diff --git "a/trainer_state.json" "b/trainer_state.json"
new file mode 100644--- /dev/null
+++ "b/trainer_state.json"
@@ -0,0 +1,3899 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 0.9999085672487885,
+  "eval_steps": 500,
+  "global_step": 2734,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.001828655024229679,
+      "grad_norm": 0.3227968599105643,
+      "learning_rate": 3.64963503649635e-08,
+      "loss": 1.0061,
+      "step": 5
+    },
+    {
+      "epoch": 0.003657310048459358,
+      "grad_norm": 0.31488398031235737,
+      "learning_rate": 7.2992700729927e-08,
+      "loss": 1.0006,
+      "step": 10
+    },
+    {
+      "epoch": 0.005485965072689037,
+      "grad_norm": 0.30358240428738464,
+      "learning_rate": 1.0948905109489052e-07,
+      "loss": 1.0021,
+      "step": 15
+    },
+    {
+      "epoch": 0.007314620096918716,
+      "grad_norm": 0.3111951888847597,
+      "learning_rate": 1.45985401459854e-07,
+      "loss": 0.9878,
+      "step": 20
+    },
+    {
+      "epoch": 0.009143275121148396,
+      "grad_norm": 0.3043844410155928,
+      "learning_rate": 1.824817518248175e-07,
+      "loss": 0.9817,
+      "step": 25
+    },
+    {
+      "epoch": 0.010971930145378074,
+      "grad_norm": 0.2806645638717227,
+      "learning_rate": 2.1897810218978103e-07,
+      "loss": 0.9839,
+      "step": 30
+    },
+    {
+      "epoch": 0.012800585169607753,
+      "grad_norm": 0.21723752788274575,
+      "learning_rate": 2.5547445255474454e-07,
+      "loss": 0.9995,
+      "step": 35
+    },
+    {
+      "epoch": 0.014629240193837432,
+      "grad_norm": 0.21121971623888727,
+      "learning_rate": 2.91970802919708e-07,
+      "loss": 0.9766,
+      "step": 40
+    },
+    {
+      "epoch": 0.016457895218067113,
+      "grad_norm": 0.1994903926968267,
+      "learning_rate": 3.284671532846715e-07,
+      "loss": 0.971,
+      "step": 45
+    },
+    {
+      "epoch": 0.018286550242296792,
+      "grad_norm": 0.13585393659539555,
+      "learning_rate": 3.64963503649635e-07,
+      "loss": 0.968,
+      "step": 50
+    },
+    {
+      "epoch": 0.020115205266526468,
+      "grad_norm": 0.09906459709464537,
+      "learning_rate": 4.0145985401459856e-07,
+      "loss": 0.9288,
+      "step": 55
+    },
+    {
+      "epoch": 0.021943860290756147,
+      "grad_norm": 0.0897922982991605,
+      "learning_rate": 4.3795620437956206e-07,
+      "loss": 0.9332,
+      "step": 60
+    },
+    {
+      "epoch": 0.023772515314985827,
+      "grad_norm": 0.09144553714947164,
+      "learning_rate": 4.744525547445255e-07,
+      "loss": 0.9349,
+      "step": 65
+    },
+    {
+      "epoch": 0.025601170339215506,
+      "grad_norm": 0.07460950044053644,
+      "learning_rate": 5.109489051094891e-07,
+      "loss": 0.9068,
+      "step": 70
+    },
+    {
+      "epoch": 0.027429825363445185,
+      "grad_norm": 0.06094759556233497,
+      "learning_rate": 5.474452554744526e-07,
+      "loss": 0.9152,
+      "step": 75
+    },
+    {
+      "epoch": 0.029258480387674864,
+      "grad_norm": 0.05097284304277641,
+      "learning_rate": 5.83941605839416e-07,
+      "loss": 0.8867,
+      "step": 80
+    },
+    {
+      "epoch": 0.031087135411904544,
+      "grad_norm": 0.047174301977798275,
+      "learning_rate": 6.204379562043796e-07,
+      "loss": 0.8998,
+      "step": 85
+    },
+    {
+      "epoch": 0.032915790436134226,
+      "grad_norm": 0.04604605991188836,
+      "learning_rate": 6.56934306569343e-07,
+      "loss": 0.8724,
+      "step": 90
+    },
+    {
+      "epoch": 0.034744445460363906,
+      "grad_norm": 0.04125560482905992,
+      "learning_rate": 6.934306569343066e-07,
+      "loss": 0.8873,
+      "step": 95
+    },
+    {
+      "epoch": 0.036573100484593585,
+      "grad_norm": 0.04243192358819402,
+      "learning_rate": 7.2992700729927e-07,
+      "loss": 0.887,
+      "step": 100
+    },
+    {
+      "epoch": 0.03840175550882326,
+      "grad_norm": 0.0390643929883548,
+      "learning_rate": 7.664233576642335e-07,
+      "loss": 0.8482,
+      "step": 105
+    },
+    {
+      "epoch": 0.040230410533052936,
+      "grad_norm": 0.03736783276564454,
+      "learning_rate": 8.029197080291971e-07,
+      "loss": 0.8745,
+      "step": 110
+    },
+    {
+      "epoch": 0.042059065557282616,
+      "grad_norm": 0.03715332356452642,
+      "learning_rate": 8.394160583941605e-07,
+      "loss": 0.8265,
+      "step": 115
+    },
+    {
+      "epoch": 0.043887720581512295,
+      "grad_norm": 0.03407556166200749,
+      "learning_rate": 8.759124087591241e-07,
+      "loss": 0.884,
+      "step": 120
+    },
+    {
+      "epoch": 0.045716375605741974,
+      "grad_norm": 0.03987890648675463,
+      "learning_rate": 9.124087591240875e-07,
+      "loss": 0.8193,
+      "step": 125
+    },
+    {
+      "epoch": 0.04754503062997165,
+      "grad_norm": 0.036064180788412246,
+      "learning_rate": 9.48905109489051e-07,
+      "loss": 0.8297,
+      "step": 130
+    },
+    {
+      "epoch": 0.04937368565420133,
+      "grad_norm": 0.032961357382233006,
+      "learning_rate": 9.854014598540146e-07,
+      "loss": 0.8122,
+      "step": 135
+    },
+    {
+      "epoch": 0.05120234067843101,
+      "grad_norm": 0.03058878038412802,
+      "learning_rate": 1.0218978102189781e-06,
+      "loss": 0.805,
+      "step": 140
+    },
+    {
+      "epoch": 0.05303099570266069,
+      "grad_norm": 0.03156463226109466,
+      "learning_rate": 1.0583941605839414e-06,
+      "loss": 0.7699,
+      "step": 145
+    },
+    {
+      "epoch": 0.05485965072689037,
+      "grad_norm": 0.028344907305577496,
+      "learning_rate": 1.0948905109489052e-06,
+      "loss": 0.771,
+      "step": 150
+    },
+    {
+      "epoch": 0.05668830575112005,
+      "grad_norm": 0.02988691657388487,
+      "learning_rate": 1.1313868613138687e-06,
+      "loss": 0.7397,
+      "step": 155
+    },
+    {
+      "epoch": 0.05851696077534973,
+      "grad_norm": 0.02598958807989773,
+      "learning_rate": 1.167883211678832e-06,
+      "loss": 0.7474,
+      "step": 160
+    },
+    {
+      "epoch": 0.06034561579957941,
+      "grad_norm": 0.02969219616431488,
+      "learning_rate": 1.2043795620437955e-06,
+      "loss": 0.7618,
+      "step": 165
+    },
+    {
+      "epoch": 0.06217427082380909,
+      "grad_norm": 0.026311165539023303,
+      "learning_rate": 1.2408759124087592e-06,
+      "loss": 0.7721,
+      "step": 170
+    },
+    {
+      "epoch": 0.06400292584803877,
+      "grad_norm": 0.02908702310592402,
+      "learning_rate": 1.2773722627737227e-06,
+      "loss": 0.7284,
+      "step": 175
+    },
+    {
+      "epoch": 0.06583158087226845,
+      "grad_norm": 0.025935017247548545,
+      "learning_rate": 1.313868613138686e-06,
+      "loss": 0.6984,
+      "step": 180
+    },
+    {
+      "epoch": 0.06766023589649812,
+      "grad_norm": 0.02433598370455952,
+      "learning_rate": 1.3503649635036497e-06,
+      "loss": 0.7146,
+      "step": 185
+    },
+    {
+      "epoch": 0.06948889092072781,
+      "grad_norm": 0.024029227725510947,
+      "learning_rate": 1.3868613138686132e-06,
+      "loss": 0.7261,
+      "step": 190
+    },
+    {
+      "epoch": 0.07131754594495748,
+      "grad_norm": 0.023570280422334007,
+      "learning_rate": 1.4233576642335765e-06,
+      "loss": 0.6886,
+      "step": 195
+    },
+    {
+      "epoch": 0.07314620096918717,
+      "grad_norm": 0.02338826061324417,
+      "learning_rate": 1.45985401459854e-06,
+      "loss": 0.6658,
+      "step": 200
+    },
+    {
+      "epoch": 0.07497485599341684,
+      "grad_norm": 0.022970612574749083,
+      "learning_rate": 1.4963503649635037e-06,
+      "loss": 0.6995,
+      "step": 205
+    },
+    {
+      "epoch": 0.07680351101764651,
+      "grad_norm": 0.023872629740325033,
+      "learning_rate": 1.532846715328467e-06,
+      "loss": 0.6604,
+      "step": 210
+    },
+    {
+      "epoch": 0.0786321660418762,
+      "grad_norm": 0.021611612058818175,
+      "learning_rate": 1.5693430656934305e-06,
+      "loss": 0.6674,
+      "step": 215
+    },
+    {
+      "epoch": 0.08046082106610587,
+      "grad_norm": 0.024354107318769958,
+      "learning_rate": 1.6058394160583942e-06,
+      "loss": 0.6581,
+      "step": 220
+    },
+    {
+      "epoch": 0.08228947609033556,
+      "grad_norm": 0.02175541031240366,
+      "learning_rate": 1.6423357664233577e-06,
+      "loss": 0.629,
+      "step": 225
+    },
+    {
+      "epoch": 0.08411813111456523,
+      "grad_norm": 0.022759840112146365,
+      "learning_rate": 1.678832116788321e-06,
+      "loss": 0.6553,
+      "step": 230
+    },
+    {
+      "epoch": 0.08594678613879492,
+      "grad_norm": 0.022809498688636545,
+      "learning_rate": 1.7153284671532845e-06,
+      "loss": 0.646,
+      "step": 235
+    },
+    {
+      "epoch": 0.08777544116302459,
+      "grad_norm": 0.021051368958793656,
+      "learning_rate": 1.7518248175182482e-06,
+      "loss": 0.6345,
+      "step": 240
+    },
+    {
+      "epoch": 0.08960409618725428,
+      "grad_norm": 0.019968386586809383,
+      "learning_rate": 1.7883211678832115e-06,
+      "loss": 0.6188,
+      "step": 245
+    },
+    {
+      "epoch": 0.09143275121148395,
+      "grad_norm": 0.020450382269383297,
+      "learning_rate": 1.824817518248175e-06,
+      "loss": 0.6656,
+      "step": 250
+    },
+    {
+      "epoch": 0.09326140623571363,
+      "grad_norm": 0.019514819010286597,
+      "learning_rate": 1.8613138686131388e-06,
+      "loss": 0.6272,
+      "step": 255
+    },
+    {
+      "epoch": 0.0950900612599433,
+      "grad_norm": 0.019308710229807086,
+      "learning_rate": 1.897810218978102e-06,
+      "loss": 0.6378,
+      "step": 260
+    },
+    {
+      "epoch": 0.09691871628417299,
+      "grad_norm": 0.020828774749267313,
+      "learning_rate": 1.9343065693430656e-06,
+      "loss": 0.6348,
+      "step": 265
+    },
+    {
+      "epoch": 0.09874737130840266,
+      "grad_norm": 0.023188325030436968,
+      "learning_rate": 1.9708029197080293e-06,
+      "loss": 0.6308,
+      "step": 270
+    },
+    {
+      "epoch": 0.10057602633263235,
+      "grad_norm": 0.0193044296167628,
+      "learning_rate": 1.9999991845459825e-06,
+      "loss": 0.6159,
+      "step": 275
+    },
+    {
+      "epoch": 0.10240468135686202,
+      "grad_norm": 0.020604114310436954,
+      "learning_rate": 1.9999706437950266e-06,
+      "loss": 0.6716,
+      "step": 280
+    },
+    {
+      "epoch": 0.10423333638109171,
+      "grad_norm": 0.01895504144956523,
+      "learning_rate": 1.9999013316731366e-06,
+      "loss": 0.5948,
+      "step": 285
+    },
+    {
+      "epoch": 0.10606199140532138,
+      "grad_norm": 0.019105099082900315,
+      "learning_rate": 1.999791251006346e-06,
+      "loss": 0.6256,
+      "step": 290
+    },
+    {
+      "epoch": 0.10789064642955107,
+      "grad_norm": 0.020122090316140033,
+      "learning_rate": 1.9996404062829256e-06,
+      "loss": 0.6394,
+      "step": 295
+    },
+    {
+      "epoch": 0.10971930145378074,
+      "grad_norm": 0.019799016950861798,
+      "learning_rate": 1.999448803653203e-06,
+      "loss": 0.6181,
+      "step": 300
+    },
+    {
+      "epoch": 0.11154795647801043,
+      "grad_norm": 0.018909198650251806,
+      "learning_rate": 1.9992164509293087e-06,
+      "loss": 0.6031,
+      "step": 305
+    },
+    {
+      "epoch": 0.1133766115022401,
+      "grad_norm": 0.0258707068397275,
+      "learning_rate": 1.99894335758486e-06,
+      "loss": 0.586,
+      "step": 310
+    },
+    {
+      "epoch": 0.11520526652646979,
+      "grad_norm": 0.019217347181826404,
+      "learning_rate": 1.998629534754574e-06,
+      "loss": 0.5914,
+      "step": 315
+    },
+    {
+      "epoch": 0.11703392155069946,
+      "grad_norm": 0.023216264837052865,
+      "learning_rate": 1.998274995233813e-06,
+      "loss": 0.5901,
+      "step": 320
+    },
+    {
+      "epoch": 0.11886257657492914,
+      "grad_norm": 0.020914447141611302,
+      "learning_rate": 1.9978797534780647e-06,
+      "loss": 0.6069,
+      "step": 325
+    },
+    {
+      "epoch": 0.12069123159915882,
+      "grad_norm": 0.019522354231767073,
+      "learning_rate": 1.9974438256023486e-06,
+      "loss": 0.6335,
+      "step": 330
+    },
+    {
+      "epoch": 0.1225198866233885,
+      "grad_norm": 0.01887216776559891,
+      "learning_rate": 1.9969672293805657e-06,
+      "loss": 0.6006,
+      "step": 335
+    },
+    {
+      "epoch": 0.12434854164761817,
+      "grad_norm": 0.022348487589019354,
+      "learning_rate": 1.9964499842447664e-06,
+      "loss": 0.599,
+      "step": 340
+    },
+    {
