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 0.00018356460828793258,
+      "loss": 1.1892,
+      "step": 5960
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 0.0001835266424951119,
+      "loss": 1.1574,
+      "step": 5970
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 0.00018348867670229125,
+      "loss": 1.1089,
+      "step": 5980
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 0.00018345071090947057,
+      "loss": 1.1851,
+      "step": 5990
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 0.0001834127451166499,
+      "loss": 1.0495,
+      "step": 6000
+    },
+    {
+      "epoch": 0.33,
+      "eval_loss": 1.1018731594085693,
+      "eval_runtime": 1.9641,
+      "eval_samples_per_second": 50.913,
+      "eval_steps_per_second": 6.619,
+      "step": 6000
+    },
+    {
+      "epoch": 0.33,
+      "mmlu_eval_accuracy": NaN,
+      "mmlu_eval_accuracy_abstract_algebra": 0.5454545454545454,
+      "mmlu_eval_accuracy_anatomy": 0.2857142857142857,
+      "mmlu_eval_accuracy_astronomy": 0.0625,
+      "mmlu_eval_accuracy_business_ethics": 0.5454545454545454,
+      "mmlu_eval_accuracy_clinical_knowledge": 0.5172413793103449,
+      "mmlu_eval_accuracy_college_biology": 0.75,
+      "mmlu_eval_accuracy_college_chemistry": 1.0,
+      "mmlu_eval_accuracy_college_computer_science": 0.6363636363636364,
+      "mmlu_eval_accuracy_college_mathematics": 0.18181818181818182,
+      "mmlu_eval_accuracy_college_medicine": 0.45454545454545453,
+      "mmlu_eval_accuracy_college_physics": 0.5454545454545454,
+      "mmlu_eval_accuracy_computer_security": 0.5454545454545454,
+      "mmlu_eval_accuracy_conceptual_physics": 0.3333333333333333,
+      "mmlu_eval_accuracy_econometrics": NaN,
+      "mmlu_eval_accuracy_electrical_engineering": NaN,
+      "mmlu_eval_accuracy_elementary_mathematics": NaN,
+      "mmlu_eval_accuracy_formal_logic": NaN,
+      "mmlu_eval_accuracy_global_facts": NaN,
+      "mmlu_eval_accuracy_high_school_biology": NaN,
+      "mmlu_eval_accuracy_high_school_chemistry": NaN,
+      "mmlu_eval_accuracy_high_school_computer_science": NaN,
+      "mmlu_eval_accuracy_high_school_european_history": NaN,
+      "mmlu_eval_accuracy_high_school_geography": NaN,
+      "mmlu_eval_accuracy_high_school_government_and_politics": NaN,
+      "mmlu_eval_accuracy_high_school_macroeconomics": NaN,
+      "mmlu_eval_accuracy_high_school_mathematics": NaN,
+      "mmlu_eval_accuracy_high_school_microeconomics": NaN,
+      "mmlu_eval_accuracy_high_school_physics": NaN,
+      "mmlu_eval_accuracy_high_school_psychology": NaN,
+      "mmlu_eval_accuracy_high_school_statistics": NaN,
+      "mmlu_eval_accuracy_high_school_us_history": NaN,
+      "mmlu_eval_accuracy_high_school_world_history": NaN,
+      "mmlu_eval_accuracy_human_aging": NaN,
+      "mmlu_eval_accuracy_human_sexuality": NaN,
+      "mmlu_eval_accuracy_international_law": NaN,
+      "mmlu_eval_accuracy_jurisprudence": NaN,
+      "mmlu_eval_accuracy_logical_fallacies": NaN,
+      "mmlu_eval_accuracy_machine_learning": NaN,
+      "mmlu_eval_accuracy_management": NaN,
+      "mmlu_eval_accuracy_marketing": NaN,
+      "mmlu_eval_accuracy_medical_genetics": NaN,
+      "mmlu_eval_accuracy_miscellaneous": NaN,
+      "mmlu_eval_accuracy_moral_disputes": NaN,
+      "mmlu_eval_accuracy_moral_scenarios": NaN,
+      "mmlu_eval_accuracy_nutrition": NaN,
+      "mmlu_eval_accuracy_philosophy": NaN,
+      "mmlu_eval_accuracy_prehistory": NaN,
+      "mmlu_eval_accuracy_professional_accounting": NaN,
+      "mmlu_eval_accuracy_professional_law": NaN,
+      "mmlu_eval_accuracy_professional_medicine": NaN,
+      "mmlu_eval_accuracy_professional_psychology": NaN,
+      "mmlu_eval_accuracy_public_relations": NaN,
+      "mmlu_eval_accuracy_security_studies": NaN,
+      "mmlu_eval_accuracy_sociology": NaN,
+      "mmlu_eval_accuracy_us_foreign_policy": NaN,
+      "mmlu_eval_accuracy_virology": NaN,
+      "mmlu_eval_accuracy_world_religions": NaN,
+      "mmlu_loss": 1.2562693644625444,
+      "step": 6000
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 0.00018337477932382925,
+      "loss": 1.2248,
+      "step": 6010
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 0.00018333681353100857,
+      "loss": 1.1279,
+      "step": 6020
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 0.0001832988477381879,
+      "loss": 1.0738,
+      "step": 6030
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 0.00018326088194536725,
+      "loss": 1.1895,
+      "step": 6040
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 0.00018322291615254657,
+      "loss": 1.0129,
+      "step": 6050
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 0.0001831849503597259,
+      "loss": 1.2221,
+      "step": 6060
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 0.00018314698456690522,
+      "loss": 1.1028,
+      "step": 6070
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 0.00018310901877408454,
+      "loss": 1.102,
+      "step": 6080
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 0.0001830710529812639,
+      "loss": 1.1225,
+      "step": 6090
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 0.00018303308718844324,
+      "loss": 1.0915,
+      "step": 6100
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 0.00018299512139562256,
+      "loss": 1.1598,
+      "step": 6110
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 0.0001829571556028019,
+      "loss": 1.1404,
+      "step": 6120
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 0.0001829191898099812,
+      "loss": 1.0737,
+      "step": 6130
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 0.00018288122401716053,
+      "loss": 1.1347,
+      "step": 6140
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 0.00018284325822433988,
+      "loss": 1.0211,
+      "step": 6150
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 0.0001828052924315192,
+      "loss": 1.2267,
+      "step": 6160
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 0.00018276732663869853,
+      "loss": 1.1446,
+      "step": 6170
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 0.00018272936084587788,
+      "loss": 1.1504,
+      "step": 6180
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 0.0001826913950530572,
+      "loss": 1.1443,
+      "step": 6190
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 0.00018265342926023653,
+      "loss": 1.0797,
+      "step": 6200
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 0.00018261546346741588,
+      "loss": 1.1665,
+      "step": 6210
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 0.0001825774976745952,
+      "loss": 1.0945,
+      "step": 6220
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 0.00018253953188177453,
+      "loss": 1.1407,
+      "step": 6230
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 0.00018250156608895385,
+      "loss": 1.1482,
+      "step": 6240
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 0.00018246360029613317,
+      "loss": 1.0432,
+      "step": 6250
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 0.00018242563450331252,
+      "loss": 1.2223,
+      "step": 6260
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 0.00018238766871049187,
+      "loss": 1.1428,
+      "step": 6270
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 0.0001823497029176712,
+      "loss": 1.0959,
+      "step": 6280
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 0.00018231173712485052,
+      "loss": 1.1745,
+      "step": 6290
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 0.00018227377133202984,
+      "loss": 1.0914,
+      "step": 6300
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 0.00018223580553920917,
+      "loss": 1.1881,
+      "step": 6310
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 0.00018219783974638852,
+      "loss": 1.1077,
+      "step": 6320
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 0.00018215987395356784,
+      "loss": 1.0773,
+      "step": 6330
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 0.00018212190816074717,
+      "loss": 1.1642,
+      "step": 6340
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 0.0001820839423679265,
+      "loss": 1.0321,
+      "step": 6350
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 0.00018204597657510584,
+      "loss": 1.168,
+      "step": 6360
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 0.0001820080107822852,
+      "loss": 1.1029,
+      "step": 6370
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 0.0001819700449894645,
+      "loss": 1.1007,
+      "step": 6380
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 0.00018193207919664384,
+      "loss": 1.2254,
+      "step": 6390
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 0.00018189411340382316,
+      "loss": 1.0357,
+      "step": 6400
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 0.00018185614761100248,
+      "loss": 1.2457,
+      "step": 6410
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 0.00018181818181818183,
+      "loss": 1.1002,
+      "step": 6420
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 0.00018178021602536116,
+      "loss": 1.1173,
+      "step": 6430
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 0.0001817422502325405,
+      "loss": 1.1747,
+      "step": 6440
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 0.00018170428443971983,
+      "loss": 1.0586,
+      "step": 6450
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 0.00018166631864689916,
+      "loss": 1.218,
+      "step": 6460
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 0.00018162835285407848,
+      "loss": 1.1095,
+      "step": 6470
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 0.00018159038706125783,
+      "loss": 1.1188,
+      "step": 6480
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 0.00018155242126843715,
+      "loss": 1.1898,
+      "step": 6490
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 0.00018151445547561648,
+      "loss": 1.0182,
+      "step": 6500
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 0.0001814764896827958,
+      "loss": 1.2594,
+      "step": 6510
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 0.00018143852388997512,
+      "loss": 1.0942,
+      "step": 6520
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 0.00018140055809715447,
+      "loss": 1.1292,
+      "step": 6530
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 0.00018136259230433382,
+      "loss": 1.1962,
+      "step": 6540
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 0.00018132462651151315,
+      "loss": 1.0887,
+      "step": 6550
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 0.00018128666071869247,
+      "loss": 1.148,
+      "step": 6560
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 0.0001812486949258718,
