diff --git "a/trainer_state.json" "b/trainer_state.json"
new file mode 100644--- /dev/null
+++ "b/trainer_state.json"
@@ -0,0 +1,9771 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 4.0,
+  "eval_steps": 500,
+  "global_step": 6932,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.0005770340450086555,
+      "grad_norm": 1.3359850523956383,
+      "learning_rate": 2.8818443804034583e-08,
+      "loss": 0.656,
+      "step": 1
+    },
+    {
+      "epoch": 0.0028851702250432777,
+      "grad_norm": 1.415611871814656,
+      "learning_rate": 1.4409221902017293e-07,
+      "loss": 0.6481,
+      "step": 5
+    },
+    {
+      "epoch": 0.005770340450086555,
+      "grad_norm": 1.212781925465119,
+      "learning_rate": 2.8818443804034586e-07,
+      "loss": 0.6386,
+      "step": 10
+    },
+    {
+      "epoch": 0.008655510675129832,
+      "grad_norm": 1.2361637923329842,
+      "learning_rate": 4.322766570605188e-07,
+      "loss": 0.65,
+      "step": 15
+    },
+    {
+      "epoch": 0.01154068090017311,
+      "grad_norm": 1.3799250131478336,
+      "learning_rate": 5.763688760806917e-07,
+      "loss": 0.6423,
+      "step": 20
+    },
+    {
+      "epoch": 0.014425851125216388,
+      "grad_norm": 1.303690207496262,
+      "learning_rate": 7.204610951008646e-07,
+      "loss": 0.6492,
+      "step": 25
+    },
+    {
+      "epoch": 0.017311021350259664,
+      "grad_norm": 1.2010767582718596,
+      "learning_rate": 8.645533141210376e-07,
+      "loss": 0.6254,
+      "step": 30
+    },
+    {
+      "epoch": 0.020196191575302943,
+      "grad_norm": 1.123148811050096,
+      "learning_rate": 1.0086455331412106e-06,
+      "loss": 0.632,
+      "step": 35
+    },
+    {
+      "epoch": 0.02308136180034622,
+      "grad_norm": 1.017563618882216,
+      "learning_rate": 1.1527377521613834e-06,
+      "loss": 0.6176,
+      "step": 40
+    },
+    {
+      "epoch": 0.025966532025389497,
+      "grad_norm": 0.9670044654743236,
+      "learning_rate": 1.2968299711815562e-06,
+      "loss": 0.6237,
+      "step": 45
+    },
+    {
+      "epoch": 0.028851702250432775,
+      "grad_norm": 0.8817997118057925,
+      "learning_rate": 1.4409221902017292e-06,
+      "loss": 0.5945,
+      "step": 50
+    },
+    {
+      "epoch": 0.031736872475476054,
+      "grad_norm": 0.7910601022781902,
+      "learning_rate": 1.5850144092219022e-06,
+      "loss": 0.5666,
+      "step": 55
+    },
+    {
+      "epoch": 0.03462204270051933,
+      "grad_norm": 0.698866883185103,
+      "learning_rate": 1.7291066282420752e-06,
+      "loss": 0.5781,
+      "step": 60
+    },
+    {
+      "epoch": 0.03750721292556261,
+      "grad_norm": 0.7228128874208073,
+      "learning_rate": 1.873198847262248e-06,
+      "loss": 0.552,
+      "step": 65
+    },
+    {
+      "epoch": 0.040392383150605886,
+      "grad_norm": 0.6829934171012417,
+      "learning_rate": 2.0172910662824213e-06,
+      "loss": 0.5538,
+      "step": 70
+    },
+    {
+      "epoch": 0.04327755337564916,
+      "grad_norm": 0.6501090725880212,
+      "learning_rate": 2.1613832853025936e-06,
+      "loss": 0.5428,
+      "step": 75
+    },
+    {
+      "epoch": 0.04616272360069244,
+      "grad_norm": 0.5921206030532731,
+      "learning_rate": 2.305475504322767e-06,
+      "loss": 0.5523,
+      "step": 80
+    },
+    {
+      "epoch": 0.04904789382573572,
+      "grad_norm": 0.5843936221943186,
+      "learning_rate": 2.4495677233429396e-06,
+      "loss": 0.5295,
+      "step": 85
+    },
+    {
+      "epoch": 0.05193306405077899,
+      "grad_norm": 0.4971214384044829,
+      "learning_rate": 2.5936599423631124e-06,
+      "loss": 0.5293,
+      "step": 90
+    },
+    {
+      "epoch": 0.054818234275822275,
+      "grad_norm": 0.502778037587032,
+      "learning_rate": 2.7377521613832852e-06,
+      "loss": 0.5192,
+      "step": 95
+    },
+    {
+      "epoch": 0.05770340450086555,
+      "grad_norm": 0.48028251458263926,
+      "learning_rate": 2.8818443804034585e-06,
+      "loss": 0.526,
+      "step": 100
+    },
+    {
+      "epoch": 0.060588574725908825,
+      "grad_norm": 0.4958418647779635,
+      "learning_rate": 3.0259365994236312e-06,
+      "loss": 0.507,
+      "step": 105
+    },
+    {
+      "epoch": 0.06347374495095211,
+      "grad_norm": 0.4605312245170545,
+      "learning_rate": 3.1700288184438045e-06,
+      "loss": 0.5209,
+      "step": 110
+    },
+    {
+      "epoch": 0.06635891517599539,
+      "grad_norm": 0.45269794748416303,
+      "learning_rate": 3.3141210374639773e-06,
+      "loss": 0.5018,
+      "step": 115
+    },
+    {
+      "epoch": 0.06924408540103866,
+      "grad_norm": 0.4518223453145298,
+      "learning_rate": 3.4582132564841505e-06,
+      "loss": 0.5014,
+      "step": 120
+    },
+    {
+      "epoch": 0.07212925562608194,
+      "grad_norm": 0.4782643771875912,
+      "learning_rate": 3.602305475504323e-06,
+      "loss": 0.5152,
+      "step": 125
+    },
+    {
+      "epoch": 0.07501442585112522,
+      "grad_norm": 0.4420045459990657,
+      "learning_rate": 3.746397694524496e-06,
+      "loss": 0.4823,
+      "step": 130
+    },
+    {
+      "epoch": 0.07789959607616849,
+      "grad_norm": 0.42381033264166024,
+      "learning_rate": 3.890489913544669e-06,
+      "loss": 0.4956,
+      "step": 135
+    },
+    {
+      "epoch": 0.08078476630121177,
+      "grad_norm": 0.4406324685320003,
+      "learning_rate": 4.0345821325648425e-06,
+      "loss": 0.4746,
+      "step": 140
+    },
+    {
+      "epoch": 0.08366993652625505,
+      "grad_norm": 0.43316241824454993,
+      "learning_rate": 4.1786743515850145e-06,
+      "loss": 0.4746,
+      "step": 145
+    },
+    {
+      "epoch": 0.08655510675129832,
+      "grad_norm": 0.4422617049175986,
+      "learning_rate": 4.322766570605187e-06,
+      "loss": 0.4791,
+      "step": 150
+    },
+    {
+      "epoch": 0.0894402769763416,
+      "grad_norm": 0.41863788062040475,
+      "learning_rate": 4.466858789625361e-06,
+      "loss": 0.5032,
+      "step": 155
+    },
+    {
+      "epoch": 0.09232544720138489,
+      "grad_norm": 0.4448228403063216,
+      "learning_rate": 4.610951008645534e-06,
+      "loss": 0.4984,
+      "step": 160
+    },
+    {
+      "epoch": 0.09521061742642815,
+      "grad_norm": 0.4280125696477632,
+      "learning_rate": 4.7550432276657065e-06,
+      "loss": 0.4966,
+      "step": 165
+    },
+    {
+      "epoch": 0.09809578765147144,
+      "grad_norm": 0.3924527289307632,
+      "learning_rate": 4.899135446685879e-06,
+      "loss": 0.4707,
+      "step": 170
+    },
+    {
+      "epoch": 0.10098095787651472,
+      "grad_norm": 0.405378064505572,
+      "learning_rate": 5.043227665706052e-06,
+      "loss": 0.4684,
+      "step": 175
+    },
+    {
+      "epoch": 0.10386612810155799,
+      "grad_norm": 0.3967929757933598,
+      "learning_rate": 5.187319884726225e-06,
+      "loss": 0.4685,
+      "step": 180
+    },
+    {
+      "epoch": 0.10675129832660127,
+      "grad_norm": 0.39908256939712566,
+      "learning_rate": 5.3314121037463985e-06,
+      "loss": 0.4777,
+      "step": 185
+    },
+    {
+      "epoch": 0.10963646855164455,
+      "grad_norm": 0.4120577956537649,
+      "learning_rate": 5.4755043227665705e-06,
+      "loss": 0.4834,
+      "step": 190
+    },
+    {
+      "epoch": 0.11252163877668782,
+      "grad_norm": 0.3970501930373392,
+      "learning_rate": 5.619596541786744e-06,
+      "loss": 0.4698,
+      "step": 195
+    },
+    {
+      "epoch": 0.1154068090017311,
+      "grad_norm": 0.41521712362317653,
+      "learning_rate": 5.763688760806917e-06,
+      "loss": 0.4854,
+      "step": 200
+    },
+    {
+      "epoch": 0.11829197922677438,
+      "grad_norm": 0.38302949433328365,
+      "learning_rate": 5.9077809798270905e-06,
+      "loss": 0.4603,
+      "step": 205
+    },
+    {
+      "epoch": 0.12117714945181765,
+      "grad_norm": 0.40356716686480254,
+      "learning_rate": 6.0518731988472625e-06,
+      "loss": 0.4617,
+      "step": 210
+    },
+    {
+      "epoch": 0.12406231967686093,
+      "grad_norm": 0.4108590792560944,
+      "learning_rate": 6.195965417867435e-06,
+      "loss": 0.4821,
+      "step": 215
+    },
+    {
+      "epoch": 0.12694748990190421,
+      "grad_norm": 0.38579700259781685,
+      "learning_rate": 6.340057636887609e-06,
+      "loss": 0.4548,
+      "step": 220
+    },
+    {
+      "epoch": 0.1298326601269475,
+      "grad_norm": 0.4008887277935931,
+      "learning_rate": 6.484149855907782e-06,
+      "loss": 0.4826,
+      "step": 225
+    },
+    {
+      "epoch": 0.13271783035199078,
+      "grad_norm": 0.4324585130384216,
+      "learning_rate": 6.6282420749279545e-06,
+      "loss": 0.4524,
+      "step": 230
+    },
+    {
+      "epoch": 0.13560300057703403,
+      "grad_norm": 0.36547940995267386,
+      "learning_rate": 6.772334293948127e-06,
+      "loss": 0.4472,
+      "step": 235
+    },
+    {
+      "epoch": 0.13848817080207732,
+      "grad_norm": 0.4021304207485496,
+      "learning_rate": 6.916426512968301e-06,
+      "loss": 0.469,
+      "step": 240
+    },
+    {
+      "epoch": 0.1413733410271206,
+      "grad_norm": 0.3986364193440875,
+      "learning_rate": 7.060518731988473e-06,
+      "loss": 0.4664,
+      "step": 245
+    },
+    {
+      "epoch": 0.14425851125216388,
+      "grad_norm": 0.3894372801996192,
+      "learning_rate": 7.204610951008646e-06,
+      "loss": 0.4596,
+      "step": 250
+    },
+    {
+      "epoch": 0.14714368147720716,
+      "grad_norm": 0.3822307947159079,
+      "learning_rate": 7.348703170028819e-06,
+      "loss": 0.4611,
+      "step": 255
+    },
+    {
+      "epoch": 0.15002885170225044,
+      "grad_norm": 0.3770508093447163,
+      "learning_rate": 7.492795389048992e-06,
+      "loss": 0.4513,
+      "step": 260
+    },
+    {
+      "epoch": 0.1529140219272937,
+      "grad_norm": 0.4040420054739231,
+      "learning_rate": 7.636887608069165e-06,
+      "loss": 0.4404,
+      "step": 265
+    },
+    {
+      "epoch": 0.15579919215233698,
+      "grad_norm": 0.3763215501882052,
+      "learning_rate": 7.780979827089338e-06,
+      "loss": 0.4484,
+      "step": 270
+    },
+    {
+      "epoch": 0.15868436237738026,
+      "grad_norm": 0.3823790521602452,
+      "learning_rate": 7.92507204610951e-06,
+      "loss": 0.4568,
+      "step": 275
+    },
+    {
+      "epoch": 0.16156953260242354,
+      "grad_norm": 0.3732843632883636,
+      "learning_rate": 8.069164265129685e-06,
+      "loss": 0.4389,
+      "step": 280
+    },
+    {
+      "epoch": 0.16445470282746683,
+      "grad_norm": 0.37767577343237896,
+      "learning_rate": 8.213256484149856e-06,
+      "loss": 0.4572,
+      "step": 285
+    },
+    {
+      "epoch": 0.1673398730525101,
+      "grad_norm": 0.3980275526344354,
+      "learning_rate": 8.357348703170029e-06,
+      "loss": 0.4625,
+      "step": 290
+    },
+    {
+      "epoch": 0.1702250432775534,
+      "grad_norm": 0.4046607258650182,
+      "learning_rate": 8.501440922190203e-06,
+      "loss": 0.488,
+      "step": 295
+    },
+    {
+      "epoch": 0.17311021350259664,
+      "grad_norm": 0.3759989273900334,
+      "learning_rate": 8.645533141210375e-06,
+      "loss": 0.4696,
+      "step": 300
+    },
+    {
+      "epoch": 0.17599538372763993,
+      "grad_norm": 0.3782703272325284,
+      "learning_rate": 8.789625360230547e-06,
+      "loss": 0.434,
+      "step": 305
+    },
+    {
+      "epoch": 0.1788805539526832,
+      "grad_norm": 0.3906713039092043,
+      "learning_rate": 8.933717579250722e-06,
+      "loss": 0.4565,
+      "step": 310
+    },
+    {
+      "epoch": 0.1817657241777265,
+      "grad_norm": 0.3805412690860938,
+      "learning_rate": 9.077809798270895e-06,
+      "loss": 0.448,
+      "step": 315
+    },
+    {
+      "epoch": 0.18465089440276977,
+      "grad_norm": 0.39402776095093084,
+      "learning_rate": 9.221902017291067e-06,
+      "loss": 0.4553,
+      "step": 320
+    },
+    {
+      "epoch": 0.18753606462781305,
+      "grad_norm": 0.45284451101047796,
+      "learning_rate": 9.36599423631124e-06,
+      "loss": 0.4447,
+      "step": 325
+    },
+    {
+      "epoch": 0.1904212348528563,
+      "grad_norm": 0.3873105480189736,
+      "learning_rate": 9.510086455331413e-06,
+      "loss": 0.4499,
+      "step": 330
+    },
+    {
+      "epoch": 0.1933064050778996,
+      "grad_norm": 0.40199355492029853,
+      "learning_rate": 9.654178674351586e-06,
+      "loss": 0.454,
+      "step": 335
+    },
+    {
+      "epoch": 0.19619157530294287,
+      "grad_norm": 0.38330387655003734,
+      "learning_rate": 9.798270893371759e-06,
+      "loss": 0.4484,
+      "step": 340
+    },
+    {
+      "epoch": 0.19907674552798615,
+      "grad_norm": 0.3790700593281963,
+      "learning_rate": 9.942363112391931e-06,
+      "loss": 0.4693,
+      "step": 345
+    },
+    {
+      "epoch": 0.20196191575302944,
+      "grad_norm": 0.4091985418093552,
+      "learning_rate": 1.0086455331412104e-05,
+      "loss": 0.4638,
+      "step": 350
+    },
+    {
+      "epoch": 0.20484708597807272,
+      "grad_norm": 0.3816167638203252,
+      "learning_rate": 1.0230547550432277e-05,
+      "loss": 0.4529,
+      "step": 355