+      "epoch": 0.12617719667184785,
+      "grad_norm": 0.018983212617096703,
+      "learning_rate": 1.9958921112843635e-06,
+      "loss": 0.6156,
+      "step": 345
+    },
+    {
+      "epoch": 0.12800585169607753,
+      "grad_norm": 0.019090029410585962,
+      "learning_rate": 1.9952936332452697e-06,
+      "loss": 0.5948,
+      "step": 350
+    },
+    {
+      "epoch": 0.12983450672030722,
+      "grad_norm": 0.01897957527393132,
+      "learning_rate": 1.9946545745289724e-06,
+      "loss": 0.6194,
+      "step": 355
+    },
+    {
+      "epoch": 0.1316631617445369,
+      "grad_norm": 0.018435118132208324,
+      "learning_rate": 1.9939749611915358e-06,
+      "loss": 0.5746,
+      "step": 360
+    },
+    {
+      "epoch": 0.13349181676876656,
+      "grad_norm": 0.018276479015225607,
+      "learning_rate": 1.9932548209425406e-06,
+      "loss": 0.5835,
+      "step": 365
+    },
+    {
+      "epoch": 0.13532047179299625,
+      "grad_norm": 0.018474140717845557,
+      "learning_rate": 1.9924941831439547e-06,
+      "loss": 0.6031,
+      "step": 370
+    },
+    {
+      "epoch": 0.13714912681722594,
+      "grad_norm": 0.01941041003273915,
+      "learning_rate": 1.9916930788089343e-06,
+      "loss": 0.59,
+      "step": 375
+    },
+    {
+      "epoch": 0.13897778184145562,
+      "grad_norm": 0.01844365587200233,
+      "learning_rate": 1.9908515406005593e-06,
+      "loss": 0.5685,
+      "step": 380
+    },
+    {
+      "epoch": 0.14080643686568528,
+      "grad_norm": 0.019506979790625125,
+      "learning_rate": 1.9899696028305044e-06,
+      "loss": 0.5604,
+      "step": 385
+    },
+    {
+      "epoch": 0.14263509188991497,
+      "grad_norm": 0.01691718921569292,
+      "learning_rate": 1.9890473014576374e-06,
+      "loss": 0.5742,
+      "step": 390
+    },
+    {
+      "epoch": 0.14446374691414465,
+      "grad_norm": 0.018202080443062315,
+      "learning_rate": 1.9880846740865525e-06,
+      "loss": 0.5841,
+      "step": 395
+    },
+    {
+      "epoch": 0.14629240193837434,
+      "grad_norm": 0.018150076943080098,
+      "learning_rate": 1.98708175996604e-06,
+      "loss": 0.5933,
+      "step": 400
+    },
+    {
+      "epoch": 0.148121056962604,
+      "grad_norm": 0.020038707130759382,
+      "learning_rate": 1.986038599987484e-06,
+      "loss": 0.6,
+      "step": 405
+    },
+    {
+      "epoch": 0.14994971198683368,
+      "grad_norm": 0.018097000889198623,
+      "learning_rate": 1.984955236683196e-06,
+      "loss": 0.5665,
+      "step": 410
+    },
+    {
+      "epoch": 0.15177836701106337,
+      "grad_norm": 0.10218606291938152,
+      "learning_rate": 1.983831714224679e-06,
+      "loss": 0.5892,
+      "step": 415
+    },
+    {
+      "epoch": 0.15360702203529303,
+      "grad_norm": 0.018706208704782058,
+      "learning_rate": 1.982668078420829e-06,
+      "loss": 0.5578,
+      "step": 420
+    },
+    {
+      "epoch": 0.15543567705952271,
+      "grad_norm": 0.018757776374688825,
+      "learning_rate": 1.981464376716066e-06,
+      "loss": 0.5836,
+      "step": 425
+    },
+    {
+      "epoch": 0.1572643320837524,
+      "grad_norm": 0.018071367089126048,
+      "learning_rate": 1.980220658188399e-06,
+      "loss": 0.5816,
+      "step": 430
+    },
+    {
+      "epoch": 0.1590929871079821,
+      "grad_norm": 0.021407977080829324,
+      "learning_rate": 1.978936973547427e-06,
+      "loss": 0.5781,
+      "step": 435
+    },
+    {
+      "epoch": 0.16092164213221175,
+      "grad_norm": 0.01743601632780526,
+      "learning_rate": 1.977613375132268e-06,
+      "loss": 0.5541,
+      "step": 440
+    },
+    {
+      "epoch": 0.16275029715644143,
+      "grad_norm": 0.01821576227060148,
+      "learning_rate": 1.976249916909429e-06,
+      "loss": 0.5723,
+      "step": 445
+    },
+    {
+      "epoch": 0.16457895218067112,
+      "grad_norm": 0.017236519272688614,
+      "learning_rate": 1.974846654470602e-06,
+      "loss": 0.5951,
+      "step": 450
+    },
+    {
+      "epoch": 0.1664076072049008,
+      "grad_norm": 0.01800426853507371,
+      "learning_rate": 1.9734036450304013e-06,
+      "loss": 0.5733,
+      "step": 455
+    },
+    {
+      "epoch": 0.16823626222913046,
+      "grad_norm": 0.020818488798787296,
+      "learning_rate": 1.971920947424026e-06,
+      "loss": 0.6015,
+      "step": 460
+    },
+    {
+      "epoch": 0.17006491725336015,
+      "grad_norm": 0.019404657628813188,
+      "learning_rate": 1.9703986221048657e-06,
+      "loss": 0.5603,
+      "step": 465
+    },
+    {
+      "epoch": 0.17189357227758983,
+      "grad_norm": 0.018952882256130753,
+      "learning_rate": 1.9688367311420326e-06,
+      "loss": 0.5891,
+      "step": 470
+    },
+    {
+      "epoch": 0.17372222730181952,
+      "grad_norm": 0.01871782938489585,
+      "learning_rate": 1.967235338217832e-06,
+      "loss": 0.6058,
+      "step": 475
+    },
+    {
+      "epoch": 0.17555088232604918,
+      "grad_norm": 0.01775688287962969,
+      "learning_rate": 1.9655945086251652e-06,
+      "loss": 0.5684,
+      "step": 480
+    },
+    {
+      "epoch": 0.17737953735027887,
+      "grad_norm": 0.017832287097653172,
+      "learning_rate": 1.9639143092648686e-06,
+      "loss": 0.5287,
+      "step": 485
+    },
+    {
+      "epoch": 0.17920819237450855,
+      "grad_norm": 0.0176333670823005,
+      "learning_rate": 1.9621948086429842e-06,
+      "loss": 0.5473,
+      "step": 490
+    },
+    {
+      "epoch": 0.18103684739873824,
+      "grad_norm": 0.021880986936722392,
+      "learning_rate": 1.960436076867968e-06,
+      "loss": 0.5864,
+      "step": 495
+    },
+    {
+      "epoch": 0.1828655024229679,
+      "grad_norm": 0.018528820749694204,
+      "learning_rate": 1.958638185647831e-06,
+      "loss": 0.5673,
+      "step": 500
+    },
+    {
+      "epoch": 0.1828655024229679,
+      "eval_runtime": 91.9151,
+      "eval_samples_per_second": 10.88,
+      "eval_steps_per_second": 0.348,
+      "step": 500
+    },
+    {
+      "epoch": 0.18469415744719758,
+      "grad_norm": 0.01926855334866488,
+      "learning_rate": 1.9568012082872147e-06,
+      "loss": 0.5854,
+      "step": 505
+    },
+    {
+      "epoch": 0.18652281247142727,
+      "grad_norm": 0.017833013528665922,
+      "learning_rate": 1.9549252196844024e-06,
+      "loss": 0.5427,
+      "step": 510
+    },
+    {
+      "epoch": 0.18835146749565695,
+      "grad_norm": 0.017981694332317954,
+      "learning_rate": 1.953010296328268e-06,
+      "loss": 0.5728,
+      "step": 515
+    },
+    {
+      "epoch": 0.1901801225198866,
+      "grad_norm": 0.018449282619064117,
+      "learning_rate": 1.9510565162951534e-06,
+      "loss": 0.5464,
+      "step": 520
+    },
+    {
+      "epoch": 0.1920087775441163,
+      "grad_norm": 0.018822846854912784,
+      "learning_rate": 1.9490639592456878e-06,
+      "loss": 0.5663,
+      "step": 525
+    },
+    {
+      "epoch": 0.19383743256834599,
+      "grad_norm": 0.017826488258388437,
+      "learning_rate": 1.9470327064215382e-06,
+      "loss": 0.5638,
+      "step": 530
+    },
+    {
+      "epoch": 0.19566608759257567,
+      "grad_norm": 0.017707034906832304,
+      "learning_rate": 1.9449628406420987e-06,
+      "loss": 0.5584,
+      "step": 535
+    },
+    {
+      "epoch": 0.19749474261680533,
+      "grad_norm": 0.01777596831107123,
+      "learning_rate": 1.9428544463011124e-06,
+      "loss": 0.5421,
+      "step": 540
+    },
+    {
+      "epoch": 0.19932339764103502,
+      "grad_norm": 0.018796275656722917,
+      "learning_rate": 1.94070760936323e-06,
+      "loss": 0.5773,
+      "step": 545
+    },
+    {
+      "epoch": 0.2011520526652647,
+      "grad_norm": 0.023809951913604018,
+      "learning_rate": 1.938522417360507e-06,
+      "loss": 0.5886,
+      "step": 550
+    },
+    {
+      "epoch": 0.2029807076894944,
+      "grad_norm": 0.01856239244366707,
+      "learning_rate": 1.936298959388832e-06,
+      "loss": 0.5427,
+      "step": 555
+    },
+    {
+      "epoch": 0.20480936271372405,
+      "grad_norm": 0.019342573858811985,
+      "learning_rate": 1.934037326104296e-06,
+      "loss": 0.6014,
+      "step": 560
+    },
+    {
+      "epoch": 0.20663801773795373,
+      "grad_norm": 0.019254423134392695,
+      "learning_rate": 1.931737609719496e-06,
+      "loss": 0.5832,
+      "step": 565
+    },
+    {
+      "epoch": 0.20846667276218342,
+      "grad_norm": 0.0172347233240028,
+      "learning_rate": 1.9293999039997745e-06,
+      "loss": 0.5857,
+      "step": 570
+    },
+    {
+      "epoch": 0.2102953277864131,
+      "grad_norm": 0.018517717683554927,
+      "learning_rate": 1.9270243042593962e-06,
+      "loss": 0.5576,
+      "step": 575
+    },
+    {
+      "epoch": 0.21212398281064276,
+      "grad_norm": 0.021411761027241353,
+      "learning_rate": 1.9246109073576627e-06,
+      "loss": 0.5432,
+      "step": 580
+    },
+    {
+      "epoch": 0.21395263783487245,
+      "grad_norm": 0.018355899159154337,
+      "learning_rate": 1.922159811694963e-06,
+      "loss": 0.5547,
+      "step": 585
+    },
+    {
+      "epoch": 0.21578129285910214,
+      "grad_norm": 0.027979414131266363,
+      "learning_rate": 1.9196711172087616e-06,
+      "loss": 0.5631,
+      "step": 590
+    },
+    {
+      "epoch": 0.21760994788333182,
+      "grad_norm": 0.017930067641649478,
+      "learning_rate": 1.9171449253695233e-06,
+      "loss": 0.5511,
+      "step": 595
+    },
+    {
+      "epoch": 0.21943860290756148,
+      "grad_norm": 0.017758784178381903,
+      "learning_rate": 1.914581339176576e-06,
+      "loss": 0.5728,
+      "step": 600
+    },
+    {
+      "epoch": 0.22126725793179117,
+      "grad_norm": 0.017899595842170893,
+      "learning_rate": 1.9119804631539114e-06,
+      "loss": 0.5768,
+      "step": 605
+    },
+    {
+      "epoch": 0.22309591295602085,
+      "grad_norm": 0.022623835387436123,
+      "learning_rate": 1.9093424033459247e-06,
+      "loss": 0.5333,
+      "step": 610
+    },
+    {
+      "epoch": 0.22492456798025054,
+      "grad_norm": 0.017241733000080472,
+      "learning_rate": 1.9066672673130878e-06,
+      "loss": 0.5414,
+      "step": 615
+    },
+    {
+      "epoch": 0.2267532230044802,
+      "grad_norm": 0.018096792246879885,
+      "learning_rate": 1.903955164127566e-06,
+      "loss": 0.5605,
+      "step": 620
+    },
+    {
+      "epoch": 0.22858187802870988,
+      "grad_norm": 0.01696067842004016,
+      "learning_rate": 1.9012062043687708e-06,
+      "loss": 0.5453,
+      "step": 625
+    },
+    {
+      "epoch": 0.23041053305293957,
+      "grad_norm": 0.01794910275393239,
+      "learning_rate": 1.8984205001188505e-06,
+      "loss": 0.5639,
+      "step": 630
+    },
+    {
+      "epoch": 0.23223918807716923,
+      "grad_norm": 0.017127765980053215,
+      "learning_rate": 1.8955981649581203e-06,
+      "loss": 0.5436,
+      "step": 635
+    },
+    {
+      "epoch": 0.23406784310139891,
+      "grad_norm": 0.017399800780216225,
+      "learning_rate": 1.8927393139604323e-06,
+      "loss": 0.5415,
+      "step": 640
+    },
+    {
+      "epoch": 0.2358964981256286,
+      "grad_norm": 0.01734636707972079,
+      "learning_rate": 1.8898440636884834e-06,
+      "loss": 0.5724,
+      "step": 645
+    },
+    {
+      "epoch": 0.2377251531498583,
+      "grad_norm": 0.01750054858979947,
+      "learning_rate": 1.886912532189061e-06,
+      "loss": 0.5651,
+      "step": 650
+    },
+    {
+      "epoch": 0.23955380817408795,
+      "grad_norm": 0.01693918950966943,
+      "learning_rate": 1.8839448389882317e-06,
+      "loss": 0.5503,
+      "step": 655
+    },
+    {
+      "epoch": 0.24138246319831763,
+      "grad_norm": 0.02237192097108235,
+      "learning_rate": 1.8809411050864691e-06,
+      "loss": 0.5485,
+      "step": 660
+    },
+    {
+      "epoch": 0.24321111822254732,
+      "grad_norm": 0.016941308766112398,
+      "learning_rate": 1.8779014529537172e-06,
+      "loss": 0.5602,
+      "step": 665
+    },
+    {
+      "epoch": 0.245039773246777,
+      "grad_norm": 0.018115129345918107,
+      "learning_rate": 1.8748260065243984e-06,
+      "loss": 0.5298,
+      "step": 670
+    },
+    {
+      "epoch": 0.24686842827100666,
+      "grad_norm": 0.01724709801350698,
+      "learning_rate": 1.8717148911923611e-06,
+      "loss": 0.553,
+      "step": 675
+    },
+    {
+      "epoch": 0.24869708329523635,
+      "grad_norm": 0.017589496791373973,
+      "learning_rate": 1.8685682338057664e-06,
+      "loss": 0.5657,
+      "step": 680
+    },
+    {
+      "epoch": 0.250525738319466,
+      "grad_norm": 0.016640608019418836,