+      "loss": 1.0687,
+      "step": 6570
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 0.00018121072913305112,
+      "loss": 1.1146,
+      "step": 6580
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 0.00018117276334023047,
+      "loss": 1.1275,
+      "step": 6590
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 0.0001811347975474098,
+      "loss": 1.0666,
+      "step": 6600
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 0.00018109683175458912,
+      "loss": 1.2092,
+      "step": 6610
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 0.00018105886596176847,
+      "loss": 1.1358,
+      "step": 6620
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 0.0001810209001689478,
+      "loss": 1.203,
+      "step": 6630
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 0.0001809829343761271,
+      "loss": 1.1757,
+      "step": 6640
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 0.00018094496858330646,
+      "loss": 1.0754,
+      "step": 6650
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 0.0001809070027904858,
+      "loss": 1.2037,
+      "step": 6660
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 0.0001808690369976651,
+      "loss": 1.1305,
+      "step": 6670
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 0.00018083107120484443,
+      "loss": 1.0914,
+      "step": 6680
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 0.00018079310541202376,
+      "loss": 1.1591,
+      "step": 6690
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 0.0001807551396192031,
+      "loss": 1.0899,
+      "step": 6700
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 0.00018071717382638246,
+      "loss": 1.1849,
+      "step": 6710
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 0.00018067920803356178,
+      "loss": 1.0921,
+      "step": 6720
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 0.0001806412422407411,
+      "loss": 1.1229,
+      "step": 6730
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 0.00018060327644792043,
+      "loss": 1.1526,
+      "step": 6740
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 0.00018056531065509975,
+      "loss": 1.0688,
+      "step": 6750
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 0.0001805273448622791,
+      "loss": 1.208,
+      "step": 6760
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 0.00018048937906945843,
+      "loss": 1.0821,
+      "step": 6770
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 0.00018045141327663775,
+      "loss": 1.0888,
+      "step": 6780
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 0.0001804134474838171,
+      "loss": 1.1763,
+      "step": 6790
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 0.00018037548169099642,
+      "loss": 1.0669,
+      "step": 6800
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 0.00018033751589817575,
+      "loss": 1.2801,
+      "step": 6810
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 0.0001802995501053551,
+      "loss": 1.0969,
+      "step": 6820
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 0.00018026158431253442,
+      "loss": 1.1022,
+      "step": 6830
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 0.00018022361851971374,
+      "loss": 1.1699,
+      "step": 6840
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 0.00018018565272689307,
+      "loss": 1.0569,
+      "step": 6850
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 0.0001801476869340724,
+      "loss": 1.1959,
+      "step": 6860
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 0.00018010972114125174,
+      "loss": 1.1223,
+      "step": 6870
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 0.0001800717553484311,
+      "loss": 1.1347,
+      "step": 6880
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 0.00018003378955561042,
+      "loss": 1.1832,
+      "step": 6890
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 0.00017999582376278974,
+      "loss": 1.0275,
+      "step": 6900
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 0.00017995785796996906,
+      "loss": 1.2386,
+      "step": 6910
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 0.00017991989217714839,
+      "loss": 1.0886,
+      "step": 6920
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 0.00017988192638432774,
+      "loss": 1.1499,
+      "step": 6930
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 0.00017984396059150706,
+      "loss": 1.1467,
+      "step": 6940
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 0.00017980599479868638,
+      "loss": 1.0457,
+      "step": 6950
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 0.00017976802900586573,
+      "loss": 1.182,
+      "step": 6960
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 0.00017973006321304506,
+      "loss": 1.1167,
+      "step": 6970
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 0.00017969209742022438,
+      "loss": 1.0978,
+      "step": 6980
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 0.00017965413162740373,
+      "loss": 1.1854,
+      "step": 6990
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 0.00017961616583458305,
+      "loss": 1.0204,
+      "step": 7000
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 0.00017957820004176238,
+      "loss": 1.1626,
+      "step": 7010
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 0.0001795402342489417,
+      "loss": 1.0955,
+      "step": 7020
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 0.00017950226845612102,
+      "loss": 1.0513,
+      "step": 7030
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 0.00017946430266330038,
+      "loss": 1.1696,
+      "step": 7040
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 0.00017942633687047973,
+      "loss": 1.0451,
+      "step": 7050
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 0.00017938837107765905,
+      "loss": 1.2098,
+      "step": 7060
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 0.00017935040528483837,
+      "loss": 1.1188,
+      "step": 7070
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 0.0001793124394920177,
+      "loss": 1.1274,
+      "step": 7080
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 0.00017927447369919702,
+      "loss": 1.1945,
+      "step": 7090
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 0.00017923650790637637,
+      "loss": 1.0446,
+      "step": 7100
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 0.0001791985421135557,
+      "loss": 1.203,
+      "step": 7110
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 0.00017916057632073502,
+      "loss": 1.0866,
+      "step": 7120
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 0.00017912261052791434,
+      "loss": 1.0894,
+      "step": 7130
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 0.0001790846447350937,
+      "loss": 1.172,
+      "step": 7140
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 0.00017904667894227301,
+      "loss": 1.034,
+      "step": 7150
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 0.00017900871314945237,
+      "loss": 1.1868,
+      "step": 7160
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 0.0001789707473566317,
+      "loss": 1.1176,
+      "step": 7170
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 0.000178932781563811,
+      "loss": 1.0997,
+      "step": 7180
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 0.00017889481577099034,
+      "loss": 1.1568,
+      "step": 7190
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 0.00017885684997816966,
+      "loss": 1.0912,
+      "step": 7200
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 0.000178818884185349,
+      "loss": 1.2338,
+      "step": 7210
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 0.00017878091839252836,
+      "loss": 1.134,
+      "step": 7220
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 0.00017874295259970768,
+      "loss": 1.1345,
+      "step": 7230
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 0.000178704986806887,
+      "loss": 1.1553,
+      "step": 7240
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 0.00017866702101406633,
+      "loss": 1.0481,
+      "step": 7250
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 0.00017862905522124565,
+      "loss": 1.1803,
+      "step": 7260
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 0.000178591089428425,
+      "loss": 1.1155,
+      "step": 7270
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 0.00017855312363560433,
+      "loss": 1.172,
+      "step": 7280
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 0.00017851515784278365,
+      "loss": 1.196,
+      "step": 7290
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 0.00017847719204996297,
+      "loss": 1.0885,
+      "step": 7300
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 0.00017843922625714233,
+      "loss": 1.2438,
+      "step": 7310
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 0.00017840126046432168,
+      "loss": 1.1048,
+      "step": 7320
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 0.000178363294671501,
+      "loss": 1.071,
+      "step": 7330
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 0.00017832532887868032,
+      "loss": 1.1621,
+      "step": 7340
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 0.00017828736308585965,
+      "loss": 1.0827,
+      "step": 7350
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 0.00017824939729303897,
+      "loss": 1.2149,
+      "step": 7360
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 0.00017821143150021832,
+      "loss": 1.1233,
+      "step": 7370
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 0.00017817346570739764,
+      "loss": 1.1542,
+      "step": 7380
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 0.00017813549991457697,
+      "loss": 1.1472,
+      "step": 7390
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 0.00017809753412175632,
+      "loss": 1.0609,
+      "step": 7400
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 0.00017805956832893564,
+      "loss": 1.1793,
+      "step": 7410
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 0.00017802160253611496,
+      "loss": 1.0775,
+      "step": 7420
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 0.00017798363674329431,
+      "loss": 1.1373,
+      "step": 7430
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 0.00017794567095047364,
+      "loss": 1.1609,
+      "step": 7440
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 0.00017790770515765296,
+      "loss": 0.9958,
+      "step": 7450
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 0.00017786973936483229,
+      "loss": 1.2479,
+      "step": 7460
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 0.0001778317735720116,
+      "loss": 1.0825,
+      "step": 7470