+    },
+    {
+      "epoch": 0.20773225620311597,
+      "grad_norm": 0.3705496517450964,
+      "learning_rate": 1.037463976945245e-05,
+      "loss": 0.4491,
+      "step": 360
+    },
+    {
+      "epoch": 0.21061742642815925,
+      "grad_norm": 0.4096345206746822,
+      "learning_rate": 1.0518731988472624e-05,
+      "loss": 0.4769,
+      "step": 365
+    },
+    {
+      "epoch": 0.21350259665320254,
+      "grad_norm": 0.3828300212257734,
+      "learning_rate": 1.0662824207492797e-05,
+      "loss": 0.446,
+      "step": 370
+    },
+    {
+      "epoch": 0.21638776687824582,
+      "grad_norm": 0.3664642342578931,
+      "learning_rate": 1.080691642651297e-05,
+      "loss": 0.4629,
+      "step": 375
+    },
+    {
+      "epoch": 0.2192729371032891,
+      "grad_norm": 0.3701318700670979,
+      "learning_rate": 1.0951008645533141e-05,
+      "loss": 0.4454,
+      "step": 380
+    },
+    {
+      "epoch": 0.22215810732833238,
+      "grad_norm": 0.3765944116998475,
+      "learning_rate": 1.1095100864553314e-05,
+      "loss": 0.445,
+      "step": 385
+    },
+    {
+      "epoch": 0.22504327755337564,
+      "grad_norm": 0.35848655613640135,
+      "learning_rate": 1.1239193083573488e-05,
+      "loss": 0.4482,
+      "step": 390
+    },
+    {
+      "epoch": 0.22792844777841892,
+      "grad_norm": 0.3522345707962982,
+      "learning_rate": 1.1383285302593661e-05,
+      "loss": 0.4488,
+      "step": 395
+    },
+    {
+      "epoch": 0.2308136180034622,
+      "grad_norm": 0.42916841303669334,
+      "learning_rate": 1.1527377521613834e-05,
+      "loss": 0.4815,
+      "step": 400
+    },
+    {
+      "epoch": 0.23369878822850548,
+      "grad_norm": 0.39333826210382833,
+      "learning_rate": 1.1671469740634007e-05,
+      "loss": 0.4479,
+      "step": 405
+    },
+    {
+      "epoch": 0.23658395845354876,
+      "grad_norm": 0.395568573594799,
+      "learning_rate": 1.1815561959654181e-05,
+      "loss": 0.4375,
+      "step": 410
+    },
+    {
+      "epoch": 0.23946912867859205,
+      "grad_norm": 0.36804463353795325,
+      "learning_rate": 1.1959654178674354e-05,
+      "loss": 0.4544,
+      "step": 415
+    },
+    {
+      "epoch": 0.2423542989036353,
+      "grad_norm": 0.4057948653659005,
+      "learning_rate": 1.2103746397694525e-05,
+      "loss": 0.4489,
+      "step": 420
+    },
+    {
+      "epoch": 0.24523946912867858,
+      "grad_norm": 0.4042926809090149,
+      "learning_rate": 1.2247838616714698e-05,
+      "loss": 0.4432,
+      "step": 425
+    },
+    {
+      "epoch": 0.24812463935372187,
+      "grad_norm": 0.3746660432560446,
+      "learning_rate": 1.239193083573487e-05,
+      "loss": 0.454,
+      "step": 430
+    },
+    {
+      "epoch": 0.2510098095787651,
+      "grad_norm": 0.39348997561248156,
+      "learning_rate": 1.2536023054755045e-05,
+      "loss": 0.4647,
+      "step": 435
+    },
+    {
+      "epoch": 0.25389497980380843,
+      "grad_norm": 0.3872848928403898,
+      "learning_rate": 1.2680115273775218e-05,
+      "loss": 0.4535,
+      "step": 440
+    },
+    {
+      "epoch": 0.2567801500288517,
+      "grad_norm": 0.3813359635416143,
+      "learning_rate": 1.282420749279539e-05,
+      "loss": 0.4508,
+      "step": 445
+    },
+    {
+      "epoch": 0.259665320253895,
+      "grad_norm": 0.40132785614459726,
+      "learning_rate": 1.2968299711815563e-05,
+      "loss": 0.4549,
+      "step": 450
+    },
+    {
+      "epoch": 0.26255049047893825,
+      "grad_norm": 0.3674608489689301,
+      "learning_rate": 1.3112391930835735e-05,
+      "loss": 0.4519,
+      "step": 455
+    },
+    {
+      "epoch": 0.26543566070398156,
+      "grad_norm": 0.3672381265982974,
+      "learning_rate": 1.3256484149855909e-05,
+      "loss": 0.4504,
+      "step": 460
+    },
+    {
+      "epoch": 0.2683208309290248,
+      "grad_norm": 0.41466539433993804,
+      "learning_rate": 1.3400576368876082e-05,
+      "loss": 0.4509,
+      "step": 465
+    },
+    {
+      "epoch": 0.27120600115406807,
+      "grad_norm": 0.37606450867708474,
+      "learning_rate": 1.3544668587896255e-05,
+      "loss": 0.4646,
+      "step": 470
+    },
+    {
+      "epoch": 0.2740911713791114,
+      "grad_norm": 0.3868252718025835,
+      "learning_rate": 1.3688760806916427e-05,
+      "loss": 0.4395,
+      "step": 475
+    },
+    {
+      "epoch": 0.27697634160415463,
+      "grad_norm": 0.36956730373024904,
+      "learning_rate": 1.3832853025936602e-05,
+      "loss": 0.4421,
+      "step": 480
+    },
+    {
+      "epoch": 0.27986151182919794,
+      "grad_norm": 0.3703399882804596,
+      "learning_rate": 1.3976945244956775e-05,
+      "loss": 0.4288,
+      "step": 485
+    },
+    {
+      "epoch": 0.2827466820542412,
+      "grad_norm": 0.4006912338102629,
+      "learning_rate": 1.4121037463976946e-05,
+      "loss": 0.4721,
+      "step": 490
+    },
+    {
+      "epoch": 0.2856318522792845,
+      "grad_norm": 0.3443154071927776,
+      "learning_rate": 1.4265129682997119e-05,
+      "loss": 0.4322,
+      "step": 495
+    },
+    {
+      "epoch": 0.28851702250432776,
+      "grad_norm": 0.37915946639466674,
+      "learning_rate": 1.4409221902017291e-05,
+      "loss": 0.4264,
+      "step": 500
+    },
+    {
+      "epoch": 0.291402192729371,
+      "grad_norm": 0.38032216321312473,
+      "learning_rate": 1.4553314121037466e-05,
+      "loss": 0.4408,
+      "step": 505
+    },
+    {
+      "epoch": 0.2942873629544143,
+      "grad_norm": 0.40009286504791697,
+      "learning_rate": 1.4697406340057639e-05,
+      "loss": 0.4616,
+      "step": 510
+    },
+    {
+      "epoch": 0.2971725331794576,
+      "grad_norm": 0.35910974753646935,
+      "learning_rate": 1.4841498559077811e-05,
+      "loss": 0.4306,
+      "step": 515
+    },
+    {
+      "epoch": 0.3000577034045009,
+      "grad_norm": 0.37319150994773553,
+      "learning_rate": 1.4985590778097984e-05,
+      "loss": 0.4497,
+      "step": 520
+    },
+    {
+      "epoch": 0.30294287362954414,
+      "grad_norm": 0.3637719085871802,
+      "learning_rate": 1.5129682997118155e-05,
+      "loss": 0.4551,
+      "step": 525
+    },
+    {
+      "epoch": 0.3058280438545874,
+      "grad_norm": 0.3575259753574912,
+      "learning_rate": 1.527377521613833e-05,
+      "loss": 0.447,
+      "step": 530
+    },
+    {
+      "epoch": 0.3087132140796307,
+      "grad_norm": 0.3596026864789941,
+      "learning_rate": 1.54178674351585e-05,
+      "loss": 0.4642,
+      "step": 535
+    },
+    {
+      "epoch": 0.31159838430467396,
+      "grad_norm": 0.3613487147969066,
+      "learning_rate": 1.5561959654178675e-05,
+      "loss": 0.4384,
+      "step": 540
+    },
+    {
+      "epoch": 0.31448355452971727,
+      "grad_norm": 0.5248264510992934,
+      "learning_rate": 1.570605187319885e-05,
+      "loss": 0.4427,
+      "step": 545
+    },
+    {
+      "epoch": 0.3173687247547605,
+      "grad_norm": 0.3648196713346363,
+      "learning_rate": 1.585014409221902e-05,
+      "loss": 0.4649,
+      "step": 550
+    },
+    {
+      "epoch": 0.32025389497980383,
+      "grad_norm": 0.40066795102765307,
+      "learning_rate": 1.5994236311239196e-05,
+      "loss": 0.439,
+      "step": 555
+    },
+    {
+      "epoch": 0.3231390652048471,
+      "grad_norm": 0.33951584959069336,
+      "learning_rate": 1.613832853025937e-05,
+      "loss": 0.4432,
+      "step": 560
+    },
+    {
+      "epoch": 0.32602423542989034,
+      "grad_norm": 0.3533117973734279,
+      "learning_rate": 1.6282420749279538e-05,
+      "loss": 0.4335,
+      "step": 565
+    },
+    {
+      "epoch": 0.32890940565493365,
+      "grad_norm": 0.37863111987341347,
+      "learning_rate": 1.6426512968299712e-05,
+      "loss": 0.428,
+      "step": 570
+    },
+    {
+      "epoch": 0.3317945758799769,
+      "grad_norm": 0.4346675097845832,
+      "learning_rate": 1.6570605187319887e-05,
+      "loss": 0.449,
+      "step": 575
+    },
+    {
+      "epoch": 0.3346797461050202,
+      "grad_norm": 0.37817008201494773,
+      "learning_rate": 1.6714697406340058e-05,
+      "loss": 0.4433,
+      "step": 580
+    },
+    {
+      "epoch": 0.33756491633006347,
+      "grad_norm": 0.3808322285847144,
+      "learning_rate": 1.6858789625360232e-05,
+      "loss": 0.4466,
+      "step": 585
+    },
+    {
+      "epoch": 0.3404500865551068,
+      "grad_norm": 0.37923635129260924,
+      "learning_rate": 1.7002881844380407e-05,
+      "loss": 0.4476,
+      "step": 590
+    },
+    {
+      "epoch": 0.34333525678015003,
+      "grad_norm": 0.3823099849897589,
+      "learning_rate": 1.7146974063400578e-05,
+      "loss": 0.4323,
+      "step": 595
+    },
+    {
+      "epoch": 0.3462204270051933,
+      "grad_norm": 0.3818824327579753,
+      "learning_rate": 1.729106628242075e-05,
+      "loss": 0.4375,
+      "step": 600
+    },
+    {
+      "epoch": 0.3491055972302366,
+      "grad_norm": 0.37210993394094566,
+      "learning_rate": 1.7435158501440924e-05,
+      "loss": 0.4305,
+      "step": 605
+    },
+    {
+      "epoch": 0.35199076745527985,
+      "grad_norm": 0.3465073067639098,
+      "learning_rate": 1.7579250720461095e-05,
+      "loss": 0.4255,
+      "step": 610
+    },
+    {
+      "epoch": 0.35487593768032316,
+      "grad_norm": 0.38378004625397016,
+      "learning_rate": 1.772334293948127e-05,
+      "loss": 0.4324,
+      "step": 615
+    },
+    {
+      "epoch": 0.3577611079053664,
+      "grad_norm": 0.37432578456970733,
+      "learning_rate": 1.7867435158501444e-05,
+      "loss": 0.4543,
+      "step": 620
+    },
+    {
+      "epoch": 0.36064627813040967,
+      "grad_norm": 0.3341493702083905,
+      "learning_rate": 1.8011527377521615e-05,
+      "loss": 0.4501,
+      "step": 625
+    },
+    {
+      "epoch": 0.363531448355453,
+      "grad_norm": 0.3618443471925012,
+      "learning_rate": 1.815561959654179e-05,
+      "loss": 0.4387,
+      "step": 630
+    },
+    {
+      "epoch": 0.36641661858049623,
+      "grad_norm": 0.3493534570885224,
+      "learning_rate": 1.829971181556196e-05,
+      "loss": 0.4483,
+      "step": 635
+    },
+    {
+      "epoch": 0.36930178880553954,
+      "grad_norm": 0.37866953974278367,
+      "learning_rate": 1.8443804034582135e-05,
+      "loss": 0.4289,
+      "step": 640
+    },
+    {
+      "epoch": 0.3721869590305828,
+      "grad_norm": 0.37227424777091433,
+      "learning_rate": 1.8587896253602306e-05,
+      "loss": 0.4306,
+      "step": 645
+    },
+    {
+      "epoch": 0.3750721292556261,
+      "grad_norm": 0.4014872351370534,
+      "learning_rate": 1.873198847262248e-05,
+      "loss": 0.4414,
+      "step": 650
+    },
+    {
+      "epoch": 0.37795729948066936,
+      "grad_norm": 0.3724847659964483,
+      "learning_rate": 1.887608069164265e-05,
+      "loss": 0.4446,
+      "step": 655
+    },
+    {
+      "epoch": 0.3808424697057126,
+      "grad_norm": 0.3783020221193775,
+      "learning_rate": 1.9020172910662826e-05,
+      "loss": 0.4608,
+      "step": 660
+    },
+    {
+      "epoch": 0.3837276399307559,
+      "grad_norm": 0.4307776328223203,
+      "learning_rate": 1.9164265129683e-05,
+      "loss": 0.4343,
+      "step": 665
+    },
+    {
+      "epoch": 0.3866128101557992,
+      "grad_norm": 0.37291676085751585,
+      "learning_rate": 1.930835734870317e-05,
+      "loss": 0.4263,
+      "step": 670
+    },
+    {
+      "epoch": 0.3894979803808425,
+      "grad_norm": 0.3765824744528829,
+      "learning_rate": 1.9452449567723343e-05,
+      "loss": 0.4258,
+      "step": 675
+    },
+    {
+      "epoch": 0.39238315060588574,
+      "grad_norm": 0.354531651779117,
+      "learning_rate": 1.9596541786743517e-05,
+      "loss": 0.4504,
+      "step": 680
+    },
+    {
+      "epoch": 0.395268320830929,
+      "grad_norm": 0.3695419929782549,
+      "learning_rate": 1.974063400576369e-05,
+      "loss": 0.4663,
+      "step": 685
+    },
+    {
+      "epoch": 0.3981534910559723,
+      "grad_norm": 0.3752799655893121,
+      "learning_rate": 1.9884726224783863e-05,
+      "loss": 0.4519,
+      "step": 690
+    },
+    {
+      "epoch": 0.40103866128101556,
+      "grad_norm": 0.3743552771364968,
+      "learning_rate": 1.9999998731825553e-05,
+      "loss": 0.4466,
+      "step": 695
+    },
+    {
+      "epoch": 0.4039238315060589,
+      "grad_norm": 0.3690677017067688,
+      "learning_rate": 1.9999954345753522e-05,
+      "loss": 0.4672,
+      "step": 700
+    },
+    {
+      "epoch": 0.4068090017311021,
+      "grad_norm": 0.3521337986777979,
+      "learning_rate": 1.9999846551280566e-05,
+      "loss": 0.4546,
+      "step": 705
+    },
+    {
+      "epoch": 0.40969417195614544,
+      "grad_norm": 0.371125205175818,
+      "learning_rate": 1.99996753490902e-05,
+      "loss": 0.4476,
+      "step": 710
+    },
+    {
+      "epoch": 0.4125793421811887,
+      "grad_norm": 0.3870027577034139,
+      "learning_rate": 1.999944074026799e-05,
+      "loss": 0.4528,
+      "step": 715
+    },
+    {
+      "epoch": 0.41546451240623195,
+      "grad_norm": 0.3694599366137213,
+      "learning_rate": 1.999914272630156e-05,
+      "loss": 0.4435,
+      "step": 720
+    },
+    {
+      "epoch": 0.41834968263127525,