+      "learning_rate": 1.8653861626619164e-06,
+      "loss": 0.56,
+      "step": 685
+    },
+    {
+      "epoch": 0.2523543933436957,
+      "grad_norm": 0.018902436433412717,
+      "learning_rate": 1.8621688075020226e-06,
+      "loss": 0.5641,
+      "step": 690
+    },
+    {
+      "epoch": 0.2541830483679254,
+      "grad_norm": 0.01808077158932347,
+      "learning_rate": 1.8589162995059163e-06,
+      "loss": 0.5122,
+      "step": 695
+    },
+    {
+      "epoch": 0.25601170339215507,
+      "grad_norm": 0.017396380550086406,
+      "learning_rate": 1.8556287712867005e-06,
+      "loss": 0.5413,
+      "step": 700
+    },
+    {
+      "epoch": 0.25784035841638475,
+      "grad_norm": 0.01806921446796445,
+      "learning_rate": 1.852306356885343e-06,
+      "loss": 0.5669,
+      "step": 705
+    },
+    {
+      "epoch": 0.25966901344061444,
+      "grad_norm": 0.01708498051929725,
+      "learning_rate": 1.84894919176521e-06,
+      "loss": 0.5475,
+      "step": 710
+    },
+    {
+      "epoch": 0.2614976684648441,
+      "grad_norm": 0.5354487931983631,
+      "learning_rate": 1.8455574128065448e-06,
+      "loss": 0.5559,
+      "step": 715
+    },
+    {
+      "epoch": 0.2633263234890738,
+      "grad_norm": 0.017660267567324688,
+      "learning_rate": 1.8421311583008846e-06,
+      "loss": 0.5907,
+      "step": 720
+    },
+    {
+      "epoch": 0.26515497851330344,
+      "grad_norm": 0.0185720719455755,
+      "learning_rate": 1.838670567945424e-06,
+      "loss": 0.5446,
+      "step": 725
+    },
+    {
+      "epoch": 0.2669836335375331,
+      "grad_norm": 0.017276383567725914,
+      "learning_rate": 1.835175782837318e-06,
+      "loss": 0.5233,
+      "step": 730
+    },
+    {
+      "epoch": 0.2688122885617628,
+      "grad_norm": 0.017330844808857257,
+      "learning_rate": 1.8316469454679297e-06,
+      "loss": 0.5286,
+      "step": 735
+    },
+    {
+      "epoch": 0.2706409435859925,
+      "grad_norm": 0.01996239596703842,
+      "learning_rate": 1.82808419971702e-06,
+      "loss": 0.542,
+      "step": 740
+    },
+    {
+      "epoch": 0.2724695986102222,
+      "grad_norm": 0.10517769141625169,
+      "learning_rate": 1.8244876908468823e-06,
+      "loss": 0.5332,
+      "step": 745
+    },
+    {
+      "epoch": 0.27429825363445187,
+      "grad_norm": 0.018487213723535005,
+      "learning_rate": 1.8208575654964178e-06,
+      "loss": 0.5471,
+      "step": 750
+    },
+    {
+      "epoch": 0.27612690865868156,
+      "grad_norm": 0.019123389142702233,
+      "learning_rate": 1.8171939716751594e-06,
+      "loss": 0.553,
+      "step": 755
+    },
+    {
+      "epoch": 0.27795556368291124,
+      "grad_norm": 0.0172420909816813,
+      "learning_rate": 1.8134970587572343e-06,
+      "loss": 0.5406,
+      "step": 760
+    },
+    {
+      "epoch": 0.2797842187071409,
+      "grad_norm": 0.018205866563671992,
+      "learning_rate": 1.8097669774752757e-06,
+      "loss": 0.5406,
+      "step": 765
+    },
+    {
+      "epoch": 0.28161287373137056,
+      "grad_norm": 0.01943978439497101,
+      "learning_rate": 1.8060038799142757e-06,
+      "loss": 0.5329,
+      "step": 770
+    },
+    {
+      "epoch": 0.28344152875560025,
+      "grad_norm": 0.017465953878362427,
+      "learning_rate": 1.8022079195053847e-06,
+      "loss": 0.5376,
+      "step": 775
+    },
+    {
+      "epoch": 0.28527018377982993,
+      "grad_norm": 0.017943158814988033,
+      "learning_rate": 1.7983792510196565e-06,
+      "loss": 0.5278,
+      "step": 780
+    },
+    {
+      "epoch": 0.2870988388040596,
+      "grad_norm": 0.017350208213509168,
+      "learning_rate": 1.7945180305617365e-06,
+      "loss": 0.5832,
+      "step": 785
+    },
+    {
+      "epoch": 0.2889274938282893,
+      "grad_norm": 0.0169590330643545,
+      "learning_rate": 1.790624415563498e-06,
+      "loss": 0.52,
+      "step": 790
+    },
+    {
+      "epoch": 0.290756148852519,
+      "grad_norm": 0.018001402384542953,
+      "learning_rate": 1.7866985647776223e-06,
+      "loss": 0.5691,
+      "step": 795
+    },
+    {
+      "epoch": 0.2925848038767487,
+      "grad_norm": 0.01746917288567252,
+      "learning_rate": 1.782740638271128e-06,
+      "loss": 0.558,
+      "step": 800
+    },
+    {
+      "epoch": 0.2944134589009783,
+      "grad_norm": 0.017038824913059244,
+      "learning_rate": 1.7787507974188408e-06,
+      "loss": 0.5234,
+      "step": 805
+    },
+    {
+      "epoch": 0.296242113925208,
+      "grad_norm": 0.017766609280282786,
+      "learning_rate": 1.7747292048968186e-06,
+      "loss": 0.5274,
+      "step": 810
+    },
+    {
+      "epoch": 0.2980707689494377,
+      "grad_norm": 0.01639240260511638,
+      "learning_rate": 1.7706760246757151e-06,
+      "loss": 0.5375,
+      "step": 815
+    },
+    {
+      "epoch": 0.29989942397366737,
+      "grad_norm": 0.01722424255408644,
+      "learning_rate": 1.766591422014096e-06,
+      "loss": 0.518,
+      "step": 820
+    },
+    {
+      "epoch": 0.30172807899789705,
+      "grad_norm": 0.017591990835254444,
+      "learning_rate": 1.7624755634517003e-06,
+      "loss": 0.5595,
+      "step": 825
+    },
+    {
+      "epoch": 0.30355673402212674,
+      "grad_norm": 0.016412558614787714,
+      "learning_rate": 1.7583286168026508e-06,
+      "loss": 0.539,
+      "step": 830
+    },
+    {
+      "epoch": 0.3053853890463564,
+      "grad_norm": 0.01746152028260052,
+      "learning_rate": 1.7541507511486113e-06,
+      "loss": 0.5376,
+      "step": 835
+    },
+    {
+      "epoch": 0.30721404407058606,
+      "grad_norm": 0.01753975494196402,
+      "learning_rate": 1.7499421368318924e-06,
+      "loss": 0.5296,
+      "step": 840
+    },
+    {
+      "epoch": 0.30904269909481574,
+      "grad_norm": 0.017159318477811224,
+      "learning_rate": 1.7457029454485066e-06,
+      "loss": 0.5355,
+      "step": 845
+    },
+    {
+      "epoch": 0.31087135411904543,
+      "grad_norm": 0.016275051040046486,
+      "learning_rate": 1.741433349841173e-06,
+      "loss": 0.5458,
+      "step": 850
+    },
+    {
+      "epoch": 0.3127000091432751,
+      "grad_norm": 0.017333324363424767,
+      "learning_rate": 1.7371335240922683e-06,
+      "loss": 0.5643,
+      "step": 855
+    },
+    {
+      "epoch": 0.3145286641675048,
+      "grad_norm": 0.01807303742926882,
+      "learning_rate": 1.7328036435167286e-06,
+      "loss": 0.544,
+      "step": 860
+    },
+    {
+      "epoch": 0.3163573191917345,
+      "grad_norm": 0.01840024338855868,
+      "learning_rate": 1.7284438846549043e-06,
+      "loss": 0.5311,
+      "step": 865
+    },
+    {
+      "epoch": 0.3181859742159642,
+      "grad_norm": 0.017090241867083487,
+      "learning_rate": 1.7240544252653594e-06,
+      "loss": 0.5229,
+      "step": 870
+    },
+    {
+      "epoch": 0.32001462924019386,
+      "grad_norm": 0.017003725321295255,
+      "learning_rate": 1.7196354443176247e-06,
+      "loss": 0.5507,
+      "step": 875
+    },
+    {
+      "epoch": 0.3218432842644235,
+      "grad_norm": 0.016900325005819092,
+      "learning_rate": 1.7151871219849006e-06,
+      "loss": 0.5169,
+      "step": 880
+    },
+    {
+      "epoch": 0.3236719392886532,
+      "grad_norm": 0.015849786576054736,
+      "learning_rate": 1.7107096396367116e-06,
+      "loss": 0.5174,
+      "step": 885
+    },
+    {
+      "epoch": 0.32550059431288286,
+      "grad_norm": 0.017500050684170834,
+      "learning_rate": 1.7062031798315099e-06,
+      "loss": 0.541,
+      "step": 890
+    },
+    {
+      "epoch": 0.32732924933711255,
+      "grad_norm": 0.017695770503772665,
+      "learning_rate": 1.7016679263092349e-06,
+      "loss": 0.5386,
+      "step": 895
+    },
+    {
+      "epoch": 0.32915790436134224,
+      "grad_norm": 0.01774455517825637,
+      "learning_rate": 1.6971040639838174e-06,
+      "loss": 0.5432,
+      "step": 900
+    },
+    {
+      "epoch": 0.3309865593855719,
+      "grad_norm": 0.016404044762877424,
+      "learning_rate": 1.6925117789356445e-06,
+      "loss": 0.5478,
+      "step": 905
+    },
+    {
+      "epoch": 0.3328152144098016,
+      "grad_norm": 0.027545162607687242,
+      "learning_rate": 1.6878912584039697e-06,
+      "loss": 0.5424,
+      "step": 910
+    },
+    {
+      "epoch": 0.3346438694340313,
+      "grad_norm": 0.01895891550461197,
+      "learning_rate": 1.6832426907792804e-06,
+      "loss": 0.5408,
+      "step": 915
+    },
+    {
+      "epoch": 0.3364725244582609,
+      "grad_norm": 0.017189740662575725,
+      "learning_rate": 1.6785662655956145e-06,
+      "loss": 0.5354,
+      "step": 920
+    },
+    {
+      "epoch": 0.3383011794824906,
+      "grad_norm": 0.016943841216029255,
+      "learning_rate": 1.6738621735228363e-06,
+      "loss": 0.549,
+      "step": 925
+    },
+    {
+      "epoch": 0.3401298345067203,
+      "grad_norm": 0.01674613231171939,
+      "learning_rate": 1.669130606358858e-06,
+      "loss": 0.5026,
+      "step": 930
+    },
+    {
+      "epoch": 0.34195848953095,
+      "grad_norm": 0.017268749046049962,
+      "learning_rate": 1.6643717570218238e-06,
+      "loss": 0.5163,
+      "step": 935
+    },
+    {
+      "epoch": 0.34378714455517967,
+      "grad_norm": 0.017271996132854857,
+      "learning_rate": 1.659585819542241e-06,
+      "loss": 0.5261,
+      "step": 940
+    },
+    {
+      "epoch": 0.34561579957940936,
+      "grad_norm": 0.016275986656171422,
+      "learning_rate": 1.65477298905507e-06,
+      "loss": 0.5151,
+      "step": 945
+    },
+    {
+      "epoch": 0.34744445460363904,
+      "grad_norm": 0.019123048788353147,
+      "learning_rate": 1.6499334617917687e-06,
+      "loss": 0.5204,
+      "step": 950
+    },
+    {
+      "epoch": 0.3492731096278687,
+      "grad_norm": 0.01660199903676317,
+      "learning_rate": 1.6450674350722905e-06,
+      "loss": 0.5102,
+      "step": 955
+    },
+    {
+      "epoch": 0.35110176465209836,
+      "grad_norm": 0.01788917493473574,
+      "learning_rate": 1.6401751072970407e-06,
+      "loss": 0.5239,
+      "step": 960
+    },
+    {
+      "epoch": 0.35293041967632804,
+      "grad_norm": 0.01798730437139649,
+      "learning_rate": 1.635256677938785e-06,
+      "loss": 0.5122,
+      "step": 965
+    },
+    {
+      "epoch": 0.35475907470055773,
+      "grad_norm": 0.01689396852560636,
+      "learning_rate": 1.630312347534518e-06,
+      "loss": 0.5312,
+      "step": 970
+    },
+    {
+      "epoch": 0.3565877297247874,
+      "grad_norm": 0.017652200481013784,
+      "learning_rate": 1.6253423176772871e-06,
+      "loss": 0.5569,
+      "step": 975
+    },
+    {
+      "epoch": 0.3584163847490171,
+      "grad_norm": 0.01739896068962979,
+      "learning_rate": 1.620346791007972e-06,
+      "loss": 0.5314,
+      "step": 980
+    },
+    {
+      "epoch": 0.3602450397732468,
+      "grad_norm": 0.018990827717724194,
+      "learning_rate": 1.6153259712070224e-06,
+      "loss": 0.5497,
+      "step": 985
+    },
+    {
+      "epoch": 0.3620736947974765,
+      "grad_norm": 0.017609123156074134,
+      "learning_rate": 1.6102800629861545e-06,
+      "loss": 0.5004,
+      "step": 990
+    },
+    {
+      "epoch": 0.36390234982170616,
+      "grad_norm": 0.016758715592028146,
+      "learning_rate": 1.6052092720800036e-06,
+      "loss": 0.521,
+      "step": 995
+    },
+    {
+      "epoch": 0.3657310048459358,
+      "grad_norm": 0.017061718198161494,
+      "learning_rate": 1.6001138052377367e-06,
+      "loss": 0.5269,
+      "step": 1000
+    },
+    {
+      "epoch": 0.3657310048459358,
+      "eval_runtime": 91.1713,
+      "eval_samples_per_second": 10.968,
+      "eval_steps_per_second": 0.351,
+      "step": 1000
+    },
+    {
+      "epoch": 0.3675596598701655,
+      "grad_norm": 0.026439825766192075,
+      "learning_rate": 1.5949938702146203e-06,
+      "loss": 0.5404,
+      "step": 1005
+    },
+    {
+      "epoch": 0.36938831489439516,
+      "grad_norm": 0.017833026829326967,
+      "learning_rate": 1.5898496757635535e-06,
+      "loss": 0.5326,
+      "step": 1010
+    },
+    {
+      "epoch": 0.37121696991862485,
+      "grad_norm": 0.01717885160020913,
+      "learning_rate": 1.5846814316265532e-06,
+      "loss": 0.5312,
+      "step": 1015
+    },
+    {
+      "epoch": 0.37304562494285454,
+      "grad_norm": 0.01959129000780149,
+      "learning_rate": 1.5794893485262044e-06,
+      "loss": 0.5128,
+      "step": 1020
+    },
+    {
+      "epoch": 0.3748742799670842,
+      "grad_norm": 0.017071760527829126,
+      "learning_rate": 1.5742736381570672e-06,
+      "loss": 0.5545,
+      "step": 1025
+    },
+    {