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 0.00017779380777919096,
+      "loss": 1.1483,
+      "step": 7480
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 0.0001777558419863703,
+      "loss": 1.1722,
+      "step": 7490
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 0.00017771787619354963,
+      "loss": 1.0618,
+      "step": 7500
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 0.00017767991040072896,
+      "loss": 1.1825,
+      "step": 7510
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 0.00017764194460790828,
+      "loss": 1.1448,
+      "step": 7520
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 0.0001776039788150876,
+      "loss": 1.102,
+      "step": 7530
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 0.00017756601302226695,
+      "loss": 1.1665,
+      "step": 7540
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 0.00017752804722944628,
+      "loss": 1.0456,
+      "step": 7550
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 0.0001774900814366256,
+      "loss": 1.2029,
+      "step": 7560
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 0.00017745211564380495,
+      "loss": 1.1128,
+      "step": 7570
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 0.00017741414985098427,
+      "loss": 1.1381,
+      "step": 7580
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 0.0001773761840581636,
+      "loss": 1.1638,
+      "step": 7590
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 0.00017733821826534295,
+      "loss": 1.0881,
+      "step": 7600
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 0.00017730025247252227,
+      "loss": 1.2373,
+      "step": 7610
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 0.0001772622866797016,
+      "loss": 1.1243,
+      "step": 7620
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 0.00017722432088688092,
+      "loss": 1.0909,
+      "step": 7630
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 0.00017718635509406024,
+      "loss": 1.1583,
+      "step": 7640
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 0.0001771483893012396,
+      "loss": 1.0165,
+      "step": 7650
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 0.00017711042350841894,
+      "loss": 1.1867,
+      "step": 7660
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 0.00017707245771559827,
+      "loss": 1.0553,
+      "step": 7670
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 0.0001770344919227776,
+      "loss": 1.1146,
+      "step": 7680
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 0.00017699652612995691,
+      "loss": 1.1711,
+      "step": 7690
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 0.00017695856033713624,
+      "loss": 1.06,
+      "step": 7700
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 0.0001769205945443156,
+      "loss": 1.2259,
+      "step": 7710
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 0.0001768826287514949,
+      "loss": 1.1026,
+      "step": 7720
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 0.00017684466295867423,
+      "loss": 1.1233,
+      "step": 7730
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 0.00017680669716585359,
+      "loss": 1.1615,
+      "step": 7740
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 0.0001767687313730329,
+      "loss": 1.0338,
+      "step": 7750
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 0.00017673076558021223,
+      "loss": 1.2506,
+      "step": 7760
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 0.00017669279978739158,
+      "loss": 1.1077,
+      "step": 7770
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 0.0001766548339945709,
+      "loss": 1.0701,
+      "step": 7780
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 0.00017661686820175023,
+      "loss": 1.1583,
+      "step": 7790
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 0.00017657890240892955,
+      "loss": 1.0209,
+      "step": 7800
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 0.00017654093661610888,
+      "loss": 1.2504,
+      "step": 7810
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 0.00017650297082328823,
+      "loss": 1.0819,
+      "step": 7820
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 0.00017646500503046758,
+      "loss": 1.1083,
+      "step": 7830
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 0.0001764270392376469,
+      "loss": 1.1461,
+      "step": 7840
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 0.00017638907344482622,
+      "loss": 1.0021,
+      "step": 7850
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 0.00017635110765200555,
+      "loss": 1.1797,
+      "step": 7860
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 0.00017631314185918487,
+      "loss": 1.1139,
+      "step": 7870
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 0.00017627517606636422,
+      "loss": 1.112,
+      "step": 7880
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 0.00017623721027354355,
+      "loss": 1.1878,
+      "step": 7890
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 0.00017619924448072287,
+      "loss": 1.0614,
+      "step": 7900
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 0.0001761612786879022,
+      "loss": 1.2,
+      "step": 7910
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 0.00017612331289508154,
+      "loss": 1.0745,
+      "step": 7920
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 0.00017608534710226087,
+      "loss": 1.091,
+      "step": 7930
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 0.00017604738130944022,
+      "loss": 1.1541,
+      "step": 7940
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 0.00017600941551661954,
+      "loss": 1.0546,
+      "step": 7950
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 0.00017597144972379886,
+      "loss": 1.1869,
+      "step": 7960
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 0.0001759334839309782,
+      "loss": 1.0857,
+      "step": 7970
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 0.0001758955181381575,
+      "loss": 1.1167,
+      "step": 7980
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 0.00017585755234533686,
+      "loss": 1.1943,
+      "step": 7990
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 0.0001758195865525162,
+      "loss": 1.0637,
+      "step": 8000
+    },
+    {
+      "epoch": 0.44,
+      "eval_loss": 1.1032294034957886,
+      "eval_runtime": 2.0046,
+      "eval_samples_per_second": 49.886,
+      "eval_steps_per_second": 6.485,
+      "step": 8000
+    },
+    {
+      "epoch": 0.44,
+      "mmlu_eval_accuracy": NaN,
+      "mmlu_eval_accuracy_abstract_algebra": 0.5454545454545454,
+      "mmlu_eval_accuracy_anatomy": 0.35714285714285715,
+      "mmlu_eval_accuracy_astronomy": 0.0625,
+      "mmlu_eval_accuracy_business_ethics": 0.5454545454545454,
+      "mmlu_eval_accuracy_clinical_knowledge": 0.41379310344827586,
+      "mmlu_eval_accuracy_college_biology": 0.8125,
+      "mmlu_eval_accuracy_college_chemistry": 1.0,
+      "mmlu_eval_accuracy_college_computer_science": 0.6363636363636364,
+      "mmlu_eval_accuracy_college_mathematics": 0.18181818181818182,
+      "mmlu_eval_accuracy_college_medicine": 0.45454545454545453,
+      "mmlu_eval_accuracy_college_physics": 0.2727272727272727,
+      "mmlu_eval_accuracy_computer_security": 0.6363636363636364,
+      "mmlu_eval_accuracy_conceptual_physics": 0.5238095238095238,
+      "mmlu_eval_accuracy_econometrics": NaN,
+      "mmlu_eval_accuracy_electrical_engineering": NaN,
+      "mmlu_eval_accuracy_elementary_mathematics": NaN,
+      "mmlu_eval_accuracy_formal_logic": NaN,
+      "mmlu_eval_accuracy_global_facts": NaN,
+      "mmlu_eval_accuracy_high_school_biology": NaN,
+      "mmlu_eval_accuracy_high_school_chemistry": NaN,
+      "mmlu_eval_accuracy_high_school_computer_science": NaN,
+      "mmlu_eval_accuracy_high_school_european_history": NaN,
+      "mmlu_eval_accuracy_high_school_geography": NaN,
+      "mmlu_eval_accuracy_high_school_government_and_politics": NaN,
+      "mmlu_eval_accuracy_high_school_macroeconomics": NaN,
+      "mmlu_eval_accuracy_high_school_mathematics": NaN,
+      "mmlu_eval_accuracy_high_school_microeconomics": NaN,
+      "mmlu_eval_accuracy_high_school_physics": NaN,
+      "mmlu_eval_accuracy_high_school_psychology": NaN,
+      "mmlu_eval_accuracy_high_school_statistics": NaN,
+      "mmlu_eval_accuracy_high_school_us_history": NaN,
+      "mmlu_eval_accuracy_high_school_world_history": NaN,
+      "mmlu_eval_accuracy_human_aging": NaN,
+      "mmlu_eval_accuracy_human_sexuality": NaN,
+      "mmlu_eval_accuracy_international_law": NaN,
+      "mmlu_eval_accuracy_jurisprudence": NaN,
+      "mmlu_eval_accuracy_logical_fallacies": NaN,
+      "mmlu_eval_accuracy_machine_learning": NaN,
+      "mmlu_eval_accuracy_management": NaN,
+      "mmlu_eval_accuracy_marketing": NaN,
+      "mmlu_eval_accuracy_medical_genetics": NaN,
+      "mmlu_eval_accuracy_miscellaneous": NaN,
+      "mmlu_eval_accuracy_moral_disputes": NaN,
+      "mmlu_eval_accuracy_moral_scenarios": NaN,
+      "mmlu_eval_accuracy_nutrition": NaN,
+      "mmlu_eval_accuracy_philosophy": NaN,
+      "mmlu_eval_accuracy_prehistory": NaN,
+      "mmlu_eval_accuracy_professional_accounting": NaN,
+      "mmlu_eval_accuracy_professional_law": NaN,
+      "mmlu_eval_accuracy_professional_medicine": NaN,
+      "mmlu_eval_accuracy_professional_psychology": NaN,
+      "mmlu_eval_accuracy_public_relations": NaN,
+      "mmlu_eval_accuracy_security_studies": NaN,
+      "mmlu_eval_accuracy_sociology": NaN,
+      "mmlu_eval_accuracy_us_foreign_policy": NaN,
+      "mmlu_eval_accuracy_virology": NaN,
+      "mmlu_eval_accuracy_world_religions": NaN,
+      "mmlu_loss": 1.1005631206401933,
+      "step": 8000
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 0.00017578162075969553,
+      "loss": 1.1715,
+      "step": 8010
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 0.00017574365496687486,
+      "loss": 1.0833,
+      "step": 8020
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 0.00017570568917405418,
+      "loss": 1.1304,
+      "step": 8030
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 0.0001756677233812335,
+      "loss": 1.1738,
+      "step": 8040
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 0.00017562975758841286,
+      "loss": 1.0631,
+      "step": 8050
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 0.00017559179179559218,
+      "loss": 1.2041,
+      "step": 8060
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 0.0001755538260027715,
+      "loss": 1.1268,
+      "step": 8070
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 0.00017551586020995083,
+      "loss": 1.1077,
+      "step": 8080
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 0.00017547789441713018,