+      "grad_norm": 0.39229103038261187,
+      "learning_rate": 1.999878130908058e-05,
+      "loss": 0.4624,
+      "step": 725
+    },
+    {
+      "epoch": 0.4212348528563185,
+      "grad_norm": 0.37197227629586244,
+      "learning_rate": 1.999835649089675e-05,
+      "loss": 0.4392,
+      "step": 730
+    },
+    {
+      "epoch": 0.4241200230813618,
+      "grad_norm": 0.36658551403514933,
+      "learning_rate": 1.9997868274443787e-05,
+      "loss": 0.4479,
+      "step": 735
+    },
+    {
+      "epoch": 0.4270051933064051,
+      "grad_norm": 0.38507395642784864,
+      "learning_rate": 1.9997316662817403e-05,
+      "loss": 0.4448,
+      "step": 740
+    },
+    {
+      "epoch": 0.4298903635314484,
+      "grad_norm": 0.36930658832139956,
+      "learning_rate": 1.9996701659515293e-05,
+      "loss": 0.4338,
+      "step": 745
+    },
+    {
+      "epoch": 0.43277553375649164,
+      "grad_norm": 0.3415004408124957,
+      "learning_rate": 1.999602326843712e-05,
+      "loss": 0.4571,
+      "step": 750
+    },
+    {
+      "epoch": 0.4356607039815349,
+      "grad_norm": 0.3616392339336591,
+      "learning_rate": 1.999528149388447e-05,
+      "loss": 0.4364,
+      "step": 755
+    },
+    {
+      "epoch": 0.4385458742065782,
+      "grad_norm": 0.3614440903798367,
+      "learning_rate": 1.999447634056084e-05,
+      "loss": 0.442,
+      "step": 760
+    },
+    {
+      "epoch": 0.44143104443162146,
+      "grad_norm": 0.3475594464185632,
+      "learning_rate": 1.9993607813571595e-05,
+      "loss": 0.4302,
+      "step": 765
+    },
+    {
+      "epoch": 0.44431621465666477,
+      "grad_norm": 0.38907218952119343,
+      "learning_rate": 1.9992675918423958e-05,
+      "loss": 0.4475,
+      "step": 770
+    },
+    {
+      "epoch": 0.447201384881708,
+      "grad_norm": 0.3532869868641272,
+      "learning_rate": 1.999168066102695e-05,
+      "loss": 0.4486,
+      "step": 775
+    },
+    {
+      "epoch": 0.4500865551067513,
+      "grad_norm": 0.37651731293442015,
+      "learning_rate": 1.9990622047691372e-05,
+      "loss": 0.4326,
+      "step": 780
+    },
+    {
+      "epoch": 0.4529717253317946,
+      "grad_norm": 0.3748786150498863,
+      "learning_rate": 1.998950008512975e-05,
+      "loss": 0.4474,
+      "step": 785
+    },
+    {
+      "epoch": 0.45585689555683784,
+      "grad_norm": 0.39902509681503917,
+      "learning_rate": 1.9988314780456305e-05,
+      "loss": 0.4568,
+      "step": 790
+    },
+    {
+      "epoch": 0.45874206578188115,
+      "grad_norm": 0.3884381984087952,
+      "learning_rate": 1.998706614118689e-05,
+      "loss": 0.4568,
+      "step": 795
+    },
+    {
+      "epoch": 0.4616272360069244,
+      "grad_norm": 0.36054442749164206,
+      "learning_rate": 1.9985754175238972e-05,
+      "loss": 0.4429,
+      "step": 800
+    },
+    {
+      "epoch": 0.4645124062319677,
+      "grad_norm": 0.3527339288196345,
+      "learning_rate": 1.998437889093155e-05,
+      "loss": 0.4437,
+      "step": 805
+    },
+    {
+      "epoch": 0.46739757645701097,
+      "grad_norm": 0.3737014283866458,
+      "learning_rate": 1.9982940296985124e-05,
+      "loss": 0.4469,
+      "step": 810
+    },
+    {
+      "epoch": 0.4702827466820542,
+      "grad_norm": 0.3836320689851165,
+      "learning_rate": 1.998143840252163e-05,
+      "loss": 0.4391,
+      "step": 815
+    },
+    {
+      "epoch": 0.47316791690709753,
+      "grad_norm": 0.36567112690649783,
+      "learning_rate": 1.997987321706438e-05,
+      "loss": 0.4414,
+      "step": 820
+    },
+    {
+      "epoch": 0.4760530871321408,
+      "grad_norm": 0.37388859283028064,
+      "learning_rate": 1.9978244750538017e-05,
+      "loss": 0.4408,
+      "step": 825
+    },
+    {
+      "epoch": 0.4789382573571841,
+      "grad_norm": 0.35288917666405584,
+      "learning_rate": 1.9976553013268428e-05,
+      "loss": 0.4432,
+      "step": 830
+    },
+    {
+      "epoch": 0.48182342758222735,
+      "grad_norm": 0.35467525608999173,
+      "learning_rate": 1.9974798015982704e-05,
+      "loss": 0.4319,
+      "step": 835
+    },
+    {
+      "epoch": 0.4847085978072706,
+      "grad_norm": 0.36246807797824815,
+      "learning_rate": 1.9972979769809044e-05,
+      "loss": 0.4421,
+      "step": 840
+    },
+    {
+      "epoch": 0.4875937680323139,
+      "grad_norm": 0.38809255131157183,
+      "learning_rate": 1.9971098286276716e-05,
+      "loss": 0.438,
+      "step": 845
+    },
+    {
+      "epoch": 0.49047893825735717,
+      "grad_norm": 0.35620554170658003,
+      "learning_rate": 1.996915357731596e-05,
+      "loss": 0.4361,
+      "step": 850
+    },
+    {
+      "epoch": 0.4933641084824005,
+      "grad_norm": 0.37319706651769974,
+      "learning_rate": 1.996714565525792e-05,
+      "loss": 0.4273,
+      "step": 855
+    },
+    {
+      "epoch": 0.49624927870744373,
+      "grad_norm": 0.3561234532855242,
+      "learning_rate": 1.996507453283456e-05,
+      "loss": 0.4405,
+      "step": 860
+    },
+    {
+      "epoch": 0.49913444893248704,
+      "grad_norm": 0.3599439467426677,
+      "learning_rate": 1.9962940223178607e-05,
+      "loss": 0.4209,
+      "step": 865
+    },
+    {
+      "epoch": 0.5020196191575302,
+      "grad_norm": 0.3708191889641566,
+      "learning_rate": 1.9960742739823433e-05,
+      "loss": 0.4375,
+      "step": 870
+    },
+    {
+      "epoch": 0.5049047893825735,
+      "grad_norm": 0.371986303129069,
+      "learning_rate": 1.9958482096702997e-05,
+      "loss": 0.44,
+      "step": 875
+    },
+    {
+      "epoch": 0.5077899596076169,
+      "grad_norm": 0.37613060788442565,
+      "learning_rate": 1.995615830815173e-05,
+      "loss": 0.444,
+      "step": 880
+    },
+    {
+      "epoch": 0.5106751298326602,
+      "grad_norm": 0.35285659183130846,
+      "learning_rate": 1.9953771388904484e-05,
+      "loss": 0.4483,
+      "step": 885
+    },
+    {
+      "epoch": 0.5135603000577034,
+      "grad_norm": 0.3744385315940648,
+      "learning_rate": 1.9951321354096395e-05,
+      "loss": 0.4307,
+      "step": 890
+    },
+    {
+      "epoch": 0.5164454702827467,
+      "grad_norm": 0.3527685970235997,
+      "learning_rate": 1.9948808219262813e-05,
+      "loss": 0.4352,
+      "step": 895
+    },
+    {
+      "epoch": 0.51933064050779,
+      "grad_norm": 0.3629898148956746,
+      "learning_rate": 1.9946232000339192e-05,
+      "loss": 0.4325,
+      "step": 900
+    },
+    {
+      "epoch": 0.5222158107328332,
+      "grad_norm": 0.33230153297658693,
+      "learning_rate": 1.994359271366101e-05,
+      "loss": 0.4358,
+      "step": 905
+    },
+    {
+      "epoch": 0.5251009809578765,
+      "grad_norm": 0.36872688475457055,
+      "learning_rate": 1.9940890375963634e-05,
+      "loss": 0.4389,
+      "step": 910
+    },
+    {
+      "epoch": 0.5279861511829198,
+      "grad_norm": 0.3500504418913126,
+      "learning_rate": 1.9938125004382226e-05,
+      "loss": 0.4368,
+      "step": 915
+    },
+    {
+      "epoch": 0.5308713214079631,
+      "grad_norm": 0.38076969493807267,
+      "learning_rate": 1.9935296616451654e-05,
+      "loss": 0.4224,
+      "step": 920
+    },
+    {
+      "epoch": 0.5337564916330063,
+      "grad_norm": 0.38809106802534876,
+      "learning_rate": 1.9932405230106352e-05,
+      "loss": 0.458,
+      "step": 925
+    },
+    {
+      "epoch": 0.5366416618580496,
+      "grad_norm": 0.34830821561529673,
+      "learning_rate": 1.992945086368022e-05,
+      "loss": 0.4427,
+      "step": 930
+    },
+    {
+      "epoch": 0.5395268320830929,
+      "grad_norm": 0.36348521739778694,
+      "learning_rate": 1.9926433535906514e-05,
+      "loss": 0.4332,
+      "step": 935
+    },
+    {
+      "epoch": 0.5424120023081361,
+      "grad_norm": 0.3516690545778123,
+      "learning_rate": 1.992335326591771e-05,
+      "loss": 0.4493,
+      "step": 940
+    },
+    {
+      "epoch": 0.5452971725331794,
+      "grad_norm": 0.3622130880303389,
+      "learning_rate": 1.99202100732454e-05,
+      "loss": 0.4399,
+      "step": 945
+    },
+    {
+      "epoch": 0.5481823427582228,
+      "grad_norm": 0.3340397775809456,
+      "learning_rate": 1.9917003977820154e-05,
+      "loss": 0.4155,
+      "step": 950
+    },
+    {
+      "epoch": 0.5510675129832661,
+      "grad_norm": 0.3373549627806811,
+      "learning_rate": 1.9913734999971402e-05,
+      "loss": 0.4343,
+      "step": 955
+    },
+    {
+      "epoch": 0.5539526832083093,
+      "grad_norm": 0.34945896571648644,
+      "learning_rate": 1.9910403160427308e-05,
+      "loss": 0.4423,
+      "step": 960
+    },
+    {
+      "epoch": 0.5568378534333526,
+      "grad_norm": 0.3477598210882681,
+      "learning_rate": 1.990700848031463e-05,
+      "loss": 0.4328,
+      "step": 965
+    },
+    {
+      "epoch": 0.5597230236583959,
+      "grad_norm": 0.3728272868907162,
+      "learning_rate": 1.990355098115859e-05,
+      "loss": 0.4407,
+      "step": 970
+    },
+    {
+      "epoch": 0.5626081938834391,
+      "grad_norm": 0.36540010757747865,
+      "learning_rate": 1.990003068488273e-05,
+      "loss": 0.4254,
+      "step": 975
+    },
+    {
+      "epoch": 0.5654933641084824,
+      "grad_norm": 0.3491338839333341,
+      "learning_rate": 1.9896447613808802e-05,
+      "loss": 0.4333,
+      "step": 980
+    },
+    {
+      "epoch": 0.5683785343335257,
+      "grad_norm": 0.3453766808693869,
+      "learning_rate": 1.9892801790656575e-05,
+      "loss": 0.4505,
+      "step": 985
+    },
+    {
+      "epoch": 0.571263704558569,
+      "grad_norm": 0.3663489965490133,
+      "learning_rate": 1.9889093238543746e-05,
+      "loss": 0.4308,
+      "step": 990
+    },
+    {
+      "epoch": 0.5741488747836122,
+      "grad_norm": 0.3679457848990051,
+      "learning_rate": 1.9885321980985758e-05,
+      "loss": 0.4594,
+      "step": 995
+    },
+    {
+      "epoch": 0.5770340450086555,
+      "grad_norm": 0.3780068047990749,
+      "learning_rate": 1.988148804189566e-05,
+      "loss": 0.4255,
+      "step": 1000
+    },
+    {
+      "epoch": 0.5799192152336988,
+      "grad_norm": 0.36727910455394897,
+      "learning_rate": 1.987759144558395e-05,
+      "loss": 0.4378,
+      "step": 1005
+    },
+    {
+      "epoch": 0.582804385458742,
+      "grad_norm": 0.36166322845028065,
+      "learning_rate": 1.9873632216758448e-05,
+      "loss": 0.4445,
+      "step": 1010
+    },
+    {
+      "epoch": 0.5856895556837853,
+      "grad_norm": 0.3412780788439251,
+      "learning_rate": 1.9869610380524098e-05,
+      "loss": 0.4285,
+      "step": 1015
+    },
+    {
+      "epoch": 0.5885747259088286,
+      "grad_norm": 0.3642868390810403,
+      "learning_rate": 1.986552596238284e-05,
+      "loss": 0.426,
+      "step": 1020
+    },
+    {
+      "epoch": 0.5914598961338718,
+      "grad_norm": 0.3513636822125392,
+      "learning_rate": 1.986137898823343e-05,
+      "loss": 0.4689,
+      "step": 1025
+    },
+    {
+      "epoch": 0.5943450663589152,
+      "grad_norm": 0.37061082820118274,
+      "learning_rate": 1.985716948437129e-05,
+      "loss": 0.4396,
+      "step": 1030
+    },
+    {
+      "epoch": 0.5972302365839585,
+      "grad_norm": 0.35578918227279527,
+      "learning_rate": 1.9852897477488335e-05,
+      "loss": 0.4347,
+      "step": 1035
+    },
+    {
+      "epoch": 0.6001154068090018,
+      "grad_norm": 0.37785556076696936,
+      "learning_rate": 1.9848562994672797e-05,
+      "loss": 0.4272,
+      "step": 1040
+    },
+    {
+      "epoch": 0.603000577034045,
+      "grad_norm": 0.36978022207552363,
+      "learning_rate": 1.984416606340907e-05,
+      "loss": 0.4326,
+      "step": 1045
+    },
+    {
+      "epoch": 0.6058857472590883,
+      "grad_norm": 0.3791504839723979,
+      "learning_rate": 1.9839706711577514e-05,
+      "loss": 0.4587,
+      "step": 1050
+    },
+    {
+      "epoch": 0.6087709174841316,
+      "grad_norm": 0.3710313865819546,
+      "learning_rate": 1.98351849674543e-05,
+      "loss": 0.457,
+      "step": 1055
+    },
+    {
+      "epoch": 0.6116560877091748,
+      "grad_norm": 0.35231701107663654,
+      "learning_rate": 1.9830600859711207e-05,
+      "loss": 0.4144,
+      "step": 1060
+    },
+    {
+      "epoch": 0.6145412579342181,
+      "grad_norm": 0.3615898765366417,
+      "learning_rate": 1.982595441741547e-05,
+      "loss": 0.4401,
+      "step": 1065
+    },
+    {
+      "epoch": 0.6174264281592614,
+      "grad_norm": 0.35977723996205,
+      "learning_rate": 1.9821245670029566e-05,
+      "loss": 0.4309,
+      "step": 1070
+    },
+    {
+      "epoch": 0.6203115983843047,
+      "grad_norm": 0.36669507423796843,
+      "learning_rate": 1.9816474647411047e-05,
+      "loss": 0.4383,
+      "step": 1075
+    },
+    {
+      "epoch": 0.6231967686093479,
+      "grad_norm": 0.35392661751275434,
+      "learning_rate": 1.9811641379812342e-05,
+      "loss": 0.446,
+      "step": 1080
+    },
+    {
+      "epoch": 0.6260819388343912,
+      "grad_norm": 0.3409833418281117,
+      "learning_rate": 1.9806745897880573e-05,
+      "loss": 0.4367,
+      "step": 1085
+    },
+    {
+      "epoch": 0.6289671090594345,
+      "grad_norm": 0.37150843192119876,
+      "learning_rate": 1.9801788232657343e-05,