+      "epoch": 0.3767029349913139,
+      "grad_norm": 0.016616879092010567,
+      "learning_rate": 1.5690345131770473e-06,
+      "loss": 0.5133,
+      "step": 1030
+    },
+    {
+      "epoch": 0.3785315900155436,
+      "grad_norm": 0.017421876511450714,
+      "learning_rate": 1.5637721871987228e-06,
+      "loss": 0.5662,
+      "step": 1035
+    },
+    {
+      "epoch": 0.3803602450397732,
+      "grad_norm": 0.016804992647736335,
+      "learning_rate": 1.5584868747806372e-06,
+      "loss": 0.5116,
+      "step": 1040
+    },
+    {
+      "epoch": 0.3821889000640029,
+      "grad_norm": 0.01698706176364593,
+      "learning_rate": 1.5531787914185497e-06,
+      "loss": 0.5218,
+      "step": 1045
+    },
+    {
+      "epoch": 0.3840175550882326,
+      "grad_norm": 0.016446976366545147,
+      "learning_rate": 1.5478481535366493e-06,
+      "loss": 0.5474,
+      "step": 1050
+    },
+    {
+      "epoch": 0.3858462101124623,
+      "grad_norm": 0.0172014312252115,
+      "learning_rate": 1.5424951784787307e-06,
+      "loss": 0.5753,
+      "step": 1055
+    },
+    {
+      "epoch": 0.38767486513669197,
+      "grad_norm": 0.016978295326201723,
+      "learning_rate": 1.537120084499333e-06,
+      "loss": 0.5395,
+      "step": 1060
+    },
+    {
+      "epoch": 0.38950352016092166,
+      "grad_norm": 0.017222675992629395,
+      "learning_rate": 1.53172309075484e-06,
+      "loss": 0.5458,
+      "step": 1065
+    },
+    {
+      "epoch": 0.39133217518515134,
+      "grad_norm": 0.018524331600037174,
+      "learning_rate": 1.5263044172945455e-06,
+      "loss": 0.5453,
+      "step": 1070
+    },
+    {
+      "epoch": 0.393160830209381,
+      "grad_norm": 0.017984727345243433,
+      "learning_rate": 1.5208642850516805e-06,
+      "loss": 0.5314,
+      "step": 1075
+    },
+    {
+      "epoch": 0.39498948523361066,
+      "grad_norm": 0.016505747515556748,
+      "learning_rate": 1.5154029158344065e-06,
+      "loss": 0.5226,
+      "step": 1080
+    },
+    {
+      "epoch": 0.39681814025784035,
+      "grad_norm": 0.01792882593381087,
+      "learning_rate": 1.5099205323167707e-06,
+      "loss": 0.5211,
+      "step": 1085
+    },
+    {
+      "epoch": 0.39864679528207003,
+      "grad_norm": 0.018749962417192022,
+      "learning_rate": 1.5044173580296265e-06,
+      "loss": 0.5636,
+      "step": 1090
+    },
+    {
+      "epoch": 0.4004754503062997,
+      "grad_norm": 0.01779289276778167,
+      "learning_rate": 1.4988936173515212e-06,
+      "loss": 0.5315,
+      "step": 1095
+    },
+    {
+      "epoch": 0.4023041053305294,
+      "grad_norm": 0.01712555873182193,
+      "learning_rate": 1.4933495354995472e-06,
+      "loss": 0.5656,
+      "step": 1100
+    },
+    {
+      "epoch": 0.4041327603547591,
+      "grad_norm": 0.017990217136661064,
+      "learning_rate": 1.4877853385201567e-06,
+      "loss": 0.527,
+      "step": 1105
+    },
+    {
+      "epoch": 0.4059614153789888,
+      "grad_norm": 0.017613449110706392,
+      "learning_rate": 1.4822012532799496e-06,
+      "loss": 0.5265,
+      "step": 1110
+    },
+    {
+      "epoch": 0.4077900704032184,
+      "grad_norm": 0.01703645287625673,
+      "learning_rate": 1.4765975074564202e-06,
+      "loss": 0.5054,
+      "step": 1115
+    },
+    {
+      "epoch": 0.4096187254274481,
+      "grad_norm": 0.03485083253319602,
+      "learning_rate": 1.470974329528675e-06,
+      "loss": 0.4972,
+      "step": 1120
+    },
+    {
+      "epoch": 0.4114473804516778,
+      "grad_norm": 0.016703063382188122,
+      "learning_rate": 1.4653319487681179e-06,
+      "loss": 0.4973,
+      "step": 1125
+    },
+    {
+      "epoch": 0.41327603547590747,
+      "grad_norm": 0.018124627758881625,
+      "learning_rate": 1.4596705952291016e-06,
+      "loss": 0.5493,
+      "step": 1130
+    },
+    {
+      "epoch": 0.41510469050013715,
+      "grad_norm": 0.017664316342146474,
+      "learning_rate": 1.4539904997395467e-06,
+      "loss": 0.5097,
+      "step": 1135
+    },
+    {
+      "epoch": 0.41693334552436684,
+      "grad_norm": 0.016644127477525594,
+      "learning_rate": 1.448291893891533e-06,
+      "loss": 0.5144,
+      "step": 1140
+    },
+    {
+      "epoch": 0.4187620005485965,
+      "grad_norm": 0.015919864749656256,
+      "learning_rate": 1.4425750100318533e-06,
+      "loss": 0.5141,
+      "step": 1145
+    },
+    {
+      "epoch": 0.4205906555728262,
+      "grad_norm": 0.01674336654176477,
+      "learning_rate": 1.4368400812525433e-06,
+      "loss": 0.4864,
+      "step": 1150
+    },
+    {
+      "epoch": 0.42241931059705584,
+      "grad_norm": 0.016582970266558143,
+      "learning_rate": 1.4310873413813754e-06,
+      "loss": 0.5193,
+      "step": 1155
+    },
+    {
+      "epoch": 0.4242479656212855,
+      "grad_norm": 0.017357555009786176,
+      "learning_rate": 1.4253170249723268e-06,
+      "loss": 0.5467,
+      "step": 1160
+    },
+    {
+      "epoch": 0.4260766206455152,
+      "grad_norm": 0.017392468171085116,
+      "learning_rate": 1.4195293672960148e-06,
+      "loss": 0.5202,
+      "step": 1165
+    },
+    {
+      "epoch": 0.4279052756697449,
+      "grad_norm": 0.016996017796632703,
+      "learning_rate": 1.413724604330104e-06,
+      "loss": 0.5203,
+      "step": 1170
+    },
+    {
+      "epoch": 0.4297339306939746,
+      "grad_norm": 0.01708376504860417,
+      "learning_rate": 1.4079029727496876e-06,
+      "loss": 0.5212,
+      "step": 1175
+    },
+    {
+      "epoch": 0.4315625857182043,
+      "grad_norm": 0.01935398417137648,
+      "learning_rate": 1.4020647099176337e-06,
+      "loss": 0.5704,
+      "step": 1180
+    },
+    {
+      "epoch": 0.43339124074243396,
+      "grad_norm": 0.019564758523905695,
+      "learning_rate": 1.3962100538749097e-06,
+      "loss": 0.5384,
+      "step": 1185
+    },
+    {
+      "epoch": 0.43521989576666364,
+      "grad_norm": 0.022048449448667892,
+      "learning_rate": 1.3903392433308763e-06,
+      "loss": 0.5434,
+      "step": 1190
+    },
+    {
+      "epoch": 0.4370485507908933,
+      "grad_norm": 0.016785376754885464,
+      "learning_rate": 1.3844525176535554e-06,
+      "loss": 0.5038,
+      "step": 1195
+    },
+    {
+      "epoch": 0.43887720581512296,
+      "grad_norm": 0.017516689226408873,
+      "learning_rate": 1.3785501168598686e-06,
+      "loss": 0.5266,
+      "step": 1200
+    },
+    {
+      "epoch": 0.44070586083935265,
+      "grad_norm": 0.018130528509261296,
+      "learning_rate": 1.3726322816058531e-06,
+      "loss": 0.4873,
+      "step": 1205
+    },
+    {
+      "epoch": 0.44253451586358233,
+      "grad_norm": 0.016659481564096815,
+      "learning_rate": 1.366699253176848e-06,
+      "loss": 0.5145,
+      "step": 1210
+    },
+    {
+      "epoch": 0.444363170887812,
+      "grad_norm": 0.016932392078214457,
+      "learning_rate": 1.360751273477658e-06,
+      "loss": 0.5442,
+      "step": 1215
+    },
+    {
+      "epoch": 0.4461918259120417,
+      "grad_norm": 0.017397635531264867,
+      "learning_rate": 1.3547885850226888e-06,
+      "loss": 0.5258,
+      "step": 1220
+    },
+    {
+      "epoch": 0.4480204809362714,
+      "grad_norm": 0.017305372507747333,
+      "learning_rate": 1.3488114309260601e-06,
+      "loss": 0.5231,
+      "step": 1225
+    },
+    {
+      "epoch": 0.4498491359605011,
+      "grad_norm": 0.017739316974001593,
+      "learning_rate": 1.342820054891693e-06,
+      "loss": 0.5052,
+      "step": 1230
+    },
+    {
+      "epoch": 0.4516777909847307,
+      "grad_norm": 0.01865225864844168,
+      "learning_rate": 1.3368147012033738e-06,
+      "loss": 0.5331,
+      "step": 1235
+    },
+    {
+      "epoch": 0.4535064460089604,
+      "grad_norm": 0.017305883692044065,
+      "learning_rate": 1.3307956147147923e-06,
+      "loss": 0.528,
+      "step": 1240
+    },
+    {
+      "epoch": 0.4553351010331901,
+      "grad_norm": 0.016404612508973574,
+      "learning_rate": 1.3247630408395617e-06,
+      "loss": 0.495,
+      "step": 1245
+    },
+    {
+      "epoch": 0.45716375605741977,
+      "grad_norm": 0.017129542073334014,
+      "learning_rate": 1.3187172255412095e-06,
+      "loss": 0.5629,
+      "step": 1250
+    },
+    {
+      "epoch": 0.45899241108164945,
+      "grad_norm": 0.017055721679601482,
+      "learning_rate": 1.3126584153231496e-06,
+      "loss": 0.5137,
+      "step": 1255
+    },
+    {
+      "epoch": 0.46082106610587914,
+      "grad_norm": 0.01769589755715341,
+      "learning_rate": 1.306586857218634e-06,
+      "loss": 0.5277,
+      "step": 1260
+    },
+    {
+      "epoch": 0.4626497211301088,
+      "grad_norm": 0.016192120011367772,
+      "learning_rate": 1.3005027987806767e-06,
+      "loss": 0.483,
+      "step": 1265
+    },
+    {
+      "epoch": 0.46447837615433846,
+      "grad_norm": 0.016933244205414207,
+      "learning_rate": 1.2944064880719632e-06,
+      "loss": 0.5426,
+      "step": 1270
+    },
+    {
+      "epoch": 0.46630703117856814,
+      "grad_norm": 0.016910956762961723,
+      "learning_rate": 1.288298173654736e-06,
+      "loss": 0.4975,
+      "step": 1275
+    },
+    {
+      "epoch": 0.46813568620279783,
+      "grad_norm": 0.017012738283461247,
+      "learning_rate": 1.282178104580659e-06,
+      "loss": 0.5552,
+      "step": 1280
+    },
+    {
+      "epoch": 0.4699643412270275,
+      "grad_norm": 0.02465811701467098,
+      "learning_rate": 1.2760465303806638e-06,
+      "loss": 0.5186,
+      "step": 1285
+    },
+    {
+      "epoch": 0.4717929962512572,
+      "grad_norm": 0.01699415071193306,
+      "learning_rate": 1.2699037010547758e-06,
+      "loss": 0.5179,
+      "step": 1290
+    },
+    {
+      "epoch": 0.4736216512754869,
+      "grad_norm": 0.029237187097736865,
+      "learning_rate": 1.26374986706192e-06,
+      "loss": 0.5027,
+      "step": 1295
+    },
+    {
+      "epoch": 0.4754503062997166,
+      "grad_norm": 0.01779652541996131,
+      "learning_rate": 1.2575852793097111e-06,
+      "loss": 0.5208,
+      "step": 1300
+    },
+    {
+      "epoch": 0.47727896132394626,
+      "grad_norm": 0.01720496967801305,
+      "learning_rate": 1.2514101891442211e-06,
+      "loss": 0.5695,
+      "step": 1305
+    },
+    {
+      "epoch": 0.4791076163481759,
+      "grad_norm": 0.0166642165550301,
+      "learning_rate": 1.245224848339733e-06,
+      "loss": 0.5422,
+      "step": 1310
+    },
+    {
+      "epoch": 0.4809362713724056,
+      "grad_norm": 0.016881306758524717,
+      "learning_rate": 1.239029509088475e-06,
+      "loss": 0.5004,
+      "step": 1315
+    },
+    {
+      "epoch": 0.48276492639663526,
+      "grad_norm": 0.017110705125781858,
+      "learning_rate": 1.232824423990336e-06,
+      "loss": 0.5282,
+      "step": 1320
+    },
+    {
+      "epoch": 0.48459358142086495,
+      "grad_norm": 0.016728308702345526,
+      "learning_rate": 1.2266098460425704e-06,
+      "loss": 0.5374,
+      "step": 1325
+    },
+    {
+      "epoch": 0.48642223644509464,
+      "grad_norm": 0.017593613342821692,
+      "learning_rate": 1.220386028629479e-06,
+      "loss": 0.5459,
+      "step": 1330
+    },
+    {
+      "epoch": 0.4882508914693243,
+      "grad_norm": 0.020687281450751747,
+      "learning_rate": 1.2141532255120793e-06,
+      "loss": 0.5363,
+      "step": 1335
+    },
+    {
+      "epoch": 0.490079546493554,
+      "grad_norm": 0.01587264987420519,
+      "learning_rate": 1.207911690817759e-06,
+      "loss": 0.5147,
+      "step": 1340
+    },
+    {
+      "epoch": 0.4919082015177837,
+      "grad_norm": 0.01647809155605945,
+      "learning_rate": 1.2016616790299162e-06,
+      "loss": 0.5368,
+      "step": 1345
+    },
+    {
+      "epoch": 0.4937368565420133,
+      "grad_norm": 0.017752895015077846,
+      "learning_rate": 1.19540344497758e-06,
+      "loss": 0.543,
+      "step": 1350
+    },
+    {
+      "epoch": 0.495565511566243,
+      "grad_norm": 0.01616175749543083,
+      "learning_rate": 1.1891372438250229e-06,
+      "loss": 0.5001,
+      "step": 1355
+    },
+    {
+      "epoch": 0.4973941665904727,
+      "grad_norm": 0.017219113142474245,
+      "learning_rate": 1.1828633310613567e-06,
+      "loss": 0.5159,
+      "step": 1360
+    },
+    {
+      "epoch": 0.4992228216147024,
+      "grad_norm": 0.016504267679499782,
+      "learning_rate": 1.1765819624901154e-06,
+      "loss": 0.4785,
+      "step": 1365
+    },
+    {
+      "epoch": 0.501051476638932,
+      "grad_norm": 0.01578671748055682,
+      "learning_rate": 1.170293394218825e-06,
+      "loss": 0.5261,
+      "step": 1370
+    },
+    {
+      "epoch": 0.5028801316631617,