+      "loss": 1.154,
+      "step": 8090
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 0.0001754399286243095,
+      "loss": 1.0583,
+      "step": 8100
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 0.00017540196283148885,
+      "loss": 1.1546,
+      "step": 8110
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 0.00017536399703866817,
+      "loss": 1.0995,
+      "step": 8120
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 0.0001753260312458475,
+      "loss": 1.1243,
+      "step": 8130
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 0.00017528806545302682,
+      "loss": 1.1597,
+      "step": 8140
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 0.00017525009966020614,
+      "loss": 1.0308,
+      "step": 8150
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 0.0001752121338673855,
+      "loss": 1.1751,
+      "step": 8160
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 0.00017517416807456482,
+      "loss": 1.1108,
+      "step": 8170
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 0.00017513620228174417,
+      "loss": 1.1264,
+      "step": 8180
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 0.0001750982364889235,
+      "loss": 1.1662,
+      "step": 8190
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 0.00017506027069610282,
+      "loss": 1.0727,
+      "step": 8200
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 0.00017502230490328214,
+      "loss": 1.1721,
+      "step": 8210
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 0.0001749843391104615,
+      "loss": 1.1738,
+      "step": 8220
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 0.0001749463733176408,
+      "loss": 1.1167,
+      "step": 8230
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 0.00017490840752482014,
+      "loss": 1.1913,
+      "step": 8240
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 0.00017487044173199946,
+      "loss": 1.0407,
+      "step": 8250
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 0.0001748324759391788,
+      "loss": 1.1789,
+      "step": 8260
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 0.00017479451014635816,
+      "loss": 1.079,
+      "step": 8270
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 0.00017475654435353748,
+      "loss": 1.0893,
+      "step": 8280
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 0.0001747185785607168,
+      "loss": 1.2122,
+      "step": 8290
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 0.00017468061276789613,
+      "loss": 1.048,
+      "step": 8300
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 0.00017464264697507545,
+      "loss": 1.2269,
+      "step": 8310
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 0.0001746046811822548,
+      "loss": 1.1305,
+      "step": 8320
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 0.00017456671538943413,
+      "loss": 1.1337,
+      "step": 8330
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 0.00017452874959661345,
+      "loss": 1.1664,
+      "step": 8340
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 0.0001744907838037928,
+      "loss": 1.0228,
+      "step": 8350
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 0.00017445281801097213,
+      "loss": 1.1606,
+      "step": 8360
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 0.00017441485221815145,
+      "loss": 1.09,
+      "step": 8370
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 0.0001743768864253308,
+      "loss": 1.1113,
+      "step": 8380
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 0.00017433892063251012,
+      "loss": 1.2006,
+      "step": 8390
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 0.00017430095483968945,
+      "loss": 1.0509,
+      "step": 8400
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 0.00017426298904686877,
+      "loss": 1.2254,
+      "step": 8410
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 0.0001742250232540481,
+      "loss": 1.1138,
+      "step": 8420
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 0.00017418705746122744,
+      "loss": 1.1118,
+      "step": 8430
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 0.0001741490916684068,
+      "loss": 1.1802,
+      "step": 8440
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 0.00017411112587558612,
+      "loss": 1.0444,
+      "step": 8450
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 0.00017407316008276544,
+      "loss": 1.2194,
+      "step": 8460
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 0.00017403519428994477,
+      "loss": 1.1218,
+      "step": 8470
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 0.0001739972284971241,
+      "loss": 1.1704,
+      "step": 8480
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 0.00017395926270430344,
+      "loss": 1.1731,
+      "step": 8490
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 0.00017392129691148276,
+      "loss": 1.0216,
+      "step": 8500
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 0.00017388333111866209,
+      "loss": 1.2453,
+      "step": 8510
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 0.00017384536532584144,
+      "loss": 1.1244,
+      "step": 8520
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 0.00017380739953302076,
+      "loss": 1.1442,
+      "step": 8530
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 0.00017376943374020008,
+      "loss": 1.1991,
+      "step": 8540
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 0.00017373146794737943,
+      "loss": 1.0731,
+      "step": 8550
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 0.00017369350215455876,
+      "loss": 1.2022,
+      "step": 8560
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 0.00017365553636173808,
+      "loss": 1.0881,
+      "step": 8570
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 0.0001736175705689174,
+      "loss": 1.1009,
+      "step": 8580
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 0.00017357960477609673,
+      "loss": 1.1683,
+      "step": 8590
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 0.00017354163898327608,
+      "loss": 1.0271,
+      "step": 8600
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 0.00017350367319045543,
+      "loss": 1.1771,
+      "step": 8610
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 0.00017346570739763475,
+      "loss": 1.1151,
+      "step": 8620
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 0.00017342774160481408,
+      "loss": 1.1111,
+      "step": 8630
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 0.0001733897758119934,
+      "loss": 1.2079,
+      "step": 8640
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 0.00017335181001917272,
+      "loss": 1.0144,
+      "step": 8650
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 0.00017331384422635207,
+      "loss": 1.1869,
+      "step": 8660
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 0.0001732758784335314,
+      "loss": 1.0914,
+      "step": 8670
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 0.00017323791264071072,
+      "loss": 1.1002,
+      "step": 8680
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 0.00017319994684789004,
+      "loss": 1.1338,
+      "step": 8690
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 0.0001731619810550694,
+      "loss": 1.0108,
+      "step": 8700
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 0.00017312401526224872,
+      "loss": 1.1879,
+      "step": 8710
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 0.00017308604946942807,
+      "loss": 1.1358,
+      "step": 8720
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 0.0001730480836766074,
+      "loss": 1.0972,
+      "step": 8730
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 0.00017301011788378672,
+      "loss": 1.1843,
+      "step": 8740
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 0.00017297215209096604,
+      "loss": 1.0276,
+      "step": 8750
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 0.00017293418629814536,
+      "loss": 1.2217,
+      "step": 8760
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 0.0001728962205053247,
+      "loss": 1.0545,
+      "step": 8770
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 0.00017285825471250406,
+      "loss": 1.0885,
+      "step": 8780
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 0.0001728202889196834,
+      "loss": 1.152,
+      "step": 8790
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 0.0001727823231268627,
+      "loss": 1.0541,
+      "step": 8800
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 0.00017274435733404203,
+      "loss": 1.2015,
+      "step": 8810
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 0.00017270639154122136,
+      "loss": 1.1192,
+      "step": 8820
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 0.0001726684257484007,
+      "loss": 1.0976,
+      "step": 8830
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 0.00017263045995558003,
+      "loss": 1.1027,
+      "step": 8840
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 0.00017259249416275935,
+      "loss": 1.0673,
+      "step": 8850
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 0.00017255452836993868,
+      "loss": 1.2016,
+      "step": 8860
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 0.00017251656257711803,
+      "loss": 1.0708,
+      "step": 8870
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 0.00017247859678429735,
+      "loss": 1.1211,
+      "step": 8880
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 0.0001724406309914767,
+      "loss": 1.1354,
+      "step": 8890
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 0.00017240266519865603,
+      "loss": 1.0757,
+      "step": 8900
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 0.00017236469940583535,
+      "loss": 1.2318,
+      "step": 8910
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 0.00017232673361301467,
+      "loss": 1.1231,
+      "step": 8920
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 0.000172288767820194,
+      "loss": 1.1097,
+      "step": 8930
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 0.00017225080202737335,
+      "loss": 1.1618,
+      "step": 8940
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 0.00017221283623455267,
+      "loss": 1.0874,
+      "step": 8950
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 0.00017217487044173202,
+      "loss": 1.2472,
+      "step": 8960
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 0.00017213690464891134,
+      "loss": 1.131,
+      "step": 8970
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 0.00017209893885609067,
+      "loss": 1.0906,
+      "step": 8980
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 0.00017206097306327,
+      "loss": 1.132,
+      "step": 8990