+      "loss": 0.4527,
+      "step": 1090
+    },
+    {
+      "epoch": 0.6318522792844777,
+      "grad_norm": 0.36569227925348624,
+      "learning_rate": 1.9796768415578564e-05,
+      "loss": 0.4436,
+      "step": 1095
+    },
+    {
+      "epoch": 0.634737449509521,
+      "grad_norm": 0.35016739834472266,
+      "learning_rate": 1.9791686478474235e-05,
+      "loss": 0.4368,
+      "step": 1100
+    },
+    {
+      "epoch": 0.6376226197345644,
+      "grad_norm": 0.36715520059954826,
+      "learning_rate": 1.978654245356825e-05,
+      "loss": 0.4354,
+      "step": 1105
+    },
+    {
+      "epoch": 0.6405077899596077,
+      "grad_norm": 0.3840761403084782,
+      "learning_rate": 1.9781336373478207e-05,
+      "loss": 0.4523,
+      "step": 1110
+    },
+    {
+      "epoch": 0.6433929601846509,
+      "grad_norm": 0.3154658430275838,
+      "learning_rate": 1.9776068271215167e-05,
+      "loss": 0.4406,
+      "step": 1115
+    },
+    {
+      "epoch": 0.6462781304096942,
+      "grad_norm": 0.3772719849710032,
+      "learning_rate": 1.9770738180183485e-05,
+      "loss": 0.4344,
+      "step": 1120
+    },
+    {
+      "epoch": 0.6491633006347375,
+      "grad_norm": 0.3381456770156173,
+      "learning_rate": 1.9765346134180567e-05,
+      "loss": 0.4324,
+      "step": 1125
+    },
+    {
+      "epoch": 0.6520484708597807,
+      "grad_norm": 0.3498102184043816,
+      "learning_rate": 1.975989216739667e-05,
+      "loss": 0.4294,
+      "step": 1130
+    },
+    {
+      "epoch": 0.654933641084824,
+      "grad_norm": 0.3449183171225279,
+      "learning_rate": 1.9754376314414686e-05,
+      "loss": 0.4139,
+      "step": 1135
+    },
+    {
+      "epoch": 0.6578188113098673,
+      "grad_norm": 0.3541093995239644,
+      "learning_rate": 1.9748798610209912e-05,
+      "loss": 0.4295,
+      "step": 1140
+    },
+    {
+      "epoch": 0.6607039815349106,
+      "grad_norm": 0.36096527367909803,
+      "learning_rate": 1.9743159090149844e-05,
+      "loss": 0.4605,
+      "step": 1145
+    },
+    {
+      "epoch": 0.6635891517599538,
+      "grad_norm": 0.334233191996567,
+      "learning_rate": 1.9737457789993936e-05,
+      "loss": 0.4368,
+      "step": 1150
+    },
+    {
+      "epoch": 0.6664743219849971,
+      "grad_norm": 0.35974161365839685,
+      "learning_rate": 1.973169474589339e-05,
+      "loss": 0.428,
+      "step": 1155
+    },
+    {
+      "epoch": 0.6693594922100404,
+      "grad_norm": 0.3568705187506466,
+      "learning_rate": 1.9725869994390912e-05,
+      "loss": 0.4417,
+      "step": 1160
+    },
+    {
+      "epoch": 0.6722446624350836,
+      "grad_norm": 0.33809479002108944,
+      "learning_rate": 1.971998357242049e-05,
+      "loss": 0.4413,
+      "step": 1165
+    },
+    {
+      "epoch": 0.6751298326601269,
+      "grad_norm": 0.3502193821967476,
+      "learning_rate": 1.971403551730715e-05,
+      "loss": 0.4564,
+      "step": 1170
+    },
+    {
+      "epoch": 0.6780150028851702,
+      "grad_norm": 0.35976087067679435,
+      "learning_rate": 1.9708025866766735e-05,
+      "loss": 0.4263,
+      "step": 1175
+    },
+    {
+      "epoch": 0.6809001731102136,
+      "grad_norm": 0.34825014902921947,
+      "learning_rate": 1.9701954658905652e-05,
+      "loss": 0.4324,
+      "step": 1180
+    },
+    {
+      "epoch": 0.6837853433352568,
+      "grad_norm": 0.3515962597578983,
+      "learning_rate": 1.9695821932220632e-05,
+      "loss": 0.4347,
+      "step": 1185
+    },
+    {
+      "epoch": 0.6866705135603001,
+      "grad_norm": 0.3518698997384743,
+      "learning_rate": 1.9689627725598495e-05,
+      "loss": 0.4141,
+      "step": 1190
+    },
+    {
+      "epoch": 0.6895556837853434,
+      "grad_norm": 0.3406825005378319,
+      "learning_rate": 1.9683372078315892e-05,
+      "loss": 0.4294,
+      "step": 1195
+    },
+    {
+      "epoch": 0.6924408540103866,
+      "grad_norm": 0.3555369233516162,
+      "learning_rate": 1.9677055030039067e-05,
+      "loss": 0.4389,
+      "step": 1200
+    },
+    {
+      "epoch": 0.6953260242354299,
+      "grad_norm": 0.4609971974027847,
+      "learning_rate": 1.967067662082359e-05,
+      "loss": 0.4479,
+      "step": 1205
+    },
+    {
+      "epoch": 0.6982111944604732,
+      "grad_norm": 0.3533102697284043,
+      "learning_rate": 1.966423689111412e-05,
+      "loss": 0.4594,
+      "step": 1210
+    },
+    {
+      "epoch": 0.7010963646855164,
+      "grad_norm": 0.41163003146138266,
+      "learning_rate": 1.965773588174414e-05,
+      "loss": 0.4211,
+      "step": 1215
+    },
+    {
+      "epoch": 0.7039815349105597,
+      "grad_norm": 0.36355249316504007,
+      "learning_rate": 1.9651173633935702e-05,
+      "loss": 0.4326,
+      "step": 1220
+    },
+    {
+      "epoch": 0.706866705135603,
+      "grad_norm": 0.361512663082872,
+      "learning_rate": 1.964455018929916e-05,
+      "loss": 0.4605,
+      "step": 1225
+    },
+    {
+      "epoch": 0.7097518753606463,
+      "grad_norm": 0.3395334945396585,
+      "learning_rate": 1.9637865589832902e-05,
+      "loss": 0.4552,
+      "step": 1230
+    },
+    {
+      "epoch": 0.7126370455856895,
+      "grad_norm": 0.3358166664046763,
+      "learning_rate": 1.963111987792311e-05,
+      "loss": 0.4418,
+      "step": 1235
+    },
+    {
+      "epoch": 0.7155222158107328,
+      "grad_norm": 0.35952927832201437,
+      "learning_rate": 1.9624313096343445e-05,
+      "loss": 0.4528,
+      "step": 1240
+    },
+    {
+      "epoch": 0.7184073860357761,
+      "grad_norm": 0.3516252150445922,
+      "learning_rate": 1.961744528825483e-05,
+      "loss": 0.4444,
+      "step": 1245
+    },
+    {
+      "epoch": 0.7212925562608193,
+      "grad_norm": 0.37486390370155487,
+      "learning_rate": 1.961051649720513e-05,
+      "loss": 0.4405,
+      "step": 1250
+    },
+    {
+      "epoch": 0.7241777264858626,
+      "grad_norm": 0.347168899689223,
+      "learning_rate": 1.96035267671289e-05,
+      "loss": 0.4197,
+      "step": 1255
+    },
+    {
+      "epoch": 0.727062896710906,
+      "grad_norm": 0.3553448619523881,
+      "learning_rate": 1.9596476142347114e-05,
+      "loss": 0.4273,
+      "step": 1260
+    },
+    {
+      "epoch": 0.7299480669359493,
+      "grad_norm": 0.3747743065569822,
+      "learning_rate": 1.958936466756685e-05,
+      "loss": 0.4318,
+      "step": 1265
+    },
+    {
+      "epoch": 0.7328332371609925,
+      "grad_norm": 0.36946705338179187,
+      "learning_rate": 1.9582192387881037e-05,
+      "loss": 0.439,
+      "step": 1270
+    },
+    {
+      "epoch": 0.7357184073860358,
+      "grad_norm": 0.3535008834153477,
+      "learning_rate": 1.9574959348768173e-05,
+      "loss": 0.4215,
+      "step": 1275
+    },
+    {
+      "epoch": 0.7386035776110791,
+      "grad_norm": 0.3528743884422751,
+      "learning_rate": 1.9567665596092e-05,
+      "loss": 0.4451,
+      "step": 1280
+    },
+    {
+      "epoch": 0.7414887478361223,
+      "grad_norm": 0.36361502852826977,
+      "learning_rate": 1.9560311176101253e-05,
+      "loss": 0.4488,
+      "step": 1285
+    },
+    {
+      "epoch": 0.7443739180611656,
+      "grad_norm": 0.3393383444681419,
+      "learning_rate": 1.9552896135429346e-05,
+      "loss": 0.4602,
+      "step": 1290
+    },
+    {
+      "epoch": 0.7472590882862089,
+      "grad_norm": 0.3504869096195613,
+      "learning_rate": 1.9545420521094082e-05,
+      "loss": 0.4213,
+      "step": 1295
+    },
+    {
+      "epoch": 0.7501442585112522,
+      "grad_norm": 0.33568870619524627,
+      "learning_rate": 1.953788438049735e-05,
+      "loss": 0.4159,
+      "step": 1300
+    },
+    {
+      "epoch": 0.7530294287362954,
+      "grad_norm": 0.349599888573523,
+      "learning_rate": 1.953028776142483e-05,
+      "loss": 0.4245,
+      "step": 1305
+    },
+    {
+      "epoch": 0.7559145989613387,
+      "grad_norm": 0.36044335412116235,
+      "learning_rate": 1.9522630712045695e-05,
+      "loss": 0.4253,
+      "step": 1310
+    },
+    {
+      "epoch": 0.758799769186382,
+      "grad_norm": 0.33120277929093034,
+      "learning_rate": 1.9514913280912284e-05,
+      "loss": 0.4196,
+      "step": 1315
+    },
+    {
+      "epoch": 0.7616849394114252,
+      "grad_norm": 0.3556004564806594,
+      "learning_rate": 1.9507135516959824e-05,
+      "loss": 0.4319,
+      "step": 1320
+    },
+    {
+      "epoch": 0.7645701096364685,
+      "grad_norm": 0.357624065192376,
+      "learning_rate": 1.949929746950609e-05,
+      "loss": 0.4464,
+      "step": 1325
+    },
+    {
+      "epoch": 0.7674552798615119,
+      "grad_norm": 0.3429194439395415,
+      "learning_rate": 1.9491399188251123e-05,
+      "loss": 0.4384,
+      "step": 1330
+    },
+    {
+      "epoch": 0.7703404500865552,
+      "grad_norm": 0.35807803811180317,
+      "learning_rate": 1.9483440723276887e-05,
+      "loss": 0.4274,
+      "step": 1335
+    },
+    {
+      "epoch": 0.7732256203115984,
+      "grad_norm": 0.3516485111182853,
+      "learning_rate": 1.9475422125046962e-05,
+      "loss": 0.4314,
+      "step": 1340
+    },
+    {
+      "epoch": 0.7761107905366417,
+      "grad_norm": 0.3448017580215496,
+      "learning_rate": 1.9467343444406234e-05,
+      "loss": 0.4305,
+      "step": 1345
+    },
+    {
+      "epoch": 0.778995960761685,
+      "grad_norm": 0.33719191373319685,
+      "learning_rate": 1.9459204732580557e-05,
+      "loss": 0.4209,
+      "step": 1350
+    },
+    {
+      "epoch": 0.7818811309867282,
+      "grad_norm": 0.357112111873714,
+      "learning_rate": 1.9451006041176437e-05,
+      "loss": 0.4614,
+      "step": 1355
+    },
+    {
+      "epoch": 0.7847663012117715,
+      "grad_norm": 0.3638827997486545,
+      "learning_rate": 1.9442747422180704e-05,
+      "loss": 0.4392,
+      "step": 1360
+    },
+    {
+      "epoch": 0.7876514714368148,
+      "grad_norm": 0.3779229355078506,
+      "learning_rate": 1.9434428927960177e-05,
+      "loss": 0.4175,
+      "step": 1365
+    },
+    {
+      "epoch": 0.790536641661858,
+      "grad_norm": 0.33870833738392225,
+      "learning_rate": 1.9426050611261343e-05,
+      "loss": 0.4378,
+      "step": 1370
+    },
+    {
+      "epoch": 0.7934218118869013,
+      "grad_norm": 0.3407658218095889,
+      "learning_rate": 1.9417612525210007e-05,
+      "loss": 0.4244,
+      "step": 1375
+    },
+    {
+      "epoch": 0.7963069821119446,
+      "grad_norm": 0.3587456861468758,
+      "learning_rate": 1.9409114723310967e-05,
+      "loss": 0.4296,
+      "step": 1380
+    },
+    {
+      "epoch": 0.7991921523369879,
+      "grad_norm": 0.3526548403605731,
+      "learning_rate": 1.9400557259447673e-05,
+      "loss": 0.4334,
+      "step": 1385
+    },
+    {
+      "epoch": 0.8020773225620311,
+      "grad_norm": 0.35126996708797126,
+      "learning_rate": 1.9391940187881885e-05,
+      "loss": 0.4455,
+      "step": 1390
+    },
+    {
+      "epoch": 0.8049624927870744,
+      "grad_norm": 0.35018120101424355,
+      "learning_rate": 1.9383263563253322e-05,
+      "loss": 0.4432,
+      "step": 1395
+    },
+    {
+      "epoch": 0.8078476630121177,
+      "grad_norm": 0.3375686338902311,
+      "learning_rate": 1.937452744057933e-05,
+      "loss": 0.4453,
+      "step": 1400
+    },
+    {
+      "epoch": 0.8107328332371609,
+      "grad_norm": 0.34947963167148055,
+      "learning_rate": 1.936573187525451e-05,
+      "loss": 0.4392,
+      "step": 1405
+    },
+    {
+      "epoch": 0.8136180034622043,
+      "grad_norm": 0.329777575614968,
+      "learning_rate": 1.9356876923050395e-05,
+      "loss": 0.4554,
+      "step": 1410
+    },
+    {
+      "epoch": 0.8165031736872476,
+      "grad_norm": 0.3481917738275054,
+      "learning_rate": 1.9347962640115078e-05,
+      "loss": 0.4172,
+      "step": 1415
+    },
+    {
+      "epoch": 0.8193883439122909,
+      "grad_norm": 0.3488032875220557,
+      "learning_rate": 1.933898908297286e-05,
+      "loss": 0.4471,
+      "step": 1420
+    },
+    {
+      "epoch": 0.8222735141373341,
+      "grad_norm": 0.35259907680989855,
+      "learning_rate": 1.9329956308523886e-05,
+      "loss": 0.4265,
+      "step": 1425
+    },
+    {
+      "epoch": 0.8251586843623774,
+      "grad_norm": 0.36772785974509603,
+      "learning_rate": 1.9320864374043805e-05,
+      "loss": 0.4346,
+      "step": 1430
+    },
+    {
+      "epoch": 0.8280438545874207,
+      "grad_norm": 0.33906704838185237,
+      "learning_rate": 1.9311713337183375e-05,
+      "loss": 0.4222,
+      "step": 1435
+    },
+    {
+      "epoch": 0.8309290248124639,
+      "grad_norm": 0.3197548387537902,
+      "learning_rate": 1.9302503255968127e-05,
+      "loss": 0.4242,
+      "step": 1440
+    },
+    {
+      "epoch": 0.8338141950375072,
+      "grad_norm": 0.3790505826213643,
+      "learning_rate": 1.929323418879798e-05,
+      "loss": 0.444,
+      "step": 1445
+    },
+    {
+      "epoch": 0.8366993652625505,
+      "grad_norm": 0.356548976452809,
+      "learning_rate": 1.928390619444687e-05,
+      "loss": 0.4389,
+      "step": 1450
+    },
+    {
+      "epoch": 0.8395845354875938,
+      "grad_norm": 0.3238309837351344,
+      "learning_rate": 1.9274519332062393e-05,
+      "loss": 0.4428,