+      "grad_norm": 0.017103542194380834,
+      "learning_rate": 1.163997882648563e-06,
+      "loss": 0.4988,
+      "step": 1375
+    },
+    {
+      "epoch": 0.5047087866873914,
+      "grad_norm": 0.0169170553425639,
+      "learning_rate": 1.1576956844635008e-06,
+      "loss": 0.5264,
+      "step": 1380
+    },
+    {
+      "epoch": 0.5065374417116211,
+      "grad_norm": 0.017221630452876273,
+      "learning_rate": 1.1513870566204427e-06,
+      "loss": 0.5423,
+      "step": 1385
+    },
+    {
+      "epoch": 0.5083660967358508,
+      "grad_norm": 0.018820113404335967,
+      "learning_rate": 1.1450722563383448e-06,
+      "loss": 0.5251,
+      "step": 1390
+    },
+    {
+      "epoch": 0.5101947517600804,
+      "grad_norm": 0.016326225259013977,
+      "learning_rate": 1.1387515410878306e-06,
+      "loss": 0.5235,
+      "step": 1395
+    },
+    {
+      "epoch": 0.5120234067843101,
+      "grad_norm": 0.017192003635164373,
+      "learning_rate": 1.1324251685806919e-06,
+      "loss": 0.5119,
+      "step": 1400
+    },
+    {
+      "epoch": 0.5138520618085398,
+      "grad_norm": 0.01614617443404074,
+      "learning_rate": 1.12609339675938e-06,
+      "loss": 0.5009,
+      "step": 1405
+    },
+    {
+      "epoch": 0.5156807168327695,
+      "grad_norm": 0.017743406738181977,
+      "learning_rate": 1.1197564837864922e-06,
+      "loss": 0.5117,
+      "step": 1410
+    },
+    {
+      "epoch": 0.5175093718569992,
+      "grad_norm": 0.01645885913887404,
+      "learning_rate": 1.113414688034242e-06,
+      "loss": 0.5273,
+      "step": 1415
+    },
+    {
+      "epoch": 0.5193380268812289,
+      "grad_norm": 0.01834517528236257,
+      "learning_rate": 1.1070682680739274e-06,
+      "loss": 0.5377,
+      "step": 1420
+    },
+    {
+      "epoch": 0.5211666819054586,
+      "grad_norm": 0.01641395755209379,
+      "learning_rate": 1.1007174826653872e-06,
+      "loss": 0.5257,
+      "step": 1425
+    },
+    {
+      "epoch": 0.5229953369296882,
+      "grad_norm": 0.01599953868820275,
+      "learning_rate": 1.0943625907464498e-06,
+      "loss": 0.514,
+      "step": 1430
+    },
+    {
+      "epoch": 0.5248239919539179,
+      "grad_norm": 0.020160517814362934,
+      "learning_rate": 1.0880038514223777e-06,
+      "loss": 0.5348,
+      "step": 1435
+    },
+    {
+      "epoch": 0.5266526469781476,
+      "grad_norm": 0.016249139274883778,
+      "learning_rate": 1.0816415239553014e-06,
+      "loss": 0.5226,
+      "step": 1440
+    },
+    {
+      "epoch": 0.5284813020023773,
+      "grad_norm": 0.017492960358966442,
+      "learning_rate": 1.0752758677536494e-06,
+      "loss": 0.5129,
+      "step": 1445
+    },
+    {
+      "epoch": 0.5303099570266069,
+      "grad_norm": 0.017023094417529788,
+      "learning_rate": 1.0689071423615707e-06,
+      "loss": 0.5141,
+      "step": 1450
+    },
+    {
+      "epoch": 0.5321386120508366,
+      "grad_norm": 0.016947613536071847,
+      "learning_rate": 1.0625356074483543e-06,
+      "loss": 0.4927,
+      "step": 1455
+    },
+    {
+      "epoch": 0.5339672670750663,
+      "grad_norm": 0.016358840825066192,
+      "learning_rate": 1.056161522797839e-06,
+      "loss": 0.5094,
+      "step": 1460
+    },
+    {
+      "epoch": 0.5357959220992959,
+      "grad_norm": 0.01738278153442922,
+      "learning_rate": 1.0497851482978246e-06,
+      "loss": 0.5574,
+      "step": 1465
+    },
+    {
+      "epoch": 0.5376245771235256,
+      "grad_norm": 0.01698692100526017,
+      "learning_rate": 1.0434067439294724e-06,
+      "loss": 0.5471,
+      "step": 1470
+    },
+    {
+      "epoch": 0.5394532321477553,
+      "grad_norm": 0.016905842026346783,
+      "learning_rate": 1.0370265697567078e-06,
+      "loss": 0.5265,
+      "step": 1475
+    },
+    {
+      "epoch": 0.541281887171985,
+      "grad_norm": 0.016017645116936763,
+      "learning_rate": 1.0306448859156153e-06,
+      "loss": 0.4743,
+      "step": 1480
+    },
+    {
+      "epoch": 0.5431105421962147,
+      "grad_norm": 0.01784235362038555,
+      "learning_rate": 1.0242619526038321e-06,
+      "loss": 0.5356,
+      "step": 1485
+    },
+    {
+      "epoch": 0.5449391972204444,
+      "grad_norm": 0.017061295623251198,
+      "learning_rate": 1.0178780300699395e-06,
+      "loss": 0.5166,
+      "step": 1490
+    },
+    {
+      "epoch": 0.5467678522446741,
+      "grad_norm": 0.0172213671791935,
+      "learning_rate": 1.0114933786028531e-06,
+      "loss": 0.522,
+      "step": 1495
+    },
+    {
+      "epoch": 0.5485965072689037,
+      "grad_norm": 0.018368691314043055,
+      "learning_rate": 1.0051082585212073e-06,
+      "loss": 0.5021,
+      "step": 1500
+    },
+    {
+      "epoch": 0.5485965072689037,
+      "eval_runtime": 90.9452,
+      "eval_samples_per_second": 10.996,
+      "eval_steps_per_second": 0.352,
+      "step": 1500
+    },
+    {
+      "epoch": 0.5504251622931334,
+      "grad_norm": 0.018847142564909176,
+      "learning_rate": 9.987229301627447e-07,
+      "loss": 0.54,
+      "step": 1505
+    },
+    {
+      "epoch": 0.5522538173173631,
+      "grad_norm": 0.01616244298204991,
+      "learning_rate": 9.923376538736983e-07,
+      "loss": 0.484,
+      "step": 1510
+    },
+    {
+      "epoch": 0.5540824723415928,
+      "grad_norm": 0.018586584011235783,
+      "learning_rate": 9.859526899981794e-07,
+      "loss": 0.5209,
+      "step": 1515
+    },
+    {
+      "epoch": 0.5559111273658225,
+      "grad_norm": 0.017323074369506024,
+      "learning_rate": 9.79568298867561e-07,
+      "loss": 0.5112,
+      "step": 1520
+    },
+    {
+      "epoch": 0.5577397823900521,
+      "grad_norm": 0.016864181491095635,
+      "learning_rate": 9.731847407898635e-07,
+      "loss": 0.5405,
+      "step": 1525
+    },
+    {
+      "epoch": 0.5595684374142817,
+      "grad_norm": 0.01757979348075452,
+      "learning_rate": 9.668022760391416e-07,
+      "loss": 0.5097,
+      "step": 1530
+    },
+    {
+      "epoch": 0.5613970924385114,
+      "grad_norm": 0.017290946509385344,
+      "learning_rate": 9.604211648448735e-07,
+      "loss": 0.506,
+      "step": 1535
+    },
+    {
+      "epoch": 0.5632257474627411,
+      "grad_norm": 0.019488130700842925,
+      "learning_rate": 9.540416673813469e-07,
+      "loss": 0.5195,
+      "step": 1540
+    },
+    {
+      "epoch": 0.5650544024869708,
+      "grad_norm": 0.017668343125548804,
+      "learning_rate": 9.476640437570561e-07,
+      "loss": 0.5097,
+      "step": 1545
+    },
+    {
+      "epoch": 0.5668830575112005,
+      "grad_norm": 0.016430740880184223,
+      "learning_rate": 9.41288554004093e-07,
+      "loss": 0.4637,
+      "step": 1550
+    },
+    {
+      "epoch": 0.5687117125354302,
+      "grad_norm": 0.016316188966994903,
+      "learning_rate": 9.349154580675457e-07,
+      "loss": 0.5234,
+      "step": 1555
+    },
+    {
+      "epoch": 0.5705403675596599,
+      "grad_norm": 0.016243301360965037,
+      "learning_rate": 9.285450157949014e-07,
+      "loss": 0.5042,
+      "step": 1560
+    },
+    {
+      "epoch": 0.5723690225838896,
+      "grad_norm": 0.017185404503474885,
+      "learning_rate": 9.221774869254495e-07,
+      "loss": 0.5355,
+      "step": 1565
+    },
+    {
+      "epoch": 0.5741976776081192,
+      "grad_norm": 0.016631229219458945,
+      "learning_rate": 9.15813131079693e-07,
+      "loss": 0.5115,
+      "step": 1570
+    },
+    {
+      "epoch": 0.5760263326323489,
+      "grad_norm": 0.016663585035506575,
+      "learning_rate": 9.094522077487629e-07,
+      "loss": 0.5157,
+      "step": 1575
+    },
+    {
+      "epoch": 0.5778549876565786,
+      "grad_norm": 0.015797414555621704,
+      "learning_rate": 9.03094976283837e-07,
+      "loss": 0.5165,
+      "step": 1580
+    },
+    {
+      "epoch": 0.5796836426808083,
+      "grad_norm": 0.01686097335264785,
+      "learning_rate": 8.967416958855665e-07,
+      "loss": 0.4786,
+      "step": 1585
+    },
+    {
+      "epoch": 0.581512297705038,
+      "grad_norm": 0.016822995245937927,
+      "learning_rate": 8.903926255935085e-07,
+      "loss": 0.5357,
+      "step": 1590
+    },
+    {
+      "epoch": 0.5833409527292677,
+      "grad_norm": 0.017322359451677507,
+      "learning_rate": 8.840480242755611e-07,
+      "loss": 0.5308,
+      "step": 1595
+    },
+    {
+      "epoch": 0.5851696077534974,
+      "grad_norm": 0.01587073701140051,
+      "learning_rate": 8.777081506174126e-07,
+      "loss": 0.4838,
+      "step": 1600
+    },
+    {
+      "epoch": 0.5869982627777269,
+      "grad_norm": 0.01682453754788706,
+      "learning_rate": 8.713732631119923e-07,
+      "loss": 0.5324,
+      "step": 1605
+    },
+    {
+      "epoch": 0.5888269178019566,
+      "grad_norm": 0.016998489801133585,
+      "learning_rate": 8.650436200489303e-07,
+      "loss": 0.5018,
+      "step": 1610
+    },
+    {
+      "epoch": 0.5906555728261863,
+      "grad_norm": 0.01636943726286715,
+      "learning_rate": 8.587194795040284e-07,
+      "loss": 0.5245,
+      "step": 1615
+    },
+    {
+      "epoch": 0.592484227850416,
+      "grad_norm": 0.01676770407505387,
+      "learning_rate": 8.524010993287363e-07,
+      "loss": 0.5162,
+      "step": 1620
+    },
+    {
+      "epoch": 0.5943128828746457,
+      "grad_norm": 0.019764893321339468,
+      "learning_rate": 8.460887371396376e-07,
+      "loss": 0.5367,
+      "step": 1625
+    },
+    {
+      "epoch": 0.5961415378988754,
+      "grad_norm": 0.016706342961335864,
+      "learning_rate": 8.397826503079488e-07,
+      "loss": 0.5043,
+      "step": 1630
+    },
+    {
+      "epoch": 0.597970192923105,
+      "grad_norm": 0.01670204571072609,
+      "learning_rate": 8.334830959490234e-07,
+      "loss": 0.5219,
+      "step": 1635
+    },
+    {
+      "epoch": 0.5997988479473347,
+      "grad_norm": 0.017044667355634275,
+      "learning_rate": 8.27190330911868e-07,
+      "loss": 0.5227,
+      "step": 1640
+    },
+    {
+      "epoch": 0.6016275029715644,
+      "grad_norm": 0.01806993489587699,
+      "learning_rate": 8.209046117686729e-07,
+      "loss": 0.5291,
+      "step": 1645
+    },
+    {
+      "epoch": 0.6034561579957941,
+      "grad_norm": 0.016322065945858236,
+      "learning_rate": 8.146261948043484e-07,
+      "loss": 0.5323,
+      "step": 1650
+    },
+    {
+      "epoch": 0.6052848130200238,
+      "grad_norm": 0.021447076161278222,
+      "learning_rate": 8.083553360060758e-07,
+      "loss": 0.5525,
+      "step": 1655
+    },
+    {
+      "epoch": 0.6071134680442535,
+      "grad_norm": 0.01651777259639759,
+      "learning_rate": 8.020922910528716e-07,
+      "loss": 0.4984,
+      "step": 1660
+    },
+    {
+      "epoch": 0.6089421230684832,
+      "grad_norm": 0.016193119419621732,
+      "learning_rate": 7.958373153051604e-07,
+      "loss": 0.5087,
+      "step": 1665
+    },
+    {
+      "epoch": 0.6107707780927129,
+      "grad_norm": 0.016530884028175274,
+      "learning_rate": 7.89590663794366e-07,
+      "loss": 0.5405,
+      "step": 1670
+    },
+    {
+      "epoch": 0.6125994331169425,
+      "grad_norm": 0.01592920822503681,
+      "learning_rate": 7.83352591212511e-07,
+      "loss": 0.4931,
+      "step": 1675
+    },
+    {
+      "epoch": 0.6144280881411721,
+      "grad_norm": 0.017436269469652713,
+      "learning_rate": 7.771233519018325e-07,
+      "loss": 0.5043,
+      "step": 1680
+    },
+    {
+      "epoch": 0.6162567431654018,
+      "grad_norm": 0.01662888442859037,
+      "learning_rate": 7.709031998444137e-07,
+      "loss": 0.524,
+      "step": 1685
+    },
+    {
+      "epoch": 0.6180853981896315,
+      "grad_norm": 0.02058323498639703,
+      "learning_rate": 7.646923886518262e-07,
+      "loss": 0.5244,
+      "step": 1690
+    },
+    {
+      "epoch": 0.6199140532138612,
+      "grad_norm": 0.018033781045547294,
+      "learning_rate": 7.584911715547906e-07,
+      "loss": 0.5117,
+      "step": 1695
+    },
+    {
+      "epoch": 0.6217427082380909,
+      "grad_norm": 0.01582858629881835,
+      "learning_rate": 7.522998013928522e-07,
+      "loss": 0.5153,
+      "step": 1700
+    },
+    {
+      "epoch": 0.6235713632623205,
+      "grad_norm": 0.015515688366334969,
+      "learning_rate": 7.461185306040711e-07,
+      "loss": 0.4966,
+      "step": 1705
+    },
+    {
+      "epoch": 0.6254000182865502,
+      "grad_norm": 0.016618705133961523,
+      "learning_rate": 7.399476112147293e-07,
+      "loss": 0.4984,
+      "step": 1710
+    },
+    {
+      "epoch": 0.6272286733107799,
+      "grad_norm": 0.018723544520206126,
+      "learning_rate": 7.337872948290574e-07,