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 0.00017202300727044934,
+      "loss": 1.0318,
+      "step": 9000
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 0.00017198504147762866,
+      "loss": 1.2357,
+      "step": 9010
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 0.000171947075684808,
+      "loss": 1.067,
+      "step": 9020
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 0.0001719091098919873,
+      "loss": 1.0932,
+      "step": 9030
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 0.00017187114409916666,
+      "loss": 1.1658,
+      "step": 9040
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 0.00017183317830634599,
+      "loss": 1.0291,
+      "step": 9050
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 0.00017179521251352534,
+      "loss": 1.183,
+      "step": 9060
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 0.00017175724672070466,
+      "loss": 1.1261,
+      "step": 9070
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 0.00017171928092788398,
+      "loss": 1.1133,
+      "step": 9080
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 0.0001716813151350633,
+      "loss": 1.2068,
+      "step": 9090
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 0.00017164334934224263,
+      "loss": 1.0777,
+      "step": 9100
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 0.00017160538354942198,
+      "loss": 1.2064,
+      "step": 9110
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 0.0001715674177566013,
+      "loss": 1.1148,
+      "step": 9120
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 0.00017152945196378065,
+      "loss": 1.1075,
+      "step": 9130
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 0.00017149148617095998,
+      "loss": 1.2026,
+      "step": 9140
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 0.0001714535203781393,
+      "loss": 1.0355,
+      "step": 9150
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 0.00017141555458531862,
+      "loss": 1.1654,
+      "step": 9160
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 0.00017137758879249798,
+      "loss": 1.1032,
+      "step": 9170
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 0.0001713396229996773,
+      "loss": 1.1445,
+      "step": 9180
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 0.00017130165720685662,
+      "loss": 1.205,
+      "step": 9190
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 0.00017126369141403595,
+      "loss": 1.0083,
+      "step": 9200
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 0.00017122572562121527,
+      "loss": 1.2079,
+      "step": 9210
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 0.00017118775982839465,
+      "loss": 1.0913,
+      "step": 9220
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 0.00017114979403557397,
+      "loss": 1.0892,
+      "step": 9230
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 0.0001711118282427533,
+      "loss": 1.1518,
+      "step": 9240
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 0.00017107386244993262,
+      "loss": 1.0212,
+      "step": 9250
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 0.00017103589665711194,
+      "loss": 1.1316,
+      "step": 9260
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 0.0001709979308642913,
+      "loss": 1.1017,
+      "step": 9270
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 0.00017095996507147061,
+      "loss": 1.1198,
+      "step": 9280
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 0.00017092199927864994,
+      "loss": 1.1801,
+      "step": 9290
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 0.0001708840334858293,
+      "loss": 1.0817,
+      "step": 9300
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 0.0001708460676930086,
+      "loss": 1.2189,
+      "step": 9310
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 0.00017080810190018794,
+      "loss": 1.0838,
+      "step": 9320
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 0.00017077013610736729,
+      "loss": 1.1358,
+      "step": 9330
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 0.0001707321703145466,
+      "loss": 1.1816,
+      "step": 9340
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 0.00017069420452172593,
+      "loss": 1.0447,
+      "step": 9350
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 0.00017065623872890526,
+      "loss": 1.1844,
+      "step": 9360
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 0.00017061827293608458,
+      "loss": 1.1397,
+      "step": 9370
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 0.00017058030714326393,
+      "loss": 1.1515,
+      "step": 9380
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 0.00017054234135044328,
+      "loss": 1.1859,
+      "step": 9390
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 0.0001705043755576226,
+      "loss": 1.0596,
+      "step": 9400
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 0.00017046640976480193,
+      "loss": 1.1949,
+      "step": 9410
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 0.00017042844397198125,
+      "loss": 1.0483,
+      "step": 9420
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 0.00017039047817916057,
+      "loss": 1.1605,
+      "step": 9430
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 0.00017035251238633993,
+      "loss": 1.2048,
+      "step": 9440
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 0.00017031454659351925,
+      "loss": 1.0345,
+      "step": 9450
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 0.00017027658080069857,
+      "loss": 1.2459,
+      "step": 9460
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 0.0001702386150078779,
+      "loss": 1.0672,
+      "step": 9470
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 0.00017020064921505725,
+      "loss": 1.097,
+      "step": 9480
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 0.00017016268342223657,
+      "loss": 1.2101,
+      "step": 9490
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 0.00017012471762941592,
+      "loss": 1.0143,
+      "step": 9500
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 0.00017008675183659524,
+      "loss": 1.2034,
+      "step": 9510
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 0.00017004878604377457,
+      "loss": 1.0942,
+      "step": 9520
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 0.0001700108202509539,
+      "loss": 1.09,
+      "step": 9530
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 0.00016997285445813321,
+      "loss": 1.1249,
+      "step": 9540
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 0.00016993488866531256,
+      "loss": 1.0765,
+      "step": 9550
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 0.00016989692287249191,
+      "loss": 1.206,
+      "step": 9560
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 0.00016985895707967124,
+      "loss": 1.1043,
+      "step": 9570
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 0.00016982099128685056,
+      "loss": 1.1091,
+      "step": 9580
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 0.00016978302549402989,
+      "loss": 1.1638,
+      "step": 9590
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 0.0001697450597012092,
+      "loss": 1.0236,
+      "step": 9600
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 0.00016970709390838856,
+      "loss": 1.1924,
+      "step": 9610
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 0.00016966912811556788,
+      "loss": 1.0758,
+      "step": 9620
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 0.0001696311623227472,
+      "loss": 1.1459,
+      "step": 9630
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 0.00016959319652992653,
+      "loss": 1.1309,
+      "step": 9640
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 0.00016955523073710588,
+      "loss": 1.1036,
+      "step": 9650
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 0.0001695172649442852,
+      "loss": 1.1954,
+      "step": 9660
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 0.00016947929915146455,
+      "loss": 1.0518,
+      "step": 9670
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 0.00016944133335864388,
+      "loss": 1.1108,
+      "step": 9680
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 0.0001694033675658232,
+      "loss": 1.1727,
+      "step": 9690
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 0.00016936540177300252,
+      "loss": 1.043,
+      "step": 9700
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 0.00016932743598018185,
+      "loss": 1.1429,
+      "step": 9710
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 0.0001692894701873612,
+      "loss": 1.0893,
+      "step": 9720
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 0.00016925150439454052,
+      "loss": 1.0873,
+      "step": 9730
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 0.00016921353860171987,
+      "loss": 1.1595,
+      "step": 9740
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 0.0001691755728088992,
+      "loss": 1.0543,
+      "step": 9750
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 0.00016913760701607852,
+      "loss": 1.2044,
+      "step": 9760
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 0.00016909964122325784,
+      "loss": 1.1167,
+      "step": 9770
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 0.0001690616754304372,
+      "loss": 1.1231,
+      "step": 9780
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 0.00016902370963761652,
+      "loss": 1.1408,
+      "step": 9790
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 0.00016898574384479584,
+      "loss": 1.0285,
+      "step": 9800
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 0.00016894777805197516,
+      "loss": 1.2118,
+      "step": 9810
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 0.00016890981225915451,
+      "loss": 1.0846,
+      "step": 9820
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 0.00016887184646633384,
+      "loss": 1.1582,
+      "step": 9830
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 0.0001688338806735132,
+      "loss": 1.1556,
+      "step": 9840
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 0.0001687959148806925,
+      "loss": 1.0019,
+      "step": 9850
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 0.00016875794908787183,
+      "loss": 1.1901,
+      "step": 9860
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 0.00016871998329505116,
+      "loss": 1.1013,
+      "step": 9870
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 0.00016868201750223048,
+      "loss": 1.1148,
+      "step": 9880
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 0.00016864405170940983,
+      "loss": 1.1592,
+      "step": 9890
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 0.00016860608591658916,