+      "step": 1455
+    },
+    {
+      "epoch": 0.842469705712637,
+      "grad_norm": 0.3328195979317483,
+      "learning_rate": 1.9265073661165407e-05,
+      "loss": 0.4418,
+      "step": 1460
+    },
+    {
+      "epoch": 0.8453548759376803,
+      "grad_norm": 0.34611626618940344,
+      "learning_rate": 1.9255569241649685e-05,
+      "loss": 0.4311,
+      "step": 1465
+    },
+    {
+      "epoch": 0.8482400461627236,
+      "grad_norm": 0.34500455296149757,
+      "learning_rate": 1.9246006133781496e-05,
+      "loss": 0.4307,
+      "step": 1470
+    },
+    {
+      "epoch": 0.8511252163877668,
+      "grad_norm": 0.3473701533278901,
+      "learning_rate": 1.9236384398199262e-05,
+      "loss": 0.4484,
+      "step": 1475
+    },
+    {
+      "epoch": 0.8540103866128101,
+      "grad_norm": 0.3409397737914092,
+      "learning_rate": 1.9226704095913143e-05,
+      "loss": 0.4247,
+      "step": 1480
+    },
+    {
+      "epoch": 0.8568955568378535,
+      "grad_norm": 0.36709716826008937,
+      "learning_rate": 1.9216965288304675e-05,
+      "loss": 0.4454,
+      "step": 1485
+    },
+    {
+      "epoch": 0.8597807270628968,
+      "grad_norm": 0.32825230009567846,
+      "learning_rate": 1.9207168037126352e-05,
+      "loss": 0.4246,
+      "step": 1490
+    },
+    {
+      "epoch": 0.86266589728794,
+      "grad_norm": 0.33313005166969667,
+      "learning_rate": 1.919731240450127e-05,
+      "loss": 0.4204,
+      "step": 1495
+    },
+    {
+      "epoch": 0.8655510675129833,
+      "grad_norm": 0.34672496584019735,
+      "learning_rate": 1.91873984529227e-05,
+      "loss": 0.4469,
+      "step": 1500
+    },
+    {
+      "epoch": 0.8684362377380266,
+      "grad_norm": 0.3342264609614132,
+      "learning_rate": 1.917742624525371e-05,
+      "loss": 0.4231,
+      "step": 1505
+    },
+    {
+      "epoch": 0.8713214079630698,
+      "grad_norm": 0.3612492143077761,
+      "learning_rate": 1.9167395844726763e-05,
+      "loss": 0.4454,
+      "step": 1510
+    },
+    {
+      "epoch": 0.8742065781881131,
+      "grad_norm": 0.3407789646703214,
+      "learning_rate": 1.915730731494332e-05,
+      "loss": 0.4293,
+      "step": 1515
+    },
+    {
+      "epoch": 0.8770917484131564,
+      "grad_norm": 0.33592086055664455,
+      "learning_rate": 1.914716071987342e-05,
+      "loss": 0.4358,
+      "step": 1520
+    },
+    {
+      "epoch": 0.8799769186381996,
+      "grad_norm": 0.36425558440653516,
+      "learning_rate": 1.91369561238553e-05,
+      "loss": 0.4378,
+      "step": 1525
+    },
+    {
+      "epoch": 0.8828620888632429,
+      "grad_norm": 0.3522122082887795,
+      "learning_rate": 1.912669359159496e-05,
+      "loss": 0.4295,
+      "step": 1530
+    },
+    {
+      "epoch": 0.8857472590882862,
+      "grad_norm": 0.324930415843359,
+      "learning_rate": 1.911637318816578e-05,
+      "loss": 0.4251,
+      "step": 1535
+    },
+    {
+      "epoch": 0.8886324293133295,
+      "grad_norm": 0.3466989166716833,
+      "learning_rate": 1.9105994979008083e-05,
+      "loss": 0.4328,
+      "step": 1540
+    },
+    {
+      "epoch": 0.8915175995383727,
+      "grad_norm": 0.3603001072615171,
+      "learning_rate": 1.9095559029928735e-05,
+      "loss": 0.4331,
+      "step": 1545
+    },
+    {
+      "epoch": 0.894402769763416,
+      "grad_norm": 0.35998678640671994,
+      "learning_rate": 1.9085065407100723e-05,
+      "loss": 0.4247,
+      "step": 1550
+    },
+    {
+      "epoch": 0.8972879399884593,
+      "grad_norm": 0.3353610544242477,
+      "learning_rate": 1.9074514177062734e-05,
+      "loss": 0.4365,
+      "step": 1555
+    },
+    {
+      "epoch": 0.9001731102135025,
+      "grad_norm": 0.3362819633369662,
+      "learning_rate": 1.9063905406718735e-05,
+      "loss": 0.4387,
+      "step": 1560
+    },
+    {
+      "epoch": 0.9030582804385459,
+      "grad_norm": 0.35146022780606057,
+      "learning_rate": 1.9053239163337553e-05,
+      "loss": 0.4294,
+      "step": 1565
+    },
+    {
+      "epoch": 0.9059434506635892,
+      "grad_norm": 0.3334044241812438,
+      "learning_rate": 1.9042515514552438e-05,
+      "loss": 0.4297,
+      "step": 1570
+    },
+    {
+      "epoch": 0.9088286208886325,
+      "grad_norm": 0.33641137519229225,
+      "learning_rate": 1.9031734528360642e-05,
+      "loss": 0.42,
+      "step": 1575
+    },
+    {
+      "epoch": 0.9117137911136757,
+      "grad_norm": 0.37493485136716287,
+      "learning_rate": 1.9020896273122987e-05,
+      "loss": 0.4254,
+      "step": 1580
+    },
+    {
+      "epoch": 0.914598961338719,
+      "grad_norm": 0.3466700992777568,
+      "learning_rate": 1.901000081756343e-05,
+      "loss": 0.4203,
+      "step": 1585
+    },
+    {
+      "epoch": 0.9174841315637623,
+      "grad_norm": 0.3601041654222663,
+      "learning_rate": 1.899904823076863e-05,
+      "loss": 0.4293,
+      "step": 1590
+    },
+    {
+      "epoch": 0.9203693017888055,
+      "grad_norm": 0.5717625955327809,
+      "learning_rate": 1.8988038582187506e-05,
+      "loss": 0.4331,
+      "step": 1595
+    },
+    {
+      "epoch": 0.9232544720138488,
+      "grad_norm": 0.3563731456994365,
+      "learning_rate": 1.8976971941630792e-05,
+      "loss": 0.4434,
+      "step": 1600
+    },
+    {
+      "epoch": 0.9261396422388921,
+      "grad_norm": 0.3133483522354351,
+      "learning_rate": 1.896584837927061e-05,
+      "loss": 0.4381,
+      "step": 1605
+    },
+    {
+      "epoch": 0.9290248124639354,
+      "grad_norm": 0.3602400529793251,
+      "learning_rate": 1.8954667965640013e-05,
+      "loss": 0.4295,
+      "step": 1610
+    },
+    {
+      "epoch": 0.9319099826889786,
+      "grad_norm": 0.3660387727678179,
+      "learning_rate": 1.894343077163254e-05,
+      "loss": 0.4137,
+      "step": 1615
+    },
+    {
+      "epoch": 0.9347951529140219,
+      "grad_norm": 0.33502056662496815,
+      "learning_rate": 1.8932136868501763e-05,
+      "loss": 0.431,
+      "step": 1620
+    },
+    {
+      "epoch": 0.9376803231390652,
+      "grad_norm": 0.3519580845164796,
+      "learning_rate": 1.8920786327860848e-05,
+      "loss": 0.4382,
+      "step": 1625
+    },
+    {
+      "epoch": 0.9405654933641084,
+      "grad_norm": 0.3609318959364458,
+      "learning_rate": 1.8909379221682083e-05,
+      "loss": 0.4263,
+      "step": 1630
+    },
+    {
+      "epoch": 0.9434506635891518,
+      "grad_norm": 0.3470987740772583,
+      "learning_rate": 1.8897915622296436e-05,
+      "loss": 0.436,
+      "step": 1635
+    },
+    {
+      "epoch": 0.9463358338141951,
+      "grad_norm": 0.3524435689745886,
+      "learning_rate": 1.8886395602393087e-05,
+      "loss": 0.4364,
+      "step": 1640
+    },
+    {
+      "epoch": 0.9492210040392384,
+      "grad_norm": 0.3438549232804042,
+      "learning_rate": 1.887481923501898e-05,
+      "loss": 0.4457,
+      "step": 1645
+    },
+    {
+      "epoch": 0.9521061742642816,
+      "grad_norm": 0.3359346965205838,
+      "learning_rate": 1.8863186593578336e-05,
+      "loss": 0.4489,
+      "step": 1650
+    },
+    {
+      "epoch": 0.9549913444893249,
+      "grad_norm": 0.33509833712084897,
+      "learning_rate": 1.8851497751832216e-05,
+      "loss": 0.4291,
+      "step": 1655
+    },
+    {
+      "epoch": 0.9578765147143682,
+      "grad_norm": 0.34585796301858324,
+      "learning_rate": 1.8839752783898033e-05,
+      "loss": 0.4449,
+      "step": 1660
+    },
+    {
+      "epoch": 0.9607616849394114,
+      "grad_norm": 0.38171505764397407,
+      "learning_rate": 1.8827951764249093e-05,
+      "loss": 0.4482,
+      "step": 1665
+    },
+    {
+      "epoch": 0.9636468551644547,
+      "grad_norm": 0.37008446077548557,
+      "learning_rate": 1.8816094767714113e-05,
+      "loss": 0.441,
+      "step": 1670
+    },
+    {
+      "epoch": 0.966532025389498,
+      "grad_norm": 0.36975245636292803,
+      "learning_rate": 1.880418186947676e-05,
+      "loss": 0.4411,
+      "step": 1675
+    },
+    {
+      "epoch": 0.9694171956145412,
+      "grad_norm": 0.36040770935696576,
+      "learning_rate": 1.8792213145075156e-05,
+      "loss": 0.4279,
+      "step": 1680
+    },
+    {
+      "epoch": 0.9723023658395845,
+      "grad_norm": 0.3555551748715796,
+      "learning_rate": 1.8780188670401415e-05,
+      "loss": 0.4201,
+      "step": 1685
+    },
+    {
+      "epoch": 0.9751875360646278,
+      "grad_norm": 0.34389008179144037,
+      "learning_rate": 1.876810852170116e-05,
+      "loss": 0.4659,
+      "step": 1690
+    },
+    {
+      "epoch": 0.9780727062896711,
+      "grad_norm": 0.34681456919275633,
+      "learning_rate": 1.875597277557303e-05,
+      "loss": 0.4214,
+      "step": 1695
+    },
+    {
+      "epoch": 0.9809578765147143,
+      "grad_norm": 0.33612247065522033,
+      "learning_rate": 1.8743781508968206e-05,
+      "loss": 0.4322,
+      "step": 1700
+    },
+    {
+      "epoch": 0.9838430467397576,
+      "grad_norm": 0.35053266185577925,
+      "learning_rate": 1.8731534799189905e-05,
+      "loss": 0.4338,
+      "step": 1705
+    },
+    {
+      "epoch": 0.986728216964801,
+      "grad_norm": 0.3554276718676979,
+      "learning_rate": 1.8719232723892916e-05,
+      "loss": 0.4387,
+      "step": 1710
+    },
+    {
+      "epoch": 0.9896133871898442,
+      "grad_norm": 0.3439024647253949,
+      "learning_rate": 1.8706875361083088e-05,
+      "loss": 0.4161,
+      "step": 1715
+    },
+    {
+      "epoch": 0.9924985574148875,
+      "grad_norm": 0.32640748056119073,
+      "learning_rate": 1.869446278911683e-05,
+      "loss": 0.4205,
+      "step": 1720
+    },
+    {
+      "epoch": 0.9953837276399308,
+      "grad_norm": 0.3673852819098302,
+      "learning_rate": 1.8681995086700654e-05,
+      "loss": 0.4264,
+      "step": 1725
+    },
+    {
+      "epoch": 0.9982688978649741,
+      "grad_norm": 0.3414077938012464,
+      "learning_rate": 1.866947233289061e-05,
+      "loss": 0.4295,
+      "step": 1730
+    },
+    {
+      "epoch": 1.0,
+      "eval_loss": 0.4312867522239685,
+      "eval_runtime": 0.6755,
+      "eval_samples_per_second": 113.987,
+      "eval_steps_per_second": 2.961,
+      "step": 1733
+    },
+    {
+      "epoch": 1.0011540680900173,
+      "grad_norm": 0.3719885030349497,
+      "learning_rate": 1.865689460709185e-05,
+      "loss": 0.4085,
+      "step": 1735
+    },
+    {
+      "epoch": 1.0040392383150605,
+      "grad_norm": 0.3653013563481502,
+      "learning_rate": 1.8644261989058082e-05,
+      "loss": 0.3663,
+      "step": 1740
+    },
+    {
+      "epoch": 1.006924408540104,
+      "grad_norm": 0.4249244844285388,
+      "learning_rate": 1.8631574558891087e-05,
+      "loss": 0.3636,
+      "step": 1745
+    },
+    {
+      "epoch": 1.009809578765147,
+      "grad_norm": 0.4065943346576567,
+      "learning_rate": 1.8618832397040192e-05,
+      "loss": 0.3602,
+      "step": 1750
+    },
+    {
+      "epoch": 1.0126947489901905,
+      "grad_norm": 0.3839053642256101,
+      "learning_rate": 1.860603558430178e-05,
+      "loss": 0.3793,
+      "step": 1755
+    },
+    {
+      "epoch": 1.0155799192152337,
+      "grad_norm": 0.36995484185295535,
+      "learning_rate": 1.859318420181877e-05,
+      "loss": 0.3438,
+      "step": 1760
+    },
+    {
+      "epoch": 1.018465089440277,
+      "grad_norm": 0.3848314010331738,
+      "learning_rate": 1.8580278331080085e-05,
+      "loss": 0.3859,
+      "step": 1765
+    },
+    {
+      "epoch": 1.0213502596653203,
+      "grad_norm": 0.3963349177249421,
+      "learning_rate": 1.8567318053920166e-05,
+      "loss": 0.3769,
+      "step": 1770
+    },
+    {
+      "epoch": 1.0242354298903635,
+      "grad_norm": 0.37297944843138187,
+      "learning_rate": 1.8554303452518436e-05,
+      "loss": 0.3579,
+      "step": 1775
+    },
+    {
+      "epoch": 1.0271206001154067,
+      "grad_norm": 0.3904119687309658,
+      "learning_rate": 1.854123460939877e-05,
+      "loss": 0.3715,
+      "step": 1780
+    },
+    {
+      "epoch": 1.0300057703404502,
+      "grad_norm": 0.3788697016458385,
+      "learning_rate": 1.8528111607429e-05,
+      "loss": 0.3728,
+      "step": 1785
+    },
+    {
+      "epoch": 1.0328909405654934,
+      "grad_norm": 0.37738365383900435,
+      "learning_rate": 1.851493452982036e-05,
+      "loss": 0.3611,
+      "step": 1790
+    },
+    {
+      "epoch": 1.0357761107905366,
+      "grad_norm": 0.39677380812493784,
+      "learning_rate": 1.8501703460126973e-05,
+      "loss": 0.3661,
+      "step": 1795
+    },
+    {
+      "epoch": 1.03866128101558,
+      "grad_norm": 0.39895561200913643,
+      "learning_rate": 1.8488418482245317e-05,
+      "loss": 0.3699,
+      "step": 1800
+    },
+    {
+      "epoch": 1.0415464512406232,
+      "grad_norm": 0.3962653634072917,
+      "learning_rate": 1.84750796804137e-05,
+      "loss": 0.3513,
+      "step": 1805
+    },
+    {
+      "epoch": 1.0444316214656664,
+      "grad_norm": 0.4045522991037393,
+      "learning_rate": 1.8461687139211712e-05,
+      "loss": 0.3621,
+      "step": 1810
+    },
+    {
+      "epoch": 1.0473167916907098,
+      "grad_norm": 0.37889291444959144,
+      "learning_rate": 1.844824094355971e-05,
+      "loss": 0.3629,