+      "loss": 0.4958,
+      "step": 1715
+    },
+    {
+      "epoch": 0.6290573283350096,
+      "grad_norm": 0.016441664454426172,
+      "learning_rate": 7.276378326189729e-07,
+      "loss": 0.4922,
+      "step": 1720
+    },
+    {
+      "epoch": 0.6308859833592393,
+      "grad_norm": 0.01729878929676495,
+      "learning_rate": 7.214994753138406e-07,
+      "loss": 0.5218,
+      "step": 1725
+    },
+    {
+      "epoch": 0.632714638383469,
+      "grad_norm": 0.01686530490874631,
+      "learning_rate": 7.153724731902505e-07,
+      "loss": 0.5323,
+      "step": 1730
+    },
+    {
+      "epoch": 0.6345432934076987,
+      "grad_norm": 0.016353842942108295,
+      "learning_rate": 7.09257076061813e-07,
+      "loss": 0.5012,
+      "step": 1735
+    },
+    {
+      "epoch": 0.6363719484319283,
+      "grad_norm": 0.016361879616358148,
+      "learning_rate": 7.031535332689721e-07,
+      "loss": 0.495,
+      "step": 1740
+    },
+    {
+      "epoch": 0.638200603456158,
+      "grad_norm": 0.016527292102103516,
+      "learning_rate": 6.970620936688403e-07,
+      "loss": 0.4972,
+      "step": 1745
+    },
+    {
+      "epoch": 0.6400292584803877,
+      "grad_norm": 0.017510098659919177,
+      "learning_rate": 6.909830056250526e-07,
+      "loss": 0.5204,
+      "step": 1750
+    },
+    {
+      "epoch": 0.6418579135046174,
+      "grad_norm": 0.01609976942743224,
+      "learning_rate": 6.849165169976386e-07,
+      "loss": 0.4774,
+      "step": 1755
+    },
+    {
+      "epoch": 0.643686568528847,
+      "grad_norm": 0.016336561106628334,
+      "learning_rate": 6.788628751329169e-07,
+      "loss": 0.5283,
+      "step": 1760
+    },
+    {
+      "epoch": 0.6455152235530767,
+      "grad_norm": 0.016618006245396773,
+      "learning_rate": 6.728223268534125e-07,
+      "loss": 0.4911,
+      "step": 1765
+    },
+    {
+      "epoch": 0.6473438785773064,
+      "grad_norm": 0.01596530642325609,
+      "learning_rate": 6.667951184477891e-07,
+      "loss": 0.5324,
+      "step": 1770
+    },
+    {
+      "epoch": 0.649172533601536,
+      "grad_norm": 0.016323553088939034,
+      "learning_rate": 6.607814956608109e-07,
+      "loss": 0.5191,
+      "step": 1775
+    },
+    {
+      "epoch": 0.6510011886257657,
+      "grad_norm": 0.016377824778227988,
+      "learning_rate": 6.547817036833207e-07,
+      "loss": 0.5055,
+      "step": 1780
+    },
+    {
+      "epoch": 0.6528298436499954,
+      "grad_norm": 0.01657761502862952,
+      "learning_rate": 6.487959871422441e-07,
+      "loss": 0.5344,
+      "step": 1785
+    },
+    {
+      "epoch": 0.6546584986742251,
+      "grad_norm": 0.02069195693629209,
+      "learning_rate": 6.428245900906155e-07,
+      "loss": 0.5115,
+      "step": 1790
+    },
+    {
+      "epoch": 0.6564871536984548,
+      "grad_norm": 0.017463377375077874,
+      "learning_rate": 6.368677559976259e-07,
+      "loss": 0.5056,
+      "step": 1795
+    },
+    {
+      "epoch": 0.6583158087226845,
+      "grad_norm": 0.018004510839168156,
+      "learning_rate": 6.309257277386974e-07,
+      "loss": 0.5288,
+      "step": 1800
+    },
+    {
+      "epoch": 0.6601444637469142,
+      "grad_norm": 0.016003113936045966,
+      "learning_rate": 6.249987475855813e-07,
+      "loss": 0.495,
+      "step": 1805
+    },
+    {
+      "epoch": 0.6619731187711438,
+      "grad_norm": 0.05467367442201051,
+      "learning_rate": 6.190870571964772e-07,
+      "loss": 0.5364,
+      "step": 1810
+    },
+    {
+      "epoch": 0.6638017737953735,
+      "grad_norm": 0.016376305642967784,
+      "learning_rate": 6.131908976061831e-07,
+      "loss": 0.5213,
+      "step": 1815
+    },
+    {
+      "epoch": 0.6656304288196032,
+      "grad_norm": 0.01661057467766724,
+      "learning_rate": 6.07310509216266e-07,
+      "loss": 0.4972,
+      "step": 1820
+    },
+    {
+      "epoch": 0.6674590838438329,
+      "grad_norm": 0.018337478686271537,
+      "learning_rate": 6.014461317852606e-07,
+      "loss": 0.501,
+      "step": 1825
+    },
+    {
+      "epoch": 0.6692877388680626,
+      "grad_norm": 0.01658447584465899,
+      "learning_rate": 5.955980044188934e-07,
+      "loss": 0.501,
+      "step": 1830
+    },
+    {
+      "epoch": 0.6711163938922923,
+      "grad_norm": 0.01665972763039353,
+      "learning_rate": 5.897663655603351e-07,
+      "loss": 0.4827,
+      "step": 1835
+    },
+    {
+      "epoch": 0.6729450489165218,
+      "grad_norm": 0.015905452330705276,
+      "learning_rate": 5.839514529804757e-07,
+      "loss": 0.49,
+      "step": 1840
+    },
+    {
+      "epoch": 0.6747737039407515,
+      "grad_norm": 0.017648443176781575,
+      "learning_rate": 5.781535037682334e-07,
+      "loss": 0.5377,
+      "step": 1845
+    },
+    {
+      "epoch": 0.6766023589649812,
+      "grad_norm": 0.020515443403825512,
+      "learning_rate": 5.723727543208866e-07,
+      "loss": 0.5121,
+      "step": 1850
+    },
+    {
+      "epoch": 0.6784310139892109,
+      "grad_norm": 0.01676845849475933,
+      "learning_rate": 5.666094403344334e-07,
+      "loss": 0.5073,
+      "step": 1855
+    },
+    {
+      "epoch": 0.6802596690134406,
+      "grad_norm": 0.018011104072379085,
+      "learning_rate": 5.608637967939846e-07,
+      "loss": 0.4708,
+      "step": 1860
+    },
+    {
+      "epoch": 0.6820883240376703,
+      "grad_norm": 0.015900263885464234,
+      "learning_rate": 5.551360579641826e-07,
+      "loss": 0.4986,
+      "step": 1865
+    },
+    {
+      "epoch": 0.6839169790619,
+      "grad_norm": 0.016849149326338082,
+      "learning_rate": 5.49426457379646e-07,
+      "loss": 0.5614,
+      "step": 1870
+    },
+    {
+      "epoch": 0.6857456340861297,
+      "grad_norm": 0.016896910336145418,
+      "learning_rate": 5.437352278354527e-07,
+      "loss": 0.5141,
+      "step": 1875
+    },
+    {
+      "epoch": 0.6875742891103593,
+      "grad_norm": 0.017256221022889148,
+      "learning_rate": 5.380626013776455e-07,
+      "loss": 0.5076,
+      "step": 1880
+    },
+    {
+      "epoch": 0.689402944134589,
+      "grad_norm": 0.0177427465442081,
+      "learning_rate": 5.324088092937715e-07,
+      "loss": 0.4884,
+      "step": 1885
+    },
+    {
+      "epoch": 0.6912315991588187,
+      "grad_norm": 0.01787008311958425,
+      "learning_rate": 5.267740821034524e-07,
+      "loss": 0.5044,
+      "step": 1890
+    },
+    {
+      "epoch": 0.6930602541830484,
+      "grad_norm": 0.016522804042283963,
+      "learning_rate": 5.211586495489845e-07,
+      "loss": 0.51,
+      "step": 1895
+    },
+    {
+      "epoch": 0.6948889092072781,
+      "grad_norm": 0.01633024915767824,
+      "learning_rate": 5.155627405859731e-07,
+      "loss": 0.497,
+      "step": 1900
+    },
+    {
+      "epoch": 0.6967175642315078,
+      "grad_norm": 0.020263855470227177,
+      "learning_rate": 5.09986583373996e-07,
+      "loss": 0.5012,
+      "step": 1905
+    },
+    {
+      "epoch": 0.6985462192557375,
+      "grad_norm": 0.01635576581749387,
+      "learning_rate": 5.044304052673014e-07,
+      "loss": 0.5192,
+      "step": 1910
+    },
+    {
+      "epoch": 0.700374874279967,
+      "grad_norm": 0.016158522614268627,
+      "learning_rate": 4.988944328055385e-07,
+      "loss": 0.4976,
+      "step": 1915
+    },
+    {
+      "epoch": 0.7022035293041967,
+      "grad_norm": 0.016149116884376268,
+      "learning_rate": 4.9337889170452e-07,
+      "loss": 0.5238,
+      "step": 1920
+    },
+    {
+      "epoch": 0.7040321843284264,
+      "grad_norm": 0.015201891264101122,
+      "learning_rate": 4.878840068470198e-07,
+      "loss": 0.4856,
+      "step": 1925
+    },
+    {
+      "epoch": 0.7058608393526561,
+      "grad_norm": 0.017161206306302912,
+      "learning_rate": 4.824100022736035e-07,
+      "loss": 0.4956,
+      "step": 1930
+    },
+    {
+      "epoch": 0.7076894943768858,
+      "grad_norm": 0.01600674616433465,
+      "learning_rate": 4.769571011734942e-07,
+      "loss": 0.5348,
+      "step": 1935
+    },
+    {
+      "epoch": 0.7095181494011155,
+      "grad_norm": 0.02483027155254057,
+      "learning_rate": 4.7152552587547203e-07,
+      "loss": 0.4926,
+      "step": 1940
+    },
+    {
+      "epoch": 0.7113468044253451,
+      "grad_norm": 0.017154306148343564,
+      "learning_rate": 4.6611549783880933e-07,
+      "loss": 0.546,
+      "step": 1945
+    },
+    {
+      "epoch": 0.7131754594495748,
+      "grad_norm": 0.016087093474122675,
+      "learning_rate": 4.6072723764424146e-07,
+      "loss": 0.504,
+      "step": 1950
+    },
+    {
+      "epoch": 0.7150041144738045,
+      "grad_norm": 0.017311434435015243,
+      "learning_rate": 4.5536096498497287e-07,
+      "loss": 0.5316,
+      "step": 1955
+    },
+    {
+      "epoch": 0.7168327694980342,
+      "grad_norm": 0.015625323180982285,
+      "learning_rate": 4.5001689865771963e-07,
+      "loss": 0.5086,
+      "step": 1960
+    },
+    {
+      "epoch": 0.7186614245222639,
+      "grad_norm": 0.016296000392688356,
+      "learning_rate": 4.4469525655378914e-07,
+      "loss": 0.4872,
+      "step": 1965
+    },
+    {
+      "epoch": 0.7204900795464936,
+      "grad_norm": 0.017442945266365154,
+      "learning_rate": 4.3939625565019413e-07,
+      "loss": 0.485,
+      "step": 1970
+    },
+    {
+      "epoch": 0.7223187345707233,
+      "grad_norm": 0.015878082725655274,
+      "learning_rate": 4.3412011200081e-07,
+      "loss": 0.478,
+      "step": 1975
+    },
+    {
+      "epoch": 0.724147389594953,
+      "grad_norm": 0.016745871083834154,
+      "learning_rate": 4.2886704072756065e-07,
+      "loss": 0.4977,
+      "step": 1980
+    },
+    {
+      "epoch": 0.7259760446191826,
+      "grad_norm": 0.016268868688411667,
+      "learning_rate": 4.2363725601165115e-07,
+      "loss": 0.5214,
+      "step": 1985
+    },
+    {
+      "epoch": 0.7278046996434123,
+      "grad_norm": 0.015781104016875658,
+      "learning_rate": 4.184309710848346e-07,
+      "loss": 0.5154,
+      "step": 1990
+    },
+    {
+      "epoch": 0.7296333546676419,
+      "grad_norm": 0.016794072327098834,
+      "learning_rate": 4.1324839822071554e-07,
+      "loss": 0.5052,
+      "step": 1995
+    },
+    {
+      "epoch": 0.7314620096918716,
+      "grad_norm": 0.016021194138622193,
+      "learning_rate": 4.080897487260977e-07,
+      "loss": 0.5181,
+      "step": 2000
+    },
+    {
+      "epoch": 0.7314620096918716,
+      "eval_runtime": 90.9981,
+      "eval_samples_per_second": 10.989,
+      "eval_steps_per_second": 0.352,
+      "step": 2000
+    },
+    {
+      "epoch": 0.7332906647161013,
+      "grad_norm": 0.018652885361705725,
+      "learning_rate": 4.029552329323689e-07,
+      "loss": 0.5148,
+      "step": 2005
+    },
+    {
+      "epoch": 0.735119319740331,
+      "grad_norm": 0.020621833393599447,
+      "learning_rate": 3.9784506018692165e-07,
+      "loss": 0.4967,
+      "step": 2010
+    },
+    {
+      "epoch": 0.7369479747645606,
+      "grad_norm": 0.017696157848433036,
+      "learning_rate": 3.927594388446209e-07,
+      "loss": 0.5045,
+      "step": 2015
+    },
+    {
+      "epoch": 0.7387766297887903,
+      "grad_norm": 0.017565692592477998,
+      "learning_rate": 3.87698576259309e-07,
+      "loss": 0.5077,
+      "step": 2020
+    },
+    {
+      "epoch": 0.74060528481302,
+      "grad_norm": 0.015762648507589124,
+      "learning_rate": 3.826626787753483e-07,
+      "loss": 0.5156,
+      "step": 2025
+    },
+    {
+      "epoch": 0.7424339398372497,
+      "grad_norm": 0.017481055733789334,
+      "learning_rate": 3.7765195171921094e-07,
+      "loss": 0.5598,
+      "step": 2030
+    },
+    {
+      "epoch": 0.7442625948614794,
+      "grad_norm": 0.02074072370962602,
+      "learning_rate": 3.726665993911058e-07,
+      "loss": 0.5268,
+      "step": 2035
+    },
+    {
+      "epoch": 0.7460912498857091,
+      "grad_norm": 0.017369175710519432,
+      "learning_rate": 3.6770682505664895e-07,
+      "loss": 0.5117,
+      "step": 2040
+    },
+    {
+      "epoch": 0.7479199049099388,
+      "grad_norm": 0.01928218423086828,
+      "learning_rate": 3.6277283093857615e-07,
+      "loss": 0.4921,
+      "step": 2045
+    },
+    {
+      "epoch": 0.7497485599341684,
+      "grad_norm": 0.018558674114664685,
+      "learning_rate": 3.5786481820849743e-07,
+      "loss": 0.4993,
+      "step": 2050
+    },
+    {
+      "epoch": 0.7515772149583981,
+      "grad_norm": 0.021517699992954543,
+      "learning_rate": 3.5298298697869475e-07,
+      "loss": 0.5148,
+      "step": 2055
+    },
+    {
+      "epoch": 0.7534058699826278,
+      "grad_norm": 0.017304528635241158,