+      "loss": 1.0152,
+      "step": 9900
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 0.0001685681201237685,
+      "loss": 1.1935,
+      "step": 9910
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 0.00016853015433094783,
+      "loss": 1.1219,
+      "step": 9920
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 0.00016849218853812715,
+      "loss": 1.0601,
+      "step": 9930
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 0.00016845422274530648,
+      "loss": 1.1514,
+      "step": 9940
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 0.00016841625695248583,
+      "loss": 0.9967,
+      "step": 9950
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 0.00016837829115966515,
+      "loss": 1.1804,
+      "step": 9960
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 0.00016834032536684447,
+      "loss": 1.106,
+      "step": 9970
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 0.0001683023595740238,
+      "loss": 1.1105,
+      "step": 9980
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 0.00016826439378120312,
+      "loss": 1.1945,
+      "step": 9990
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 0.00016822642798838247,
+      "loss": 1.0313,
+      "step": 10000
+    },
+    {
+      "epoch": 0.55,
+      "eval_loss": 1.093135952949524,
+      "eval_runtime": 1.9627,
+      "eval_samples_per_second": 50.95,
+      "eval_steps_per_second": 6.624,
+      "step": 10000
+    },
+    {
+      "epoch": 0.55,
+      "mmlu_eval_accuracy": NaN,
+      "mmlu_eval_accuracy_abstract_algebra": 0.45454545454545453,
+      "mmlu_eval_accuracy_anatomy": 0.35714285714285715,
+      "mmlu_eval_accuracy_astronomy": 0.125,
+      "mmlu_eval_accuracy_business_ethics": 0.6363636363636364,
+      "mmlu_eval_accuracy_clinical_knowledge": 0.41379310344827586,
+      "mmlu_eval_accuracy_college_biology": 0.75,
+      "mmlu_eval_accuracy_college_chemistry": 0.875,
+      "mmlu_eval_accuracy_college_computer_science": 0.6363636363636364,
+      "mmlu_eval_accuracy_college_mathematics": 0.18181818181818182,
+      "mmlu_eval_accuracy_college_medicine": 0.5454545454545454,
+      "mmlu_eval_accuracy_college_physics": 0.18181818181818182,
+      "mmlu_eval_accuracy_computer_security": 0.45454545454545453,
+      "mmlu_eval_accuracy_conceptual_physics": 0.47619047619047616,
+      "mmlu_eval_accuracy_econometrics": NaN,
+      "mmlu_eval_accuracy_electrical_engineering": NaN,
+      "mmlu_eval_accuracy_elementary_mathematics": NaN,
+      "mmlu_eval_accuracy_formal_logic": NaN,
+      "mmlu_eval_accuracy_global_facts": NaN,
+      "mmlu_eval_accuracy_high_school_biology": NaN,
+      "mmlu_eval_accuracy_high_school_chemistry": NaN,
+      "mmlu_eval_accuracy_high_school_computer_science": NaN,
+      "mmlu_eval_accuracy_high_school_european_history": NaN,
+      "mmlu_eval_accuracy_high_school_geography": NaN,
+      "mmlu_eval_accuracy_high_school_government_and_politics": NaN,
+      "mmlu_eval_accuracy_high_school_macroeconomics": NaN,
+      "mmlu_eval_accuracy_high_school_mathematics": NaN,
+      "mmlu_eval_accuracy_high_school_microeconomics": NaN,
+      "mmlu_eval_accuracy_high_school_physics": NaN,
+      "mmlu_eval_accuracy_high_school_psychology": NaN,
+      "mmlu_eval_accuracy_high_school_statistics": NaN,
+      "mmlu_eval_accuracy_high_school_us_history": NaN,
+      "mmlu_eval_accuracy_high_school_world_history": NaN,
+      "mmlu_eval_accuracy_human_aging": NaN,
+      "mmlu_eval_accuracy_human_sexuality": NaN,
+      "mmlu_eval_accuracy_international_law": NaN,
+      "mmlu_eval_accuracy_jurisprudence": NaN,
+      "mmlu_eval_accuracy_logical_fallacies": NaN,
+      "mmlu_eval_accuracy_machine_learning": NaN,
+      "mmlu_eval_accuracy_management": NaN,
+      "mmlu_eval_accuracy_marketing": NaN,
+      "mmlu_eval_accuracy_medical_genetics": NaN,
+      "mmlu_eval_accuracy_miscellaneous": NaN,
+      "mmlu_eval_accuracy_moral_disputes": NaN,
+      "mmlu_eval_accuracy_moral_scenarios": NaN,
+      "mmlu_eval_accuracy_nutrition": NaN,
+      "mmlu_eval_accuracy_philosophy": NaN,
+      "mmlu_eval_accuracy_prehistory": NaN,
+      "mmlu_eval_accuracy_professional_accounting": NaN,
+      "mmlu_eval_accuracy_professional_law": NaN,
+      "mmlu_eval_accuracy_professional_medicine": NaN,
+      "mmlu_eval_accuracy_professional_psychology": NaN,
+      "mmlu_eval_accuracy_public_relations": NaN,
+      "mmlu_eval_accuracy_security_studies": NaN,
+      "mmlu_eval_accuracy_sociology": NaN,
+      "mmlu_eval_accuracy_us_foreign_policy": NaN,
+      "mmlu_eval_accuracy_virology": NaN,
+      "mmlu_eval_accuracy_world_religions": NaN,
+      "mmlu_loss": 0.9934630032318333,
+      "step": 10000
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 0.00016818846219556182,
+      "loss": 1.2753,
+      "step": 10010
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 0.00016815049640274115,
+      "loss": 1.0814,
+      "step": 10020
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 0.00016811253060992047,
+      "loss": 1.0591,
+      "step": 10030
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 0.0001680745648170998,
+      "loss": 1.1769,
+      "step": 10040
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 0.00016803659902427912,
+      "loss": 1.0637,
+      "step": 10050
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 0.00016799863323145847,
+      "loss": 1.1419,
+      "step": 10060
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 0.0001679606674386378,
+      "loss": 1.1077,
+      "step": 10070
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 0.00016792270164581714,
+      "loss": 1.1202,
+      "step": 10080
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 0.00016788473585299646,
+      "loss": 1.1092,
+      "step": 10090
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 0.0001678467700601758,
+      "loss": 1.0123,
+      "step": 10100
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 0.00016780880426735514,
+      "loss": 1.1836,
+      "step": 10110
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 0.00016777083847453446,
+      "loss": 1.0693,
+      "step": 10120
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 0.00016773287268171378,
+      "loss": 1.082,
+      "step": 10130
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 0.0001676949068888931,
+      "loss": 1.1454,
+      "step": 10140
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 0.00016765694109607243,
+      "loss": 1.0528,
+      "step": 10150
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 0.00016761897530325178,
+      "loss": 1.1933,
+      "step": 10160
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 0.00016758100951043113,
+      "loss": 1.1221,
+      "step": 10170
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 0.00016754304371761046,
+      "loss": 1.117,
+      "step": 10180
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 0.00016750507792478978,
+      "loss": 1.1451,
+      "step": 10190
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 0.0001674671121319691,
+      "loss": 1.0643,
+      "step": 10200
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 0.00016742914633914843,
+      "loss": 1.2249,
+      "step": 10210
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 0.00016739118054632778,
+      "loss": 1.115,
+      "step": 10220
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 0.0001673532147535071,
+      "loss": 1.0937,
+      "step": 10230
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 0.00016731524896068642,
+      "loss": 1.1612,
+      "step": 10240
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 0.00016727728316786575,
+      "loss": 1.0649,
+      "step": 10250
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 0.0001672393173750451,
+      "loss": 1.2074,
+      "step": 10260
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 0.00016720135158222442,
+      "loss": 1.0892,
+      "step": 10270
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 0.00016716338578940377,
+      "loss": 1.1302,
+      "step": 10280
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 0.0001671254199965831,
+      "loss": 1.1513,
+      "step": 10290
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 0.00016708745420376242,
+      "loss": 1.0549,
+      "step": 10300
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 0.00016704948841094174,
+      "loss": 1.2296,
+      "step": 10310
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 0.00016701152261812107,
+      "loss": 1.0671,
+      "step": 10320
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 0.00016697355682530042,
+      "loss": 1.1102,
+      "step": 10330
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 0.00016693559103247977,
+      "loss": 1.1001,
+      "step": 10340
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 0.0001668976252396591,
+      "loss": 1.0419,
+      "step": 10350
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 0.0001668596594468384,
+      "loss": 1.2167,
+      "step": 10360
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 0.00016682169365401774,
+      "loss": 1.1062,
+      "step": 10370
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 0.00016678372786119706,
+      "loss": 1.0865,
+      "step": 10380
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 0.0001667457620683764,
+      "loss": 1.1414,
+      "step": 10390
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 0.00016670779627555573,
+      "loss": 1.0388,
+      "step": 10400
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 0.00016666983048273506,
+      "loss": 1.1287,
+      "step": 10410
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 0.00016663186468991438,
+      "loss": 1.0894,
+      "step": 10420
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 0.00016659389889709373,
+      "loss": 1.1156,
+      "step": 10430
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 0.00016655593310427306,
+      "loss": 1.1405,
+      "step": 10440
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 0.0001665179673114524,
+      "loss": 1.0181,
+      "step": 10450
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 0.00016648000151863173,
+      "loss": 1.1604,
+      "step": 10460
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 0.00016644203572581105,
+      "loss": 1.0799,
+      "step": 10470
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 0.00016640406993299038,
+      "loss": 1.1437,
+      "step": 10480
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 0.0001663661041401697,
+      "loss": 1.1852,
+      "step": 10490
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 0.00016632813834734905,
+      "loss": 1.0712,
+      "step": 10500
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 0.00016629017255452837,