+      "step": 1815
+    },
+    {
+      "epoch": 1.050201961915753,
+      "grad_norm": 0.36255794337001546,
+      "learning_rate": 1.843474117871826e-05,
+      "loss": 0.356,
+      "step": 1820
+    },
+    {
+      "epoch": 1.0530871321407964,
+      "grad_norm": 0.36607303065989255,
+      "learning_rate": 1.84211879302876e-05,
+      "loss": 0.3481,
+      "step": 1825
+    },
+    {
+      "epoch": 1.0559723023658396,
+      "grad_norm": 0.38272410808497676,
+      "learning_rate": 1.84075812842071e-05,
+      "loss": 0.3509,
+      "step": 1830
+    },
+    {
+      "epoch": 1.0588574725908828,
+      "grad_norm": 0.359466522831376,
+      "learning_rate": 1.8393921326754724e-05,
+      "loss": 0.3691,
+      "step": 1835
+    },
+    {
+      "epoch": 1.0617426428159262,
+      "grad_norm": 0.3758374522906589,
+      "learning_rate": 1.8380208144546473e-05,
+      "loss": 0.3718,
+      "step": 1840
+    },
+    {
+      "epoch": 1.0646278130409694,
+      "grad_norm": 0.3762140791163445,
+      "learning_rate": 1.836644182453584e-05,
+      "loss": 0.3777,
+      "step": 1845
+    },
+    {
+      "epoch": 1.0675129832660126,
+      "grad_norm": 0.4061862324473009,
+      "learning_rate": 1.835262245401326e-05,
+      "loss": 0.3618,
+      "step": 1850
+    },
+    {
+      "epoch": 1.070398153491056,
+      "grad_norm": 0.361301220342186,
+      "learning_rate": 1.833875012060555e-05,
+      "loss": 0.3974,
+      "step": 1855
+    },
+    {
+      "epoch": 1.0732833237160992,
+      "grad_norm": 0.3769202277748975,
+      "learning_rate": 1.8324824912275355e-05,
+      "loss": 0.3495,
+      "step": 1860
+    },
+    {
+      "epoch": 1.0761684939411424,
+      "grad_norm": 0.3605517647936932,
+      "learning_rate": 1.8310846917320602e-05,
+      "loss": 0.3481,
+      "step": 1865
+    },
+    {
+      "epoch": 1.0790536641661859,
+      "grad_norm": 0.384557592260759,
+      "learning_rate": 1.8296816224373926e-05,
+      "loss": 0.3431,
+      "step": 1870
+    },
+    {
+      "epoch": 1.081938834391229,
+      "grad_norm": 0.39969729027596174,
+      "learning_rate": 1.828273292240211e-05,
+      "loss": 0.361,
+      "step": 1875
+    },
+    {
+      "epoch": 1.0848240046162723,
+      "grad_norm": 0.3893084314526289,
+      "learning_rate": 1.8268597100705534e-05,
+      "loss": 0.3533,
+      "step": 1880
+    },
+    {
+      "epoch": 1.0877091748413157,
+      "grad_norm": 0.39190402847282024,
+      "learning_rate": 1.8254408848917587e-05,
+      "loss": 0.3528,
+      "step": 1885
+    },
+    {
+      "epoch": 1.0905943450663589,
+      "grad_norm": 0.3991376675237472,
+      "learning_rate": 1.824016825700412e-05,
+      "loss": 0.3547,
+      "step": 1890
+    },
+    {
+      "epoch": 1.0934795152914023,
+      "grad_norm": 0.3647535186718755,
+      "learning_rate": 1.8225875415262852e-05,
+      "loss": 0.3752,
+      "step": 1895
+    },
+    {
+      "epoch": 1.0963646855164455,
+      "grad_norm": 0.36492225390564653,
+      "learning_rate": 1.8211530414322834e-05,
+      "loss": 0.3705,
+      "step": 1900
+    },
+    {
+      "epoch": 1.0992498557414887,
+      "grad_norm": 0.38796798500765634,
+      "learning_rate": 1.8197133345143828e-05,
+      "loss": 0.3691,
+      "step": 1905
+    },
+    {
+      "epoch": 1.1021350259665321,
+      "grad_norm": 0.3827359128376877,
+      "learning_rate": 1.8182684299015764e-05,
+      "loss": 0.3689,
+      "step": 1910
+    },
+    {
+      "epoch": 1.1050201961915753,
+      "grad_norm": 0.38752794638536536,
+      "learning_rate": 1.816818336755816e-05,
+      "loss": 0.3567,
+      "step": 1915
+    },
+    {
+      "epoch": 1.1079053664166185,
+      "grad_norm": 0.38630104069610516,
+      "learning_rate": 1.815363064271952e-05,
+      "loss": 0.3483,
+      "step": 1920
+    },
+    {
+      "epoch": 1.110790536641662,
+      "grad_norm": 0.37210418058912814,
+      "learning_rate": 1.813902621677676e-05,
+      "loss": 0.3543,
+      "step": 1925
+    },
+    {
+      "epoch": 1.1136757068667051,
+      "grad_norm": 0.37713275848642197,
+      "learning_rate": 1.812437018233464e-05,
+      "loss": 0.3531,
+      "step": 1930
+    },
+    {
+      "epoch": 1.1165608770917483,
+      "grad_norm": 0.36232403860634815,
+      "learning_rate": 1.8109662632325152e-05,
+      "loss": 0.3658,
+      "step": 1935
+    },
+    {
+      "epoch": 1.1194460473167918,
+      "grad_norm": 0.3466007749649004,
+      "learning_rate": 1.809490366000694e-05,
+      "loss": 0.3717,
+      "step": 1940
+    },
+    {
+      "epoch": 1.122331217541835,
+      "grad_norm": 0.37512115226860376,
+      "learning_rate": 1.8080093358964727e-05,
+      "loss": 0.3684,
+      "step": 1945
+    },
+    {
+      "epoch": 1.1252163877668782,
+      "grad_norm": 0.4257633969210461,
+      "learning_rate": 1.806523182310868e-05,
+      "loss": 0.365,
+      "step": 1950
+    },
+    {
+      "epoch": 1.1281015579919216,
+      "grad_norm": 0.3907894746639846,
+      "learning_rate": 1.8050319146673856e-05,
+      "loss": 0.3539,
+      "step": 1955
+    },
+    {
+      "epoch": 1.1309867282169648,
+      "grad_norm": 0.3868414285848145,
+      "learning_rate": 1.8035355424219586e-05,
+      "loss": 0.3619,
+      "step": 1960
+    },
+    {
+      "epoch": 1.133871898442008,
+      "grad_norm": 0.391737905658893,
+      "learning_rate": 1.8020340750628872e-05,
+      "loss": 0.3556,
+      "step": 1965
+    },
+    {
+      "epoch": 1.1367570686670514,
+      "grad_norm": 0.39546663065808646,
+      "learning_rate": 1.800527522110779e-05,
+      "loss": 0.3608,
+      "step": 1970
+    },
+    {
+      "epoch": 1.1396422388920946,
+      "grad_norm": 0.3982732465681356,
+      "learning_rate": 1.7990158931184892e-05,
+      "loss": 0.3553,
+      "step": 1975
+    },
+    {
+      "epoch": 1.1425274091171378,
+      "grad_norm": 0.3738040441095041,
+      "learning_rate": 1.79749919767106e-05,
+      "loss": 0.3599,
+      "step": 1980
+    },
+    {
+      "epoch": 1.1454125793421812,
+      "grad_norm": 0.36487443451311957,
+      "learning_rate": 1.7959774453856576e-05,
+      "loss": 0.3521,
+      "step": 1985
+    },
+    {
+      "epoch": 1.1482977495672244,
+      "grad_norm": 0.40333453889614557,
+      "learning_rate": 1.794450645911514e-05,
+      "loss": 0.3651,
+      "step": 1990
+    },
+    {
+      "epoch": 1.1511829197922678,
+      "grad_norm": 0.5753350465511803,
+      "learning_rate": 1.7929188089298653e-05,
+      "loss": 0.369,
+      "step": 1995
+    },
+    {
+      "epoch": 1.154068090017311,
+      "grad_norm": 0.3727451234003791,
+      "learning_rate": 1.791381944153889e-05,
+      "loss": 0.3536,
+      "step": 2000
+    },
+    {
+      "epoch": 1.1569532602423542,
+      "grad_norm": 0.3787668346391338,
+      "learning_rate": 1.789840061328643e-05,
+      "loss": 0.3765,
+      "step": 2005
+    },
+    {
+      "epoch": 1.1598384304673977,
+      "grad_norm": 0.36284152770529254,
+      "learning_rate": 1.788293170231005e-05,
+      "loss": 0.3455,
+      "step": 2010
+    },
+    {
+      "epoch": 1.1627236006924409,
+      "grad_norm": 0.38508608974804587,
+      "learning_rate": 1.7867412806696087e-05,
+      "loss": 0.3716,
+      "step": 2015
+    },
+    {
+      "epoch": 1.165608770917484,
+      "grad_norm": 0.3959297634423943,
+      "learning_rate": 1.785184402484782e-05,
+      "loss": 0.3715,
+      "step": 2020
+    },
+    {
+      "epoch": 1.1684939411425275,
+      "grad_norm": 0.4224675659582224,
+      "learning_rate": 1.7836225455484865e-05,
+      "loss": 0.3844,
+      "step": 2025
+    },
+    {
+      "epoch": 1.1713791113675707,
+      "grad_norm": 0.36797685694286536,
+      "learning_rate": 1.7820557197642513e-05,
+      "loss": 0.3691,
+      "step": 2030
+    },
+    {
+      "epoch": 1.1742642815926139,
+      "grad_norm": 0.3831267176121955,
+      "learning_rate": 1.7804839350671145e-05,
+      "loss": 0.3571,
+      "step": 2035
+    },
+    {
+      "epoch": 1.1771494518176573,
+      "grad_norm": 0.3905138892178264,
+      "learning_rate": 1.778907201423557e-05,
+      "loss": 0.3636,
+      "step": 2040
+    },
+    {
+      "epoch": 1.1800346220427005,
+      "grad_norm": 0.3912936170187795,
+      "learning_rate": 1.7773255288314398e-05,
+      "loss": 0.3574,
+      "step": 2045
+    },
+    {
+      "epoch": 1.182919792267744,
+      "grad_norm": 0.3754522878062524,
+      "learning_rate": 1.775738927319942e-05,
+      "loss": 0.3874,
+      "step": 2050
+    },
+    {
+      "epoch": 1.185804962492787,
+      "grad_norm": 0.36965375437322023,
+      "learning_rate": 1.774147406949496e-05,
+      "loss": 0.3717,
+      "step": 2055
+    },
+    {
+      "epoch": 1.1886901327178303,
+      "grad_norm": 0.4220256804649789,
+      "learning_rate": 1.7725509778117242e-05,
+      "loss": 0.3541,
+      "step": 2060
+    },
+    {
+      "epoch": 1.1915753029428737,
+      "grad_norm": 0.38295286541947393,
+      "learning_rate": 1.770949650029374e-05,
+      "loss": 0.3574,
+      "step": 2065
+    },
+    {
+      "epoch": 1.194460473167917,
+      "grad_norm": 0.37402840183729646,
+      "learning_rate": 1.7693434337562557e-05,
+      "loss": 0.3738,
+      "step": 2070
+    },
+    {
+      "epoch": 1.1973456433929601,
+      "grad_norm": 0.385808424420751,
+      "learning_rate": 1.7677323391771766e-05,
+      "loss": 0.3646,
+      "step": 2075
+    },
+    {
+      "epoch": 1.2002308136180035,
+      "grad_norm": 0.37207777332282577,
+      "learning_rate": 1.766116376507876e-05,
+      "loss": 0.3699,
+      "step": 2080
+    },
+    {
+      "epoch": 1.2031159838430467,
+      "grad_norm": 0.39725920392186914,
+      "learning_rate": 1.764495555994962e-05,
+      "loss": 0.3736,
+      "step": 2085
+    },
+    {
+      "epoch": 1.20600115406809,
+      "grad_norm": 0.3885483852856526,
+      "learning_rate": 1.7628698879158444e-05,
+      "loss": 0.3493,
+      "step": 2090
+    },
+    {
+      "epoch": 1.2088863242931334,
+      "grad_norm": 0.3779946551242946,
+      "learning_rate": 1.7612393825786723e-05,
+      "loss": 0.3758,
+      "step": 2095
+    },
+    {
+      "epoch": 1.2117714945181766,
+      "grad_norm": 0.3625482575463983,
+      "learning_rate": 1.7596040503222665e-05,
+      "loss": 0.3745,
+      "step": 2100
+    },
+    {
+      "epoch": 1.2146566647432198,
+      "grad_norm": 0.38330347138439946,
+      "learning_rate": 1.7579639015160545e-05,
+      "loss": 0.362,
+      "step": 2105
+    },
+    {
+      "epoch": 1.2175418349682632,
+      "grad_norm": 0.4177686333809965,
+      "learning_rate": 1.7563189465600047e-05,
+      "loss": 0.3699,
+      "step": 2110
+    },
+    {
+      "epoch": 1.2204270051933064,
+      "grad_norm": 0.3707075705845332,
+      "learning_rate": 1.7546691958845617e-05,
+      "loss": 0.3485,
+      "step": 2115
+    },
+    {
+      "epoch": 1.2233121754183496,
+      "grad_norm": 0.37675133161938845,
+      "learning_rate": 1.7530146599505782e-05,
+      "loss": 0.3738,
+      "step": 2120
+    },
+    {
+      "epoch": 1.226197345643393,
+      "grad_norm": 0.4073716130331381,
+      "learning_rate": 1.75135534924925e-05,
+      "loss": 0.3726,
+      "step": 2125
+    },
+    {
+      "epoch": 1.2290825158684362,
+      "grad_norm": 0.377632182741206,
+      "learning_rate": 1.7496912743020487e-05,
+      "loss": 0.3606,
+      "step": 2130
+    },
+    {
+      "epoch": 1.2319676860934794,
+      "grad_norm": 0.41027591746654574,
+      "learning_rate": 1.748022445660656e-05,
+      "loss": 0.3677,
+      "step": 2135
+    },
+    {
+      "epoch": 1.2348528563185228,
+      "grad_norm": 0.454127691618852,
+      "learning_rate": 1.7463488739068952e-05,
+      "loss": 0.356,
+      "step": 2140
+    },
+    {
+      "epoch": 1.237738026543566,
+      "grad_norm": 0.38902494056071213,
+      "learning_rate": 1.744670569652666e-05,
+      "loss": 0.356,
+      "step": 2145
+    },
+    {
+      "epoch": 1.2406231967686094,
+      "grad_norm": 0.3547365448603461,
+      "learning_rate": 1.742987543539876e-05,
+      "loss": 0.3717,
+      "step": 2150
+    },
+    {
+      "epoch": 1.2435083669936526,
+      "grad_norm": 0.37046329915555815,
+      "learning_rate": 1.741299806240373e-05,
+      "loss": 0.3675,
+      "step": 2155
+    },
+    {
+      "epoch": 1.2463935372186958,
+      "grad_norm": 0.3869703816603672,
+      "learning_rate": 1.7396073684558788e-05,
+      "loss": 0.3665,
+      "step": 2160
+    },
+    {
+      "epoch": 1.2492787074437393,
+      "grad_norm": 0.38188698969658114,
+      "learning_rate": 1.737910240917919e-05,
+      "loss": 0.3624,
+      "step": 2165
+    },
+    {
+      "epoch": 1.2521638776687825,
+      "grad_norm": 0.37535923832323237,
+      "learning_rate": 1.7362084343877576e-05,
+      "loss": 0.368,
+      "step": 2170
+    },
+    {
+      "epoch": 1.2550490478938257,
+      "grad_norm": 0.36714065996796413,
+      "learning_rate": 1.7345019596563268e-05,
+      "loss": 0.3667,
+      "step": 2175
+    },
+    {
+      "epoch": 1.257934218118869,
+      "grad_norm": 0.4157188737740098,
+      "learning_rate": 1.732790827544159e-05,
+      "loss": 0.3737,
+      "step": 2180
+    },
+    {
+      "epoch": 1.2608193883439123,