+      "learning_rate": 3.481275362939634e-07,
+      "loss": 0.5014,
+      "step": 2060
+    },
+    {
+      "epoch": 0.7552345250068575,
+      "grad_norm": 0.017708137711997186,
+      "learning_rate": 3.4329866412349574e-07,
+      "loss": 0.4996,
+      "step": 2065
+    },
+    {
+      "epoch": 0.7570631800310872,
+      "grad_norm": 0.017363757612465638,
+      "learning_rate": 3.384965673528102e-07,
+      "loss": 0.4901,
+      "step": 2070
+    },
+    {
+      "epoch": 0.7588918350553168,
+      "grad_norm": 0.016291702170797662,
+      "learning_rate": 3.3372144177572303e-07,
+      "loss": 0.4858,
+      "step": 2075
+    },
+    {
+      "epoch": 0.7607204900795465,
+      "grad_norm": 0.016749355612971098,
+      "learning_rate": 3.289734820863661e-07,
+      "loss": 0.5311,
+      "step": 2080
+    },
+    {
+      "epoch": 0.7625491451037761,
+      "grad_norm": 0.017456437290198287,
+      "learning_rate": 3.242528818712479e-07,
+      "loss": 0.5231,
+      "step": 2085
+    },
+    {
+      "epoch": 0.7643778001280058,
+      "grad_norm": 0.01679395900909603,
+      "learning_rate": 3.1955983360136154e-07,
+      "loss": 0.5136,
+      "step": 2090
+    },
+    {
+      "epoch": 0.7662064551522355,
+      "grad_norm": 0.017092110352329166,
+      "learning_rate": 3.1489452862433485e-07,
+      "loss": 0.5455,
+      "step": 2095
+    },
+    {
+      "epoch": 0.7680351101764652,
+      "grad_norm": 0.0155875203902004,
+      "learning_rate": 3.102571571566328e-07,
+      "loss": 0.5261,
+      "step": 2100
+    },
+    {
+      "epoch": 0.7698637652006949,
+      "grad_norm": 0.01568859802236798,
+      "learning_rate": 3.0564790827579813e-07,
+      "loss": 0.5082,
+      "step": 2105
+    },
+    {
+      "epoch": 0.7716924202249246,
+      "grad_norm": 0.016963190335988948,
+      "learning_rate": 3.0106696991274274e-07,
+      "loss": 0.507,
+      "step": 2110
+    },
+    {
+      "epoch": 0.7735210752491543,
+      "grad_norm": 0.017664773921940914,
+      "learning_rate": 2.965145288440879e-07,
+      "loss": 0.5065,
+      "step": 2115
+    },
+    {
+      "epoch": 0.7753497302733839,
+      "grad_norm": 0.016840036004908854,
+      "learning_rate": 2.919907706845458e-07,
+      "loss": 0.4984,
+      "step": 2120
+    },
+    {
+      "epoch": 0.7771783852976136,
+      "grad_norm": 0.015867263595970734,
+      "learning_rate": 2.874958798793523e-07,
+      "loss": 0.4734,
+      "step": 2125
+    },
+    {
+      "epoch": 0.7790070403218433,
+      "grad_norm": 0.01603437452296936,
+      "learning_rate": 2.8303003969674865e-07,
+      "loss": 0.4975,
+      "step": 2130
+    },
+    {
+      "epoch": 0.780835695346073,
+      "grad_norm": 0.01607458200267235,
+      "learning_rate": 2.785934322205068e-07,
+      "loss": 0.4989,
+      "step": 2135
+    },
+    {
+      "epoch": 0.7826643503703027,
+      "grad_norm": 0.01878886008248898,
+      "learning_rate": 2.74186238342505e-07,
+      "loss": 0.4996,
+      "step": 2140
+    },
+    {
+      "epoch": 0.7844930053945324,
+      "grad_norm": 0.017646908227235164,
+      "learning_rate": 2.698086377553559e-07,
+      "loss": 0.4906,
+      "step": 2145
+    },
+    {
+      "epoch": 0.786321660418762,
+      "grad_norm": 0.017266971738315123,
+      "learning_rate": 2.654608089450764e-07,
+      "loss": 0.4911,
+      "step": 2150
+    },
+    {
+      "epoch": 0.7881503154429916,
+      "grad_norm": 0.022308407354678553,
+      "learning_rate": 2.611429291838111e-07,
+      "loss": 0.5103,
+      "step": 2155
+    },
+    {
+      "epoch": 0.7899789704672213,
+      "grad_norm": 0.016594781179714146,
+      "learning_rate": 2.568551745226056e-07,
+      "loss": 0.5239,
+      "step": 2160
+    },
+    {
+      "epoch": 0.791807625491451,
+      "grad_norm": 0.016516160236593653,
+      "learning_rate": 2.5259771978422875e-07,
+      "loss": 0.4686,
+      "step": 2165
+    },
+    {
+      "epoch": 0.7936362805156807,
+      "grad_norm": 0.017275446193756597,
+      "learning_rate": 2.4837073855604184e-07,
+      "loss": 0.5391,
+      "step": 2170
+    },
+    {
+      "epoch": 0.7954649355399104,
+      "grad_norm": 0.01605505090938619,
+      "learning_rate": 2.4417440318292416e-07,
+      "loss": 0.5048,
+      "step": 2175
+    },
+    {
+      "epoch": 0.7972935905641401,
+      "grad_norm": 0.016433057958384803,
+      "learning_rate": 2.400088847602445e-07,
+      "loss": 0.5022,
+      "step": 2180
+    },
+    {
+      "epoch": 0.7991222455883698,
+      "grad_norm": 0.01603168317181438,
+      "learning_rate": 2.3587435312688542e-07,
+      "loss": 0.4895,
+      "step": 2185
+    },
+    {
+      "epoch": 0.8009509006125994,
+      "grad_norm": 0.016097573954801322,
+      "learning_rate": 2.3177097685831848e-07,
+      "loss": 0.5052,
+      "step": 2190
+    },
+    {
+      "epoch": 0.8027795556368291,
+      "grad_norm": 0.016347735703314027,
+      "learning_rate": 2.2769892325973105e-07,
+      "loss": 0.5075,
+      "step": 2195
+    },
+    {
+      "epoch": 0.8046082106610588,
+      "grad_norm": 0.017046114120373577,
+      "learning_rate": 2.2365835835920487e-07,
+      "loss": 0.5166,
+      "step": 2200
+    },
+    {
+      "epoch": 0.8064368656852885,
+      "grad_norm": 0.01741287073919635,
+      "learning_rate": 2.1964944690094655e-07,
+      "loss": 0.5151,
+      "step": 2205
+    },
+    {
+      "epoch": 0.8082655207095182,
+      "grad_norm": 0.016486547927818302,
+      "learning_rate": 2.156723523385705e-07,
+      "loss": 0.4899,
+      "step": 2210
+    },
+    {
+      "epoch": 0.8100941757337479,
+      "grad_norm": 0.016184797537776692,
+      "learning_rate": 2.1172723682843473e-07,
+      "loss": 0.5092,
+      "step": 2215
+    },
+    {
+      "epoch": 0.8119228307579776,
+      "grad_norm": 0.01711620716878606,
+      "learning_rate": 2.0781426122302915e-07,
+      "loss": 0.5437,
+      "step": 2220
+    },
+    {
+      "epoch": 0.8137514857822072,
+      "grad_norm": 0.016654399103508962,
+      "learning_rate": 2.0393358506441694e-07,
+      "loss": 0.5026,
+      "step": 2225
+    },
+    {
+      "epoch": 0.8155801408064368,
+      "grad_norm": 0.016593144465982552,
+      "learning_rate": 2.0008536657773045e-07,
+      "loss": 0.5351,
+      "step": 2230
+    },
+    {
+      "epoch": 0.8174087958306665,
+      "grad_norm": 0.016662361195834745,
+      "learning_rate": 1.9626976266471885e-07,
+      "loss": 0.5042,
+      "step": 2235
+    },
+    {
+      "epoch": 0.8192374508548962,
+      "grad_norm": 0.016895196634510556,
+      "learning_rate": 1.9248692889735163e-07,
+      "loss": 0.5079,
+      "step": 2240
+    },
+    {
+      "epoch": 0.8210661058791259,
+      "grad_norm": 0.01577277436380118,
+      "learning_rate": 1.8873701951147536e-07,
+      "loss": 0.4928,
+      "step": 2245
+    },
+    {
+      "epoch": 0.8228947609033556,
+      "grad_norm": 0.01665132620507756,
+      "learning_rate": 1.8502018740052494e-07,
+      "loss": 0.5245,
+      "step": 2250
+    },
+    {
+      "epoch": 0.8247234159275852,
+      "grad_norm": 0.01627720307587572,
+      "learning_rate": 1.8133658410928986e-07,
+      "loss": 0.5218,
+      "step": 2255
+    },
+    {
+      "epoch": 0.8265520709518149,
+      "grad_norm": 0.015675398692590027,
+      "learning_rate": 1.7768635982773529e-07,
+      "loss": 0.4942,
+      "step": 2260
+    },
+    {
+      "epoch": 0.8283807259760446,
+      "grad_norm": 0.018334396785114376,
+      "learning_rate": 1.740696633848785e-07,
+      "loss": 0.5222,
+      "step": 2265
+    },
+    {
+      "epoch": 0.8302093810002743,
+      "grad_norm": 0.018580145771728913,
+      "learning_rate": 1.7048664224272024e-07,
+      "loss": 0.5171,
+      "step": 2270
+    },
+    {
+      "epoch": 0.832038036024504,
+      "grad_norm": 0.01648108908479592,
+      "learning_rate": 1.669374424902339e-07,
+      "loss": 0.511,
+      "step": 2275
+    },
+    {
+      "epoch": 0.8338666910487337,
+      "grad_norm": 0.01709058845132961,
+      "learning_rate": 1.6342220883740666e-07,
+      "loss": 0.5271,
+      "step": 2280
+    },
+    {
+      "epoch": 0.8356953460729634,
+      "grad_norm": 0.016890861901410495,
+      "learning_rate": 1.5994108460934119e-07,
+      "loss": 0.53,
+      "step": 2285
+    },
+    {
+      "epoch": 0.837524001097193,
+      "grad_norm": 0.017261660088763515,
+      "learning_rate": 1.5649421174041188e-07,
+      "loss": 0.546,
+      "step": 2290
+    },
+    {
+      "epoch": 0.8393526561214227,
+      "grad_norm": 0.019044402874809352,
+      "learning_rate": 1.530817307684761e-07,
+      "loss": 0.5103,
+      "step": 2295
+    },
+    {
+      "epoch": 0.8411813111456524,
+      "grad_norm": 0.016606680354726594,
+      "learning_rate": 1.497037808291458e-07,
+      "loss": 0.4805,
+      "step": 2300
+    },
+    {
+      "epoch": 0.843009966169882,
+      "grad_norm": 0.01693730029318611,
+      "learning_rate": 1.4636049965011487e-07,
+      "loss": 0.5172,
+      "step": 2305
+    },
+    {
+      "epoch": 0.8448386211941117,
+      "grad_norm": 0.017379308863508768,
+      "learning_rate": 1.4305202354554134e-07,
+      "loss": 0.513,
+      "step": 2310
+    },
+    {
+      "epoch": 0.8466672762183414,
+      "grad_norm": 0.0159779921920061,
+      "learning_rate": 1.397784874104917e-07,
+      "loss": 0.5128,
+      "step": 2315
+    },
+    {
+      "epoch": 0.848495931242571,
+      "grad_norm": 0.016643061385526618,
+      "learning_rate": 1.365400247154409e-07,
+      "loss": 0.522,
+      "step": 2320
+    },
+    {
+      "epoch": 0.8503245862668007,
+      "grad_norm": 0.01687648593312963,
+      "learning_rate": 1.3333676750082824e-07,
+      "loss": 0.5058,
+      "step": 2325
+    },
+    {
+      "epoch": 0.8521532412910304,
+      "grad_norm": 0.016032393851703408,
+      "learning_rate": 1.3016884637167568e-07,
+      "loss": 0.5144,
+      "step": 2330
+    },
+    {
+      "epoch": 0.8539818963152601,
+      "grad_norm": 0.01663236614977319,
+      "learning_rate": 1.2703639049226345e-07,
+      "loss": 0.4997,
+      "step": 2335
+    },
+    {
+      "epoch": 0.8558105513394898,
+      "grad_norm": 0.017084159814229206,
+      "learning_rate": 1.2393952758086056e-07,
+      "loss": 0.5289,
+      "step": 2340
+    },
+    {
+      "epoch": 0.8576392063637195,
+      "grad_norm": 0.046795259518188144,
+      "learning_rate": 1.2087838390452064e-07,
+      "loss": 0.5325,
+      "step": 2345
+    },
+    {
+      "epoch": 0.8594678613879492,
+      "grad_norm": 0.018858056639567515,
+      "learning_rate": 1.1785308427393181e-07,
+      "loss": 0.516,
+      "step": 2350
+    },
+    {
+      "epoch": 0.8612965164121789,
+      "grad_norm": 0.016973486165728708,
+      "learning_rate": 1.1486375203832855e-07,
+      "loss": 0.4821,
+      "step": 2355
+    },
+    {
+      "epoch": 0.8631251714364085,
+      "grad_norm": 0.016440061194393316,
+      "learning_rate": 1.1191050908046241e-07,
+      "loss": 0.517,
+      "step": 2360
+    },
+    {
+      "epoch": 0.8649538264606382,
+      "grad_norm": 0.01634626038462328,
+      "learning_rate": 1.089934758116322e-07,
+      "loss": 0.5237,
+      "step": 2365
+    },
+    {
+      "epoch": 0.8667824814848679,
+      "grad_norm": 0.015975121144821074,
+      "learning_rate": 1.0611277116677487e-07,
+      "loss": 0.5168,
+      "step": 2370
+    },
+    {
+      "epoch": 0.8686111365090976,
+      "grad_norm": 0.015818032515093727,
+      "learning_rate": 1.032685125996161e-07,
+      "loss": 0.5197,
+      "step": 2375
+    },
+    {
+      "epoch": 0.8704397915333273,
+      "grad_norm": 0.016139801070050402,
+      "learning_rate": 1.0046081607788126e-07,
+      "loss": 0.5052,
+      "step": 2380
+    },
+    {
+      "epoch": 0.8722684465575569,
+      "grad_norm": 0.017649696893816486,
+      "learning_rate": 9.76897960785673e-08,
+      "loss": 0.5239,
+      "step": 2385
+    },
+    {
+      "epoch": 0.8740971015817866,
+      "grad_norm": 0.016689526904007688,
+      "learning_rate": 9.495556558327544e-08,
+      "loss": 0.5248,
+      "step": 2390
+    },
+    {
+      "epoch": 0.8759257566060162,
+      "grad_norm": 0.017233811816289347,
+      "learning_rate": 9.22582360736035e-08,
+      "loss": 0.5252,
+      "step": 2395
+    },
+    {
+      "epoch": 0.8777544116302459,
+      "grad_norm": 0.016467691562862877,
+      "learning_rate": 8.959791752660262e-08,
+      "loss": 0.4985,
+      "step": 2400
+    },
+    {
+      "epoch": 0.8795830666544756,
+      "grad_norm": 0.01651889684672315,
+      "learning_rate": 8.697471841029113e-08,