+      "loss": 1.1855,
+      "step": 10510
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 0.00016625220676170772,
+      "loss": 1.1211,
+      "step": 10520
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 0.00016621424096888705,
+      "loss": 1.0731,
+      "step": 10530
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 0.00016617627517606637,
+      "loss": 1.1385,
+      "step": 10540
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 0.0001661383093832457,
+      "loss": 1.0333,
+      "step": 10550
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 0.00016610034359042504,
+      "loss": 1.2008,
+      "step": 10560
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 0.00016606237779760437,
+      "loss": 1.0792,
+      "step": 10570
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 0.0001660244120047837,
+      "loss": 1.1494,
+      "step": 10580
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 0.00016598644621196302,
+      "loss": 1.1602,
+      "step": 10590
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 0.00016594848041914237,
+      "loss": 1.0484,
+      "step": 10600
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 0.0001659105146263217,
+      "loss": 1.2558,
+      "step": 10610
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 0.00016587254883350104,
+      "loss": 1.0949,
+      "step": 10620
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 0.00016583458304068036,
+      "loss": 1.1117,
+      "step": 10630
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 0.00016579661724785969,
+      "loss": 1.1609,
+      "step": 10640
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 0.000165758651455039,
+      "loss": 1.0533,
+      "step": 10650
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 0.00016572068566221833,
+      "loss": 1.2188,
+      "step": 10660
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 0.00016568271986939768,
+      "loss": 1.0741,
+      "step": 10670
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 0.000165644754076577,
+      "loss": 1.0587,
+      "step": 10680
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 0.00016560678828375636,
+      "loss": 1.1442,
+      "step": 10690
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 0.00016556882249093568,
+      "loss": 1.0176,
+      "step": 10700
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 0.000165530856698115,
+      "loss": 1.1438,
+      "step": 10710
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 0.00016549289090529433,
+      "loss": 1.1068,
+      "step": 10720
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 0.00016545492511247368,
+      "loss": 1.1599,
+      "step": 10730
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 0.000165416959319653,
+      "loss": 1.1086,
+      "step": 10740
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 0.00016537899352683233,
+      "loss": 1.017,
+      "step": 10750
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 0.00016534102773401165,
+      "loss": 1.1783,
+      "step": 10760
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 0.00016530306194119097,
+      "loss": 1.0951,
+      "step": 10770
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 0.00016526509614837032,
+      "loss": 1.1006,
+      "step": 10780
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 0.00016522713035554967,
+      "loss": 1.1423,
+      "step": 10790
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 0.000165189164562729,
+      "loss": 1.0442,
+      "step": 10800
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 0.00016515119876990832,
+      "loss": 1.1991,
+      "step": 10810
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 0.00016511323297708764,
+      "loss": 1.0897,
+      "step": 10820
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 0.00016507526718426697,
+      "loss": 1.1232,
+      "step": 10830
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 0.00016503730139144632,
+      "loss": 1.1423,
+      "step": 10840
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 0.00016499933559862564,
+      "loss": 1.0857,
+      "step": 10850
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 0.000164961369805805,
+      "loss": 1.1523,
+      "step": 10860
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 0.00016492340401298432,
+      "loss": 1.0755,
+      "step": 10870
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 0.00016488543822016364,
+      "loss": 1.1343,
+      "step": 10880
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 0.00016484747242734296,
+      "loss": 1.1869,
+      "step": 10890
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 0.0001648095066345223,
+      "loss": 0.9949,
+      "step": 10900
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 0.00016477154084170164,
+      "loss": 1.1829,
+      "step": 10910
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 0.00016473357504888096,
+      "loss": 1.066,
+      "step": 10920
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 0.00016469560925606028,
+      "loss": 1.1363,
+      "step": 10930
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 0.0001646576434632396,
+      "loss": 1.1399,
+      "step": 10940
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 0.00016461967767041896,
+      "loss": 1.0054,
+      "step": 10950
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 0.0001645817118775983,
+      "loss": 1.1933,
+      "step": 10960
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 0.00016454374608477763,
+      "loss": 1.081,
+      "step": 10970
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 0.00016450578029195695,
+      "loss": 1.1207,
+      "step": 10980
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 0.00016446781449913628,
+      "loss": 1.1576,
+      "step": 10990
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 0.0001644298487063156,
+      "loss": 1.0303,
+      "step": 11000
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 0.00016439188291349495,
+      "loss": 1.2308,
+      "step": 11010
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 0.00016435391712067428,
+      "loss": 1.0759,
+      "step": 11020
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 0.0001643159513278536,
+      "loss": 1.1499,
+      "step": 11030
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 0.00016427798553503295,
+      "loss": 1.2058,
+      "step": 11040
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 0.00016424001974221227,
+      "loss": 1.0445,
+      "step": 11050
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 0.00016420205394939162,
+      "loss": 1.1934,
+      "step": 11060
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 0.00016416408815657095,
+      "loss": 1.0907,
+      "step": 11070
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 0.00016412612236375027,
+      "loss": 1.083,
+      "step": 11080
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 0.0001640881565709296,
+      "loss": 1.1749,
+      "step": 11090
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 0.00016405019077810892,
+      "loss": 1.0066,
+      "step": 11100
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 0.00016401222498528827,
+      "loss": 1.1799,
+      "step": 11110
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 0.00016397425919246762,
+      "loss": 1.0522,
+      "step": 11120
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 0.00016393629339964694,
+      "loss": 1.0808,
+      "step": 11130
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 0.00016389832760682626,
+      "loss": 1.2018,
+      "step": 11140
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 0.0001638603618140056,
+      "loss": 1.0684,
+      "step": 11150
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 0.0001638223960211849,
+      "loss": 1.1679,
+      "step": 11160
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 0.00016378443022836426,
+      "loss": 1.0899,
+      "step": 11170
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 0.00016374646443554359,
+      "loss": 1.0833,
+      "step": 11180
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 0.0001637084986427229,
+      "loss": 1.182,
+      "step": 11190
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 0.00016367053284990223,
+      "loss": 1.0758,
+      "step": 11200
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 0.00016363256705708158,
+      "loss": 1.1437,
+      "step": 11210
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 0.0001635946012642609,
+      "loss": 1.1106,
+      "step": 11220
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 0.00016355663547144026,
+      "loss": 1.1053,
+      "step": 11230
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 0.00016351866967861958,
+      "loss": 1.1546,
+      "step": 11240
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 0.0001634807038857989,
+      "loss": 1.0076,
+      "step": 11250
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 0.00016344273809297823,
+      "loss": 1.1533,
+      "step": 11260
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 0.00016340477230015755,
+      "loss": 1.119,
+      "step": 11270
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 0.0001633668065073369,
+      "loss": 1.1071,
+      "step": 11280
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 0.00016332884071451622,
+      "loss": 1.1484,
+      "step": 11290
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 0.00016329087492169558,
+      "loss": 1.0669,
+      "step": 11300
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 0.0001632529091288749,
+      "loss": 1.2377,
+      "step": 11310
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 0.00016321494333605422,
+      "loss": 1.064,
+      "step": 11320
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 0.00016317697754323355,
+      "loss": 1.0834,
+      "step": 11330
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 0.0001631390117504129,
+      "loss": 1.1403,
+      "step": 11340
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 0.00016310104595759222,
+      "loss": 1.0699,
+      "step": 11350
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 0.00016306308016477154,
+      "loss": 1.1507,
+      "step": 11360
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 0.00016302511437195087,
+      "loss": 1.0728,
+      "step": 11370
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 0.00016298714857913022,
+      "loss": 1.0695,
+      "step": 11380
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 0.00016294918278630954,
+      "loss": 1.1562,
+      "step": 11390
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 0.0001629112169934889,
+      "loss": 1.0224,
+      "step": 11400