+      "grad_norm": 0.3811681622112457,
+      "learning_rate": 1.7310750489013198e-05,
+      "loss": 0.3544,
+      "step": 2185
+    },
+    {
+      "epoch": 1.2637045585689557,
+      "grad_norm": 0.3753050362248373,
+      "learning_rate": 1.729354634607336e-05,
+      "loss": 0.3664,
+      "step": 2190
+    },
+    {
+      "epoch": 1.266589728793999,
+      "grad_norm": 0.36855745458054423,
+      "learning_rate": 1.7276295955711295e-05,
+      "loss": 0.3637,
+      "step": 2195
+    },
+    {
+      "epoch": 1.269474899019042,
+      "grad_norm": 0.3650539628680322,
+      "learning_rate": 1.725899942730947e-05,
+      "loss": 0.3645,
+      "step": 2200
+    },
+    {
+      "epoch": 1.2723600692440855,
+      "grad_norm": 0.3831096440319005,
+      "learning_rate": 1.724165687054291e-05,
+      "loss": 0.3618,
+      "step": 2205
+    },
+    {
+      "epoch": 1.2752452394691287,
+      "grad_norm": 0.3603410399079337,
+      "learning_rate": 1.7224268395378493e-05,
+      "loss": 0.3453,
+      "step": 2210
+    },
+    {
+      "epoch": 1.278130409694172,
+      "grad_norm": 0.37269959302348715,
+      "learning_rate": 1.720683411207426e-05,
+      "loss": 0.3576,
+      "step": 2215
+    },
+    {
+      "epoch": 1.2810155799192153,
+      "grad_norm": 0.38173221937480023,
+      "learning_rate": 1.718935413117872e-05,
+      "loss": 0.371,
+      "step": 2220
+    },
+    {
+      "epoch": 1.2839007501442585,
+      "grad_norm": 0.3722436971668028,
+      "learning_rate": 1.7171828563530146e-05,
+      "loss": 0.3663,
+      "step": 2225
+    },
+    {
+      "epoch": 1.2867859203693017,
+      "grad_norm": 0.39837683099640625,
+      "learning_rate": 1.7154257520255864e-05,
+      "loss": 0.3745,
+      "step": 2230
+    },
+    {
+      "epoch": 1.2896710905943451,
+      "grad_norm": 0.41009144762310795,
+      "learning_rate": 1.713664111277156e-05,
+      "loss": 0.3666,
+      "step": 2235
+    },
+    {
+      "epoch": 1.2925562608193883,
+      "grad_norm": 0.36668782585448506,
+      "learning_rate": 1.711897945278057e-05,
+      "loss": 0.3738,
+      "step": 2240
+    },
+    {
+      "epoch": 1.2954414310444315,
+      "grad_norm": 0.41098358203090174,
+      "learning_rate": 1.710127265227316e-05,
+      "loss": 0.3668,
+      "step": 2245
+    },
+    {
+      "epoch": 1.298326601269475,
+      "grad_norm": 0.3862089819354528,
+      "learning_rate": 1.7083520823525836e-05,
+      "loss": 0.3661,
+      "step": 2250
+    },
+    {
+      "epoch": 1.3012117714945182,
+      "grad_norm": 0.4033977142486018,
+      "learning_rate": 1.706572407910062e-05,
+      "loss": 0.3688,
+      "step": 2255
+    },
+    {
+      "epoch": 1.3040969417195614,
+      "grad_norm": 0.4003446173157873,
+      "learning_rate": 1.7047882531844338e-05,
+      "loss": 0.377,
+      "step": 2260
+    },
+    {
+      "epoch": 1.3069821119446048,
+      "grad_norm": 0.3702354075920004,
+      "learning_rate": 1.7029996294887904e-05,
+      "loss": 0.3774,
+      "step": 2265
+    },
+    {
+      "epoch": 1.309867282169648,
+      "grad_norm": 0.3915858903728003,
+      "learning_rate": 1.7012065481645604e-05,
+      "loss": 0.3671,
+      "step": 2270
+    },
+    {
+      "epoch": 1.3127524523946912,
+      "grad_norm": 0.3721772452216554,
+      "learning_rate": 1.699409020581438e-05,
+      "loss": 0.3553,
+      "step": 2275
+    },
+    {
+      "epoch": 1.3156376226197346,
+      "grad_norm": 0.3726078669899795,
+      "learning_rate": 1.6976070581373098e-05,
+      "loss": 0.3781,
+      "step": 2280
+    },
+    {
+      "epoch": 1.3185227928447778,
+      "grad_norm": 0.3851091572746381,
+      "learning_rate": 1.6958006722581838e-05,
+      "loss": 0.3764,
+      "step": 2285
+    },
+    {
+      "epoch": 1.321407963069821,
+      "grad_norm": 0.3999997138224476,
+      "learning_rate": 1.693989874398116e-05,
+      "loss": 0.3719,
+      "step": 2290
+    },
+    {
+      "epoch": 1.3242931332948644,
+      "grad_norm": 0.36767006912934236,
+      "learning_rate": 1.6921746760391387e-05,
+      "loss": 0.3748,
+      "step": 2295
+    },
+    {
+      "epoch": 1.3271783035199076,
+      "grad_norm": 0.3730179310503956,
+      "learning_rate": 1.6903550886911873e-05,
+      "loss": 0.3608,
+      "step": 2300
+    },
+    {
+      "epoch": 1.3300634737449508,
+      "grad_norm": 0.3672274323141715,
+      "learning_rate": 1.688531123892027e-05,
+      "loss": 0.3443,
+      "step": 2305
+    },
+    {
+      "epoch": 1.3329486439699942,
+      "grad_norm": 0.36049125372252444,
+      "learning_rate": 1.6867027932071786e-05,
+      "loss": 0.3606,
+      "step": 2310
+    },
+    {
+      "epoch": 1.3358338141950374,
+      "grad_norm": 0.37209341340715907,
+      "learning_rate": 1.684870108229849e-05,
+      "loss": 0.3866,
+      "step": 2315
+    },
+    {
+      "epoch": 1.3387189844200809,
+      "grad_norm": 0.4041341082410308,
+      "learning_rate": 1.6830330805808524e-05,
+      "loss": 0.3552,
+      "step": 2320
+    },
+    {
+      "epoch": 1.341604154645124,
+      "grad_norm": 0.44070233485279253,
+      "learning_rate": 1.681191721908541e-05,
+      "loss": 0.3641,
+      "step": 2325
+    },
+    {
+      "epoch": 1.3444893248701673,
+      "grad_norm": 0.3782903985125648,
+      "learning_rate": 1.6793460438887294e-05,
+      "loss": 0.3702,
+      "step": 2330
+    },
+    {
+      "epoch": 1.3473744950952107,
+      "grad_norm": 0.38093409058463656,
+      "learning_rate": 1.67749605822462e-05,
+      "loss": 0.3738,
+      "step": 2335
+    },
+    {
+      "epoch": 1.3502596653202539,
+      "grad_norm": 0.4076843344147931,
+      "learning_rate": 1.6756417766467294e-05,
+      "loss": 0.3713,
+      "step": 2340
+    },
+    {
+      "epoch": 1.3531448355452973,
+      "grad_norm": 0.4114981370603978,
+      "learning_rate": 1.6737832109128142e-05,
+      "loss": 0.3626,
+      "step": 2345
+    },
+    {
+      "epoch": 1.3560300057703405,
+      "grad_norm": 0.4070630768126986,
+      "learning_rate": 1.6719203728077963e-05,
+      "loss": 0.3579,
+      "step": 2350
+    },
+    {
+      "epoch": 1.3589151759953837,
+      "grad_norm": 0.3980167515551806,
+      "learning_rate": 1.670053274143689e-05,
+      "loss": 0.3688,
+      "step": 2355
+    },
+    {
+      "epoch": 1.3618003462204271,
+      "grad_norm": 0.4117026797519244,
+      "learning_rate": 1.6681819267595193e-05,
+      "loss": 0.3631,
+      "step": 2360
+    },
+    {
+      "epoch": 1.3646855164454703,
+      "grad_norm": 0.3794143818195765,
+      "learning_rate": 1.6663063425212564e-05,
+      "loss": 0.3658,
+      "step": 2365
+    },
+    {
+      "epoch": 1.3675706866705135,
+      "grad_norm": 0.4389143548454321,
+      "learning_rate": 1.6644265333217347e-05,
+      "loss": 0.3622,
+      "step": 2370
+    },
+    {
+      "epoch": 1.370455856895557,
+      "grad_norm": 0.34975549009666657,
+      "learning_rate": 1.6625425110805774e-05,
+      "loss": 0.3636,
+      "step": 2375
+    },
+    {
+      "epoch": 1.3733410271206001,
+      "grad_norm": 0.4000657679860908,
+      "learning_rate": 1.660654287744123e-05,
+      "loss": 0.3655,
+      "step": 2380
+    },
+    {
+      "epoch": 1.3762261973456433,
+      "grad_norm": 0.39327391261319994,
+      "learning_rate": 1.6587618752853484e-05,
+      "loss": 0.3673,
+      "step": 2385
+    },
+    {
+      "epoch": 1.3791113675706868,
+      "grad_norm": 0.3803854498603796,
+      "learning_rate": 1.656865285703794e-05,
+      "loss": 0.3644,
+      "step": 2390
+    },
+    {
+      "epoch": 1.38199653779573,
+      "grad_norm": 0.36271341863389506,
+      "learning_rate": 1.6549645310254848e-05,
+      "loss": 0.3597,
+      "step": 2395
+    },
+    {
+      "epoch": 1.3848817080207732,
+      "grad_norm": 0.39017089645929526,
+      "learning_rate": 1.653059623302858e-05,
+      "loss": 0.3584,
+      "step": 2400
+    },
+    {
+      "epoch": 1.3877668782458166,
+      "grad_norm": 0.377459677055801,
+      "learning_rate": 1.6511505746146837e-05,
+      "loss": 0.3603,
+      "step": 2405
+    },
+    {
+      "epoch": 1.3906520484708598,
+      "grad_norm": 0.36253902214360645,
+      "learning_rate": 1.6492373970659897e-05,
+      "loss": 0.3622,
+      "step": 2410
+    },
+    {
+      "epoch": 1.393537218695903,
+      "grad_norm": 0.3792755236074643,
+      "learning_rate": 1.647320102787984e-05,
+      "loss": 0.3579,
+      "step": 2415
+    },
+    {
+      "epoch": 1.3964223889209464,
+      "grad_norm": 0.4106370547161479,
+      "learning_rate": 1.645398703937979e-05,
+      "loss": 0.3548,
+      "step": 2420
+    },
+    {
+      "epoch": 1.3993075591459896,
+      "grad_norm": 0.3961446879313167,
+      "learning_rate": 1.6434732126993137e-05,
+      "loss": 0.3788,
+      "step": 2425
+    },
+    {
+      "epoch": 1.4021927293710328,
+      "grad_norm": 0.41444403209474506,
+      "learning_rate": 1.6415436412812742e-05,
+      "loss": 0.3935,
+      "step": 2430
+    },
+    {
+      "epoch": 1.4050778995960762,
+      "grad_norm": 0.45443169110667836,
+      "learning_rate": 1.6396100019190216e-05,
+      "loss": 0.3672,
+      "step": 2435
+    },
+    {
+      "epoch": 1.4079630698211194,
+      "grad_norm": 0.3704577288597062,
+      "learning_rate": 1.6376723068735096e-05,
+      "loss": 0.3614,
+      "step": 2440
+    },
+    {
+      "epoch": 1.4108482400461626,
+      "grad_norm": 0.37885398698326,
+      "learning_rate": 1.6357305684314083e-05,
+      "loss": 0.3473,
+      "step": 2445
+    },
+    {
+      "epoch": 1.413733410271206,
+      "grad_norm": 0.41962586363127974,
+      "learning_rate": 1.6337847989050276e-05,
+      "loss": 0.3658,
+      "step": 2450
+    },
+    {
+      "epoch": 1.4166185804962492,
+      "grad_norm": 0.3851208552370083,
+      "learning_rate": 1.6318350106322363e-05,
+      "loss": 0.3699,
+      "step": 2455
+    },
+    {
+      "epoch": 1.4195037507212924,
+      "grad_norm": 0.3548157462238281,
+      "learning_rate": 1.6298812159763868e-05,
+      "loss": 0.3681,
+      "step": 2460
+    },
+    {
+      "epoch": 1.4223889209463358,
+      "grad_norm": 0.37651463846589284,
+      "learning_rate": 1.6279234273262357e-05,
+      "loss": 0.3516,
+      "step": 2465
+    },
+    {
+      "epoch": 1.425274091171379,
+      "grad_norm": 0.35900318539287046,
+      "learning_rate": 1.6259616570958638e-05,
+      "loss": 0.3704,
+      "step": 2470
+    },
+    {
+      "epoch": 1.4281592613964225,
+      "grad_norm": 0.4024381546192383,
+      "learning_rate": 1.6239959177246e-05,
+      "loss": 0.3638,
+      "step": 2475
+    },
+    {
+      "epoch": 1.4310444316214657,
+      "grad_norm": 0.398944280818137,
+      "learning_rate": 1.6220262216769393e-05,
+      "loss": 0.3813,
+      "step": 2480
+    },
+    {
+      "epoch": 1.4339296018465089,
+      "grad_norm": 0.3831395961609494,
+      "learning_rate": 1.6200525814424666e-05,
+      "loss": 0.3693,
+      "step": 2485
+    },
+    {
+      "epoch": 1.4368147720715523,
+      "grad_norm": 0.3661492085432,
+      "learning_rate": 1.6180750095357766e-05,
+      "loss": 0.3776,
+      "step": 2490
+    },
+    {
+      "epoch": 1.4396999422965955,
+      "grad_norm": 0.36586438881277866,
+      "learning_rate": 1.6160935184963937e-05,
+      "loss": 0.3503,
+      "step": 2495
+    },
+    {
+      "epoch": 1.442585112521639,
+      "grad_norm": 0.3784608295775924,
+      "learning_rate": 1.614108120888693e-05,
+      "loss": 0.3715,
+      "step": 2500
+    },
+    {
+      "epoch": 1.445470282746682,
+      "grad_norm": 0.364821310782025,
+      "learning_rate": 1.6121188293018198e-05,
+      "loss": 0.3774,
+      "step": 2505
+    },
+    {
+      "epoch": 1.4483554529717253,
+      "grad_norm": 0.38118287581212157,
+      "learning_rate": 1.6101256563496133e-05,
+      "loss": 0.3608,
+      "step": 2510
+    },
+    {
+      "epoch": 1.4512406231967687,
+      "grad_norm": 0.37900358798651623,
+      "learning_rate": 1.608128614670521e-05,
+      "loss": 0.3889,
+      "step": 2515
+    },
+    {
+      "epoch": 1.454125793421812,
+      "grad_norm": 0.40251780895398115,
+      "learning_rate": 1.6061277169275227e-05,
+      "loss": 0.3606,
+      "step": 2520
+    },
+    {
+      "epoch": 1.4570109636468551,
+      "grad_norm": 0.38327828759400334,
+      "learning_rate": 1.6041229758080496e-05,
+      "loss": 0.3635,
+      "step": 2525
+    },
+    {
+      "epoch": 1.4598961338718985,
+      "grad_norm": 0.37246590219746933,
+      "learning_rate": 1.6021144040239017e-05,
+      "loss": 0.3766,
+      "step": 2530
+    },
+    {
+      "epoch": 1.4627813040969417,
+      "grad_norm": 0.3833944550693614,
+      "learning_rate": 1.60010201431117e-05,
+      "loss": 0.3612,
+      "step": 2535
+    },
+    {
+      "epoch": 1.465666474321985,
+      "grad_norm": 0.36363911288467893,
+      "learning_rate": 1.598085819430155e-05,
+      "loss": 0.3711,
+      "step": 2540
+    },
+    {
+      "epoch": 1.4685516445470284,
+      "grad_norm": 0.3588573344819394,
+      "learning_rate": 1.596065832165283e-05,
+      "loss": 0.3726,
+      "step": 2545
+    },
+    {
+      "epoch": 1.4714368147720716,
+      "grad_norm": 0.3705849995740623,