+      "loss": 0.524,
+      "step": 2405
+    },
+    {
+      "epoch": 0.8814117216787053,
+      "grad_norm": 0.016156095985749932,
+      "learning_rate": 8.43887456792326e-08,
+      "loss": 0.4901,
+      "step": 2410
+    },
+    {
+      "epoch": 0.883240376702935,
+      "grad_norm": 0.0161374597629895,
+      "learning_rate": 8.18401047701761e-08,
+      "loss": 0.4887,
+      "step": 2415
+    },
+    {
+      "epoch": 0.8850690317271647,
+      "grad_norm": 0.016358879304863485,
+      "learning_rate": 7.932889959775613e-08,
+      "loss": 0.5046,
+      "step": 2420
+    },
+    {
+      "epoch": 0.8868976867513944,
+      "grad_norm": 0.016602693944506797,
+      "learning_rate": 7.685523255025561e-08,
+      "loss": 0.4958,
+      "step": 2425
+    },
+    {
+      "epoch": 0.888726341775624,
+      "grad_norm": 0.017647296627426368,
+      "learning_rate": 7.441920448543237e-08,
+      "loss": 0.5058,
+      "step": 2430
+    },
+    {
+      "epoch": 0.8905549967998537,
+      "grad_norm": 0.01691246446397513,
+      "learning_rate": 7.202091472640615e-08,
+      "loss": 0.5436,
+      "step": 2435
+    },
+    {
+      "epoch": 0.8923836518240834,
+      "grad_norm": 0.01640777983658524,
+      "learning_rate": 6.966046105760825e-08,
+      "loss": 0.5208,
+      "step": 2440
+    },
+    {
+      "epoch": 0.8942123068483131,
+      "grad_norm": 0.018188465902255493,
+      "learning_rate": 6.733793972079682e-08,
+      "loss": 0.5493,
+      "step": 2445
+    },
+    {
+      "epoch": 0.8960409618725428,
+      "grad_norm": 0.01735244725018571,
+      "learning_rate": 6.505344541113045e-08,
+      "loss": 0.538,
+      "step": 2450
+    },
+    {
+      "epoch": 0.8978696168967725,
+      "grad_norm": 0.01907248055322319,
+      "learning_rate": 6.28070712733082e-08,
+      "loss": 0.4968,
+      "step": 2455
+    },
+    {
+      "epoch": 0.8996982719210022,
+      "grad_norm": 0.017525190208813646,
+      "learning_rate": 6.059890889777198e-08,
+      "loss": 0.4947,
+      "step": 2460
+    },
+    {
+      "epoch": 0.9015269269452317,
+      "grad_norm": 0.017477268236289163,
+      "learning_rate": 5.8429048316972595e-08,
+      "loss": 0.5184,
+      "step": 2465
+    },
+    {
+      "epoch": 0.9033555819694614,
+      "grad_norm": 0.01765761665903124,
+      "learning_rate": 5.629757800169732e-08,
+      "loss": 0.5242,
+      "step": 2470
+    },
+    {
+      "epoch": 0.9051842369936911,
+      "grad_norm": 0.02991941815923375,
+      "learning_rate": 5.420458485746438e-08,
+      "loss": 0.4866,
+      "step": 2475
+    },
+    {
+      "epoch": 0.9070128920179208,
+      "grad_norm": 0.01738704014681301,
+      "learning_rate": 5.2150154220979014e-08,
+      "loss": 0.493,
+      "step": 2480
+    },
+    {
+      "epoch": 0.9088415470421505,
+      "grad_norm": 0.0154700641596835,
+      "learning_rate": 5.0134369856653534e-08,
+      "loss": 0.4952,
+      "step": 2485
+    },
+    {
+      "epoch": 0.9106702020663802,
+      "grad_norm": 0.019645793234805923,
+      "learning_rate": 4.815731395319278e-08,
+      "loss": 0.4988,
+      "step": 2490
+    },
+    {
+      "epoch": 0.9124988570906098,
+      "grad_norm": 0.016274549218470255,
+      "learning_rate": 4.62190671202427e-08,
+      "loss": 0.4975,
+      "step": 2495
+    },
+    {
+      "epoch": 0.9143275121148395,
+      "grad_norm": 0.017490591509614194,
+      "learning_rate": 4.43197083851039e-08,
+      "loss": 0.5028,
+      "step": 2500
+    },
+    {
+      "epoch": 0.9143275121148395,
+      "eval_runtime": 90.9126,
+      "eval_samples_per_second": 11.0,
+      "eval_steps_per_second": 0.352,
+      "step": 2500
+    },
+    {
+      "epoch": 0.9161561671390692,
+      "grad_norm": 0.017313625675891377,
+      "learning_rate": 4.245931518950929e-08,
+      "loss": 0.5157,
+      "step": 2505
+    },
+    {
+      "epoch": 0.9179848221632989,
+      "grad_norm": 0.01617359584229841,
+      "learning_rate": 4.063796338646663e-08,
+      "loss": 0.493,
+      "step": 2510
+    },
+    {
+      "epoch": 0.9198134771875286,
+      "grad_norm": 0.01775109563425394,
+      "learning_rate": 3.88557272371659e-08,
+      "loss": 0.5006,
+      "step": 2515
+    },
+    {
+      "epoch": 0.9216421322117583,
+      "grad_norm": 0.016942766298411845,
+      "learning_rate": 3.71126794079516e-08,
+      "loss": 0.4891,
+      "step": 2520
+    },
+    {
+      "epoch": 0.923470787235988,
+      "grad_norm": 0.01669809254371766,
+      "learning_rate": 3.540889096735944e-08,
+      "loss": 0.4932,
+      "step": 2525
+    },
+    {
+      "epoch": 0.9252994422602177,
+      "grad_norm": 0.015935278806104767,
+      "learning_rate": 3.374443138321936e-08,
+      "loss": 0.4938,
+      "step": 2530
+    },
+    {
+      "epoch": 0.9271280972844473,
+      "grad_norm": 0.01588111976104216,
+      "learning_rate": 3.2119368519822554e-08,
+      "loss": 0.491,
+      "step": 2535
+    },
+    {
+      "epoch": 0.9289567523086769,
+      "grad_norm": 0.016464611141449353,
+      "learning_rate": 3.053376863515511e-08,
+      "loss": 0.4936,
+      "step": 2540
+    },
+    {
+      "epoch": 0.9307854073329066,
+      "grad_norm": 0.017147830632735245,
+      "learning_rate": 2.898769637819587e-08,
+      "loss": 0.5115,
+      "step": 2545
+    },
+    {
+      "epoch": 0.9326140623571363,
+      "grad_norm": 0.017573715112945552,
+      "learning_rate": 2.748121478628074e-08,
+      "loss": 0.5219,
+      "step": 2550
+    },
+    {
+      "epoch": 0.934442717381366,
+      "grad_norm": 0.016524355529867023,
+      "learning_rate": 2.6014385282532747e-08,
+      "loss": 0.4987,
+      "step": 2555
+    },
+    {
+      "epoch": 0.9362713724055957,
+      "grad_norm": 0.017124543654619927,
+      "learning_rate": 2.4587267673357394e-08,
+      "loss": 0.5105,
+      "step": 2560
+    },
+    {
+      "epoch": 0.9381000274298253,
+      "grad_norm": 0.016149663796524754,
+      "learning_rate": 2.3199920146004138e-08,
+      "loss": 0.5,
+      "step": 2565
+    },
+    {
+      "epoch": 0.939928682454055,
+      "grad_norm": 0.016291206513456682,
+      "learning_rate": 2.185239926619431e-08,
+      "loss": 0.4927,
+      "step": 2570
+    },
+    {
+      "epoch": 0.9417573374782847,
+      "grad_norm": 0.01680274375559089,
+      "learning_rate": 2.0544759975814396e-08,
+      "loss": 0.4861,
+      "step": 2575
+    },
+    {
+      "epoch": 0.9435859925025144,
+      "grad_norm": 0.016921842070277682,
+      "learning_rate": 1.927705559067616e-08,
+      "loss": 0.5503,
+      "step": 2580
+    },
+    {
+      "epoch": 0.9454146475267441,
+      "grad_norm": 0.016224256682366604,
+      "learning_rate": 1.804933779834239e-08,
+      "loss": 0.49,
+      "step": 2585
+    },
+    {
+      "epoch": 0.9472433025509738,
+      "grad_norm": 0.018285696958157516,
+      "learning_rate": 1.686165665602046e-08,
+      "loss": 0.5188,
+      "step": 2590
+    },
+    {
+      "epoch": 0.9490719575752035,
+      "grad_norm": 0.017055017276686454,
+      "learning_rate": 1.5714060588520093e-08,
+      "loss": 0.4859,
+      "step": 2595
+    },
+    {
+      "epoch": 0.9509006125994331,
+      "grad_norm": 0.01689023600800715,
+      "learning_rate": 1.4606596386279924e-08,
+      "loss": 0.5014,
+      "step": 2600
+    },
+    {
+      "epoch": 0.9527292676236628,
+      "grad_norm": 0.01601967393957623,
+      "learning_rate": 1.353930920345936e-08,
+      "loss": 0.5282,
+      "step": 2605
+    },
+    {
+      "epoch": 0.9545579226478925,
+      "grad_norm": 0.016101505378654526,
+      "learning_rate": 1.2512242556097397e-08,
+      "loss": 0.5189,
+      "step": 2610
+    },
+    {
+      "epoch": 0.9563865776721222,
+      "grad_norm": 0.01691060821905496,
+      "learning_rate": 1.1525438320338477e-08,
+      "loss": 0.4874,
+      "step": 2615
+    },
+    {
+      "epoch": 0.9582152326963518,
+      "grad_norm": 0.01834433493706182,
+      "learning_rate": 1.057893673072563e-08,
+      "loss": 0.5311,
+      "step": 2620
+    },
+    {
+      "epoch": 0.9600438877205815,
+      "grad_norm": 0.016167148545609827,
+      "learning_rate": 9.672776378558789e-09,
+      "loss": 0.4991,
+      "step": 2625
+    },
+    {
+      "epoch": 0.9618725427448112,
+      "grad_norm": 0.016019839174249213,
+      "learning_rate": 8.806994210322604e-09,
+      "loss": 0.5021,
+      "step": 2630
+    },
+    {
+      "epoch": 0.9637011977690408,
+      "grad_norm": 0.01808500836287581,
+      "learning_rate": 7.981625526179315e-09,
+      "loss": 0.5222,
+      "step": 2635
+    },
+    {
+      "epoch": 0.9655298527932705,
+      "grad_norm": 0.017430609481273694,
+      "learning_rate": 7.19670397852945e-09,
+      "loss": 0.5023,
+      "step": 2640
+    },
+    {
+      "epoch": 0.9673585078175002,
+      "grad_norm": 0.01593831619408055,
+      "learning_rate": 6.452261570640161e-09,
+      "loss": 0.5094,
+      "step": 2645
+    },
+    {
+      "epoch": 0.9691871628417299,
+      "grad_norm": 0.01688269537341714,
+      "learning_rate": 5.74832865534014e-09,
+      "loss": 0.5122,
+      "step": 2650
+    },
+    {
+      "epoch": 0.9710158178659596,
+      "grad_norm": 0.01679344374513018,
+      "learning_rate": 5.084933933781732e-09,
+      "loss": 0.4909,
+      "step": 2655
+    },
+    {
+      "epoch": 0.9728444728901893,
+      "grad_norm": 0.016337547526303887,
+      "learning_rate": 4.46210445427142e-09,
+      "loss": 0.5182,
+      "step": 2660
+    },
+    {
+      "epoch": 0.974673127914419,
+      "grad_norm": 0.016071199631250827,
+      "learning_rate": 3.8798656111663774e-09,
+      "loss": 0.4673,
+      "step": 2665
+    },
+    {
+      "epoch": 0.9765017829386486,
+      "grad_norm": 0.016850394608535636,
+      "learning_rate": 3.3382411438392934e-09,
+      "loss": 0.5229,
+      "step": 2670
+    },
+    {
+      "epoch": 0.9783304379628783,
+      "grad_norm": 0.018091858546274816,
+      "learning_rate": 2.8372531357104823e-09,
+      "loss": 0.4865,
+      "step": 2675
+    },
+    {
+      "epoch": 0.980159092987108,
+      "grad_norm": 0.01590421627516969,
+      "learning_rate": 2.376922013347715e-09,
+      "loss": 0.4972,
+      "step": 2680
+    },
+    {
+      "epoch": 0.9819877480113377,
+      "grad_norm": 0.015768265813029696,
+      "learning_rate": 1.9572665456326633e-09,
+      "loss": 0.5025,
+      "step": 2685
+    },
+    {
+      "epoch": 0.9838164030355674,
+      "grad_norm": 0.017044611692703958,
+      "learning_rate": 1.5783038429965089e-09,
+      "loss": 0.5066,
+      "step": 2690
+    },
+    {
+      "epoch": 0.9856450580597971,
+      "grad_norm": 0.018431227029253552,
+      "learning_rate": 1.2400493567217285e-09,
+      "loss": 0.4936,
+      "step": 2695
+    },
+    {
+      "epoch": 0.9874737130840266,
+      "grad_norm": 0.01564301064743967,
+      "learning_rate": 9.425168783123716e-10,
+      "loss": 0.5125,
+      "step": 2700
+    },
+    {
+      "epoch": 0.9893023681082563,
+      "grad_norm": 0.018690003863648913,
+      "learning_rate": 6.857185389315123e-10,
+      "loss": 0.482,
+      "step": 2705
+    },
+    {
+      "epoch": 0.991131023132486,
+      "grad_norm": 0.016649965785577904,
+      "learning_rate": 4.696648089068667e-10,
+      "loss": 0.4715,
+      "step": 2710
+    },
+    {
+      "epoch": 0.9929596781567157,
+      "grad_norm": 0.019135876300316445,
+      "learning_rate": 2.9436449730391165e-10,
+      "loss": 0.5493,
+      "step": 2715
+    },
+    {
+      "epoch": 0.9947883331809454,
+      "grad_norm": 0.017746127346522297,
+      "learning_rate": 1.598247515663953e-10,
+      "loss": 0.527,
+      "step": 2720
+    },
+    {
+      "epoch": 0.9966169882051751,
+      "grad_norm": 0.016878019098499255,
+      "learning_rate": 6.605105722534698e-11,
+      "loss": 0.5334,
+      "step": 2725
+    },
+    {
+      "epoch": 0.9984456432294048,
+      "grad_norm": 0.016123109599214404,
+      "learning_rate": 1.304723767492355e-11,
+      "loss": 0.5114,
+      "step": 2730
+    },
+    {
+      "epoch": 0.9999085672487885,
+      "step": 2734,
+      "total_flos": 3.8701158076063416e+18,
+      "train_loss": 0.554877914079542,
+      "train_runtime": 127427.5888,
+      "train_samples_per_second": 1.373,
+      "train_steps_per_second": 0.021
+    }
+  ],
+  "logging_steps": 5,
+  "max_steps": 2734,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 1,
+  "save_steps": 50,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": true
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 3.8701158076063416e+18,
+  "train_batch_size": 2,
+  "trial_name": null,
+  "trial_params": null
+}