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 0.00016287325120066821,
+      "loss": 1.1824,
+      "step": 11410
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 0.00016283528540784754,
+      "loss": 1.1097,
+      "step": 11420
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 0.00016279731961502686,
+      "loss": 1.0889,
+      "step": 11430
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 0.00016275935382220618,
+      "loss": 1.1409,
+      "step": 11440
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 0.00016272138802938554,
+      "loss": 1.0545,
+      "step": 11450
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 0.00016268342223656486,
+      "loss": 1.2208,
+      "step": 11460
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 0.0001626454564437442,
+      "loss": 1.104,
+      "step": 11470
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 0.00016260749065092353,
+      "loss": 1.1207,
+      "step": 11480
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 0.00016256952485810286,
+      "loss": 1.1701,
+      "step": 11490
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 0.00016253155906528218,
+      "loss": 1.0237,
+      "step": 11500
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 0.00016249359327246153,
+      "loss": 1.2081,
+      "step": 11510
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 0.00016245562747964085,
+      "loss": 1.1031,
+      "step": 11520
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 0.00016241766168682018,
+      "loss": 1.0649,
+      "step": 11530
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 0.0001623796958939995,
+      "loss": 1.1853,
+      "step": 11540
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 0.00016234173010117882,
+      "loss": 1.0277,
+      "step": 11550
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 0.00016230376430835817,
+      "loss": 1.133,
+      "step": 11560
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 0.00016226579851553753,
+      "loss": 1.1233,
+      "step": 11570
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 0.00016222783272271685,
+      "loss": 1.0954,
+      "step": 11580
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 0.00016218986692989617,
+      "loss": 1.1522,
+      "step": 11590
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 0.0001621519011370755,
+      "loss": 1.0041,
+      "step": 11600
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 0.00016211393534425482,
+      "loss": 1.1471,
+      "step": 11610
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 0.00016207596955143417,
+      "loss": 1.0746,
+      "step": 11620
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 0.0001620380037586135,
+      "loss": 1.0909,
+      "step": 11630
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 0.00016200003796579284,
+      "loss": 1.1264,
+      "step": 11640
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 0.00016196207217297217,
+      "loss": 1.0273,
+      "step": 11650
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 0.0001619241063801515,
+      "loss": 1.2157,
+      "step": 11660
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 0.00016188614058733081,
+      "loss": 1.055,
+      "step": 11670
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 0.00016184817479451016,
+      "loss": 1.0756,
+      "step": 11680
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 0.0001618102090016895,
+      "loss": 1.1673,
+      "step": 11690
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 0.0001617722432088688,
+      "loss": 1.0647,
+      "step": 11700
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 0.00016173427741604813,
+      "loss": 1.1893,
+      "step": 11710
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 0.00016169631162322746,
+      "loss": 1.1097,
+      "step": 11720
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 0.0001616583458304068,
+      "loss": 1.1296,
+      "step": 11730
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 0.00016162038003758616,
+      "loss": 1.1569,
+      "step": 11740
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 0.00016158241424476548,
+      "loss": 1.0083,
+      "step": 11750
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 0.0001615444484519448,
+      "loss": 1.124,
+      "step": 11760
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 0.00016150648265912413,
+      "loss": 1.0624,
+      "step": 11770
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 0.00016146851686630345,
+      "loss": 1.1091,
+      "step": 11780
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 0.0001614305510734828,
+      "loss": 1.1548,
+      "step": 11790
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 0.00016139258528066213,
+      "loss": 1.0428,
+      "step": 11800
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 0.00016135461948784145,
+      "loss": 1.1567,
+      "step": 11810
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 0.0001613166536950208,
+      "loss": 1.085,
+      "step": 11820
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 0.00016127868790220012,
+      "loss": 1.0887,
+      "step": 11830
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 0.00016124072210937945,
+      "loss": 1.1699,
+      "step": 11840
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 0.0001612027563165588,
+      "loss": 1.0464,
+      "step": 11850
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 0.00016116479052373812,
+      "loss": 1.2486,
+      "step": 11860
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 0.00016112682473091745,
+      "loss": 1.1073,
+      "step": 11870
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 0.00016108885893809677,
+      "loss": 1.1094,
+      "step": 11880
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 0.0001610508931452761,
+      "loss": 1.1772,
+      "step": 11890
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 0.00016101292735245544,
+      "loss": 1.0588,
+      "step": 11900
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 0.0001609749615596348,
+      "loss": 1.2318,
+      "step": 11910
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 0.00016093699576681412,
+      "loss": 1.0654,
+      "step": 11920
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 0.00016089902997399344,
+      "loss": 1.0555,
+      "step": 11930
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 0.00016086106418117276,
+      "loss": 1.1654,
+      "step": 11940
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 0.0001608230983883521,
+      "loss": 1.0351,
+      "step": 11950
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 0.00016078513259553144,
+      "loss": 1.1783,
+      "step": 11960
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 0.00016074716680271076,
+      "loss": 1.1161,
+      "step": 11970
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 0.00016070920100989008,
+      "loss": 1.1364,
+      "step": 11980
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 0.00016067123521706943,
+      "loss": 1.1486,
+      "step": 11990
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 0.00016063326942424876,
+      "loss": 0.9943,
+      "step": 12000
+    },
+    {
+      "epoch": 0.66,
+      "eval_loss": 1.0959264039993286,
+      "eval_runtime": 2.0019,
+      "eval_samples_per_second": 49.952,
+      "eval_steps_per_second": 6.494,
+      "step": 12000
+    },
+    {
+      "epoch": 0.66,
+      "mmlu_eval_accuracy": NaN,
+      "mmlu_eval_accuracy_abstract_algebra": 0.45454545454545453,
+      "mmlu_eval_accuracy_anatomy": 0.42857142857142855,
+      "mmlu_eval_accuracy_astronomy": 0.125,
+      "mmlu_eval_accuracy_business_ethics": 0.6363636363636364,
+      "mmlu_eval_accuracy_clinical_knowledge": 0.4827586206896552,
+      "mmlu_eval_accuracy_college_biology": 0.8125,
+      "mmlu_eval_accuracy_college_chemistry": 1.0,
+      "mmlu_eval_accuracy_college_computer_science": 0.5454545454545454,
+      "mmlu_eval_accuracy_college_mathematics": 0.18181818181818182,
+      "mmlu_eval_accuracy_college_medicine": 0.5,
+      "mmlu_eval_accuracy_college_physics": 0.36363636363636365,
+      "mmlu_eval_accuracy_computer_security": 0.6363636363636364,
+      "mmlu_eval_accuracy_conceptual_physics": 0.42857142857142855,
+      "mmlu_eval_accuracy_econometrics": NaN,
+      "mmlu_eval_accuracy_electrical_engineering": NaN,
+      "mmlu_eval_accuracy_elementary_mathematics": NaN,
+      "mmlu_eval_accuracy_formal_logic": NaN,
+      "mmlu_eval_accuracy_global_facts": NaN,
+      "mmlu_eval_accuracy_high_school_biology": NaN,
+      "mmlu_eval_accuracy_high_school_chemistry": NaN,
+      "mmlu_eval_accuracy_high_school_computer_science": NaN,
+      "mmlu_eval_accuracy_high_school_european_history": NaN,
+      "mmlu_eval_accuracy_high_school_geography": NaN,
+      "mmlu_eval_accuracy_high_school_government_and_politics": NaN,
+      "mmlu_eval_accuracy_high_school_macroeconomics": NaN,
+      "mmlu_eval_accuracy_high_school_mathematics": NaN,
+      "mmlu_eval_accuracy_high_school_microeconomics": NaN,
+      "mmlu_eval_accuracy_high_school_physics": NaN,
+      "mmlu_eval_accuracy_high_school_psychology": NaN,
+      "mmlu_eval_accuracy_high_school_statistics": NaN,
+      "mmlu_eval_accuracy_high_school_us_history": NaN,
+      "mmlu_eval_accuracy_high_school_world_history": NaN,
+      "mmlu_eval_accuracy_human_aging": NaN,
+      "mmlu_eval_accuracy_human_sexuality": NaN,
+      "mmlu_eval_accuracy_international_law": NaN,
+      "mmlu_eval_accuracy_jurisprudence": NaN,
+      "mmlu_eval_accuracy_logical_fallacies": NaN,
+      "mmlu_eval_accuracy_machine_learning": NaN,
+      "mmlu_eval_accuracy_management": NaN,
+      "mmlu_eval_accuracy_marketing": NaN,
+      "mmlu_eval_accuracy_medical_genetics": NaN,
+      "mmlu_eval_accuracy_miscellaneous": NaN,
+      "mmlu_eval_accuracy_moral_disputes": NaN,
+      "mmlu_eval_accuracy_moral_scenarios": NaN,
+      "mmlu_eval_accuracy_nutrition": NaN,
+      "mmlu_eval_accuracy_philosophy": NaN,
+      "mmlu_eval_accuracy_prehistory": NaN,
+      "mmlu_eval_accuracy_professional_accounting": NaN,
+      "mmlu_eval_accuracy_professional_law": NaN,
+      "mmlu_eval_accuracy_professional_medicine": NaN,
+      "mmlu_eval_accuracy_professional_psychology": NaN,
+      "mmlu_eval_accuracy_public_relations": NaN,
+      "mmlu_eval_accuracy_security_studies": NaN,
+      "mmlu_eval_accuracy_sociology": NaN,
+      "mmlu_eval_accuracy_us_foreign_policy": NaN,
+      "mmlu_eval_accuracy_virology": NaN,
+      "mmlu_eval_accuracy_world_religions": NaN,
+      "mmlu_loss": 0.9960953282037129,
+      "step": 12000
+    }
+  ],
+  "max_steps": 54309,
+  "num_train_epochs": 3,
+  "total_flos": 6.187341352301232e+18,
+  "trial_name": null,
+  "trial_params": null
+}