+      "learning_rate": 1.5940420653250293e-05,
+      "loss": 0.3642,
+      "step": 2550
+    },
+    {
+      "epoch": 1.4743219849971148,
+      "grad_norm": 0.38341287260652185,
+      "learning_rate": 1.5920145317418346e-05,
+      "loss": 0.3714,
+      "step": 2555
+    },
+    {
+      "epoch": 1.4772071552221582,
+      "grad_norm": 0.38734703763461453,
+      "learning_rate": 1.589983244272024e-05,
+      "loss": 0.3836,
+      "step": 2560
+    },
+    {
+      "epoch": 1.4800923254472014,
+      "grad_norm": 0.36011456145391224,
+      "learning_rate": 1.5879482157957245e-05,
+      "loss": 0.3669,
+      "step": 2565
+    },
+    {
+      "epoch": 1.4829774956722446,
+      "grad_norm": 0.3784776076526531,
+      "learning_rate": 1.5859094592167858e-05,
+      "loss": 0.3634,
+      "step": 2570
+    },
+    {
+      "epoch": 1.485862665897288,
+      "grad_norm": 0.38334066789201254,
+      "learning_rate": 1.5838669874626963e-05,
+      "loss": 0.3836,
+      "step": 2575
+    },
+    {
+      "epoch": 1.4887478361223312,
+      "grad_norm": 0.3769831368967705,
+      "learning_rate": 1.5818208134845022e-05,
+      "loss": 0.3592,
+      "step": 2580
+    },
+    {
+      "epoch": 1.4916330063473744,
+      "grad_norm": 0.411783030147672,
+      "learning_rate": 1.579770950256724e-05,
+      "loss": 0.3794,
+      "step": 2585
+    },
+    {
+      "epoch": 1.4945181765724178,
+      "grad_norm": 0.5027841057723603,
+      "learning_rate": 1.577717410777276e-05,
+      "loss": 0.3601,
+      "step": 2590
+    },
+    {
+      "epoch": 1.497403346797461,
+      "grad_norm": 0.38774582541265035,
+      "learning_rate": 1.5756602080673843e-05,
+      "loss": 0.3753,
+      "step": 2595
+    },
+    {
+      "epoch": 1.5002885170225042,
+      "grad_norm": 0.37739115096678616,
+      "learning_rate": 1.5735993551715006e-05,
+      "loss": 0.3737,
+      "step": 2600
+    },
+    {
+      "epoch": 1.5031736872475476,
+      "grad_norm": 0.3902916999812962,
+      "learning_rate": 1.571534865157224e-05,
+      "loss": 0.3632,
+      "step": 2605
+    },
+    {
+      "epoch": 1.5060588574725908,
+      "grad_norm": 0.3868041357607236,
+      "learning_rate": 1.5694667511152143e-05,
+      "loss": 0.3809,
+      "step": 2610
+    },
+    {
+      "epoch": 1.508944027697634,
+      "grad_norm": 0.37704578010786705,
+      "learning_rate": 1.5673950261591133e-05,
+      "loss": 0.3694,
+      "step": 2615
+    },
+    {
+      "epoch": 1.5118291979226774,
+      "grad_norm": 0.3607652817962323,
+      "learning_rate": 1.5653197034254564e-05,
+      "loss": 0.3519,
+      "step": 2620
+    },
+    {
+      "epoch": 1.5147143681477209,
+      "grad_norm": 0.37320845213700304,
+      "learning_rate": 1.563240796073594e-05,
+      "loss": 0.3606,
+      "step": 2625
+    },
+    {
+      "epoch": 1.5175995383727638,
+      "grad_norm": 0.43098838699804737,
+      "learning_rate": 1.5611583172856045e-05,
+      "loss": 0.3741,
+      "step": 2630
+    },
+    {
+      "epoch": 1.5204847085978073,
+      "grad_norm": 0.3812237058754953,
+      "learning_rate": 1.5590722802662143e-05,
+      "loss": 0.3486,
+      "step": 2635
+    },
+    {
+      "epoch": 1.5233698788228507,
+      "grad_norm": 0.3799133932625459,
+      "learning_rate": 1.55698269824271e-05,
+      "loss": 0.3657,
+      "step": 2640
+    },
+    {
+      "epoch": 1.5262550490478937,
+      "grad_norm": 0.34887334582855944,
+      "learning_rate": 1.5548895844648583e-05,
+      "loss": 0.3656,
+      "step": 2645
+    },
+    {
+      "epoch": 1.529140219272937,
+      "grad_norm": 0.3798697034141352,
+      "learning_rate": 1.552792952204819e-05,
+      "loss": 0.3714,
+      "step": 2650
+    },
+    {
+      "epoch": 1.5320253894979805,
+      "grad_norm": 0.3597061950949025,
+      "learning_rate": 1.550692814757063e-05,
+      "loss": 0.3651,
+      "step": 2655
+    },
+    {
+      "epoch": 1.5349105597230237,
+      "grad_norm": 0.39911942641093723,
+      "learning_rate": 1.5485891854382868e-05,
+      "loss": 0.3662,
+      "step": 2660
+    },
+    {
+      "epoch": 1.537795729948067,
+      "grad_norm": 0.4048874005019976,
+      "learning_rate": 1.5464820775873285e-05,
+      "loss": 0.378,
+      "step": 2665
+    },
+    {
+      "epoch": 1.5406809001731103,
+      "grad_norm": 0.3614850352345454,
+      "learning_rate": 1.544371504565082e-05,
+      "loss": 0.3511,
+      "step": 2670
+    },
+    {
+      "epoch": 1.5435660703981535,
+      "grad_norm": 0.3852211514058087,
+      "learning_rate": 1.5422574797544166e-05,
+      "loss": 0.3571,
+      "step": 2675
+    },
+    {
+      "epoch": 1.5464512406231967,
+      "grad_norm": 0.36466553350789727,
+      "learning_rate": 1.5401400165600848e-05,
+      "loss": 0.3654,
+      "step": 2680
+    },
+    {
+      "epoch": 1.5493364108482401,
+      "grad_norm": 0.4640557855511497,
+      "learning_rate": 1.5380191284086442e-05,
+      "loss": 0.3663,
+      "step": 2685
+    },
+    {
+      "epoch": 1.5522215810732833,
+      "grad_norm": 0.40382378120892326,
+      "learning_rate": 1.5358948287483688e-05,
+      "loss": 0.395,
+      "step": 2690
+    },
+    {
+      "epoch": 1.5551067512983265,
+      "grad_norm": 0.34945555100262243,
+      "learning_rate": 1.533767131049164e-05,
+      "loss": 0.3507,
+      "step": 2695
+    },
+    {
+      "epoch": 1.55799192152337,
+      "grad_norm": 0.4017093497659257,
+      "learning_rate": 1.5316360488024838e-05,
+      "loss": 0.3666,
+      "step": 2700
+    },
+    {
+      "epoch": 1.5608770917484132,
+      "grad_norm": 0.36824354949459775,
+      "learning_rate": 1.5295015955212397e-05,
+      "loss": 0.3629,
+      "step": 2705
+    },
+    {
+      "epoch": 1.5637622619734564,
+      "grad_norm": 0.37118819752646043,
+      "learning_rate": 1.5273637847397216e-05,
+      "loss": 0.3619,
+      "step": 2710
+    },
+    {
+      "epoch": 1.5666474321984998,
+      "grad_norm": 0.380235084250298,
+      "learning_rate": 1.5252226300135075e-05,
+      "loss": 0.3627,
+      "step": 2715
+    },
+    {
+      "epoch": 1.569532602423543,
+      "grad_norm": 0.35479734123182555,
+      "learning_rate": 1.5230781449193787e-05,
+      "loss": 0.3499,
+      "step": 2720
+    },
+    {
+      "epoch": 1.5724177726485862,
+      "grad_norm": 0.37537141405956453,
+      "learning_rate": 1.5209303430552353e-05,
+      "loss": 0.3689,
+      "step": 2725
+    },
+    {
+      "epoch": 1.5753029428736296,
+      "grad_norm": 0.38314418131207195,
+      "learning_rate": 1.5187792380400066e-05,
+      "loss": 0.3568,
+      "step": 2730
+    },
+    {
+      "epoch": 1.5781881130986728,
+      "grad_norm": 0.4015489085451092,
+      "learning_rate": 1.5166248435135687e-05,
+      "loss": 0.3584,
+      "step": 2735
+    },
+    {
+      "epoch": 1.581073283323716,
+      "grad_norm": 0.413370518767082,
+      "learning_rate": 1.5144671731366537e-05,
+      "loss": 0.3627,
+      "step": 2740
+    },
+    {
+      "epoch": 1.5839584535487594,
+      "grad_norm": 0.3855856313069642,
+      "learning_rate": 1.5123062405907689e-05,
+      "loss": 0.3576,
+      "step": 2745
+    },
+    {
+      "epoch": 1.5868436237738026,
+      "grad_norm": 0.40248976719527124,
+      "learning_rate": 1.510142059578103e-05,
+      "loss": 0.371,
+      "step": 2750
+    },
+    {
+      "epoch": 1.5897287939988458,
+      "grad_norm": 0.3896121946823667,
+      "learning_rate": 1.5079746438214452e-05,
+      "loss": 0.3569,
+      "step": 2755
+    },
+    {
+      "epoch": 1.5926139642238892,
+      "grad_norm": 0.3601348358274979,
+      "learning_rate": 1.5058040070640947e-05,
+      "loss": 0.3689,
+      "step": 2760
+    },
+    {
+      "epoch": 1.5954991344489324,
+      "grad_norm": 0.35986769610870906,
+      "learning_rate": 1.5036301630697747e-05,
+      "loss": 0.3555,
+      "step": 2765
+    },
+    {
+      "epoch": 1.5983843046739756,
+      "grad_norm": 0.36475176711784657,
+      "learning_rate": 1.5014531256225459e-05,
+      "loss": 0.3605,
+      "step": 2770
+    },
+    {
+      "epoch": 1.601269474899019,
+      "grad_norm": 0.37930345881140104,
+      "learning_rate": 1.499272908526717e-05,
+      "loss": 0.3728,
+      "step": 2775
+    },
+    {
+      "epoch": 1.6041546451240625,
+      "grad_norm": 0.38456807910427965,
+      "learning_rate": 1.4970895256067593e-05,
+      "loss": 0.3736,
+      "step": 2780
+    },
+    {
+      "epoch": 1.6070398153491054,
+      "grad_norm": 0.4121207010530726,
+      "learning_rate": 1.4949029907072179e-05,
+      "loss": 0.3855,
+      "step": 2785
+    },
+    {
+      "epoch": 1.6099249855741489,
+      "grad_norm": 0.36551396208539605,
+      "learning_rate": 1.4927133176926245e-05,
+      "loss": 0.3662,
+      "step": 2790
+    },
+    {
+      "epoch": 1.6128101557991923,
+      "grad_norm": 0.3716986354984255,
+      "learning_rate": 1.4905205204474088e-05,
+      "loss": 0.3782,
+      "step": 2795
+    },
+    {
+      "epoch": 1.6156953260242353,
+      "grad_norm": 0.6240855472300652,
+      "learning_rate": 1.488324612875811e-05,
+      "loss": 0.3464,
+      "step": 2800
+    },
+    {
+      "epoch": 1.6185804962492787,
+      "grad_norm": 0.33838650586561203,
+      "learning_rate": 1.4861256089017934e-05,
+      "loss": 0.3571,
+      "step": 2805
+    },
+    {
+      "epoch": 1.621465666474322,
+      "grad_norm": 0.3772401322980198,
+      "learning_rate": 1.483923522468952e-05,
+      "loss": 0.3582,
+      "step": 2810
+    },
+    {
+      "epoch": 1.6243508366993653,
+      "grad_norm": 0.3686319357750962,
+      "learning_rate": 1.4817183675404292e-05,
+      "loss": 0.3705,
+      "step": 2815
+    },
+    {
+      "epoch": 1.6272360069244085,
+      "grad_norm": 0.3968174151919166,
+      "learning_rate": 1.4795101580988228e-05,
+      "loss": 0.3806,
+      "step": 2820
+    },
+    {
+      "epoch": 1.630121177149452,
+      "grad_norm": 0.3902887780136523,
+      "learning_rate": 1.4772989081460997e-05,
+      "loss": 0.3626,
+      "step": 2825
+    },
+    {
+      "epoch": 1.6330063473744951,
+      "grad_norm": 0.3746341856550175,
+      "learning_rate": 1.4750846317035071e-05,
+      "loss": 0.3672,
+      "step": 2830
+    },
+    {
+      "epoch": 1.6358915175995383,
+      "grad_norm": 0.3994993753443447,
+      "learning_rate": 1.4728673428114808e-05,
+      "loss": 0.3688,
+      "step": 2835
+    },
+    {
+      "epoch": 1.6387766878245817,
+      "grad_norm": 0.37287704407061595,
+      "learning_rate": 1.4706470555295605e-05,
+      "loss": 0.3554,
+      "step": 2840
+    },
+    {
+      "epoch": 1.641661858049625,
+      "grad_norm": 0.37701343446505425,
+      "learning_rate": 1.4684237839362963e-05,
+      "loss": 0.3654,
+      "step": 2845
+    },
+    {
+      "epoch": 1.6445470282746681,
+      "grad_norm": 0.38168823604260177,
+      "learning_rate": 1.4661975421291623e-05,
+      "loss": 0.3586,
+      "step": 2850
+    },
+    {
+      "epoch": 1.6474321984997116,
+      "grad_norm": 0.39975859055427077,
+      "learning_rate": 1.4639683442244663e-05,
+      "loss": 0.3716,
+      "step": 2855
+    },
+    {
+      "epoch": 1.6503173687247548,
+      "grad_norm": 0.37090442126154205,
+      "learning_rate": 1.4617362043572607e-05,
+      "loss": 0.3588,
+      "step": 2860
+    },
+    {
+      "epoch": 1.653202538949798,
+      "grad_norm": 0.3519117177291488,
+      "learning_rate": 1.4595011366812512e-05,
+      "loss": 0.3663,
+      "step": 2865
+    },
+    {
+      "epoch": 1.6560877091748414,
+      "grad_norm": 0.3737325432685507,
+      "learning_rate": 1.4572631553687103e-05,
+      "loss": 0.3582,
+      "step": 2870
+    },
+    {
+      "epoch": 1.6589728793998846,
+      "grad_norm": 0.38870437566944893,
+      "learning_rate": 1.4550222746103835e-05,
+      "loss": 0.351,
+      "step": 2875
+    },
+    {
+      "epoch": 1.6618580496249278,
+      "grad_norm": 0.3617260019248044,
+      "learning_rate": 1.4527785086154027e-05,
+      "loss": 0.3696,
+      "step": 2880
+    },
+    {
+      "epoch": 1.6647432198499712,
+      "grad_norm": 0.38332448310467415,
+      "learning_rate": 1.4505318716111943e-05,
+      "loss": 0.373,
+      "step": 2885
+    },
+    {
+      "epoch": 1.6676283900750144,
+      "grad_norm": 0.3747344452694502,
+      "learning_rate": 1.448282377843389e-05,
+      "loss": 0.3664,
+      "step": 2890
+    },
+    {
+      "epoch": 1.6705135603000576,
+      "grad_norm": 0.36202726723998613,
+      "learning_rate": 1.4460300415757323e-05,
+      "loss": 0.3766,
+      "step": 2895
+    },
+    {
+      "epoch": 1.673398730525101,
+      "grad_norm": 0.37923473785013656,
+      "learning_rate": 1.4437748770899936e-05,
+      "loss": 0.3814,
+      "step": 2900
+    },
+    {
+      "epoch": 1.6762839007501442,
+      "grad_norm": 0.3873960582212302,
+      "learning_rate": 1.4415168986858754e-05,
+      "loss": 0.3745,
+      "step": 2905
+    },
+    {
+      "epoch": 1.6791690709751874,
+      "grad_norm": 0.4013398358287003,
+      "learning_rate": 1.4392561206809232e-05,
+      "loss": 0.3798,
+      "step": 2910
+    },
+    {
+      "epoch": 1.6820542412002308,
+      "grad_norm": 0.40802241994691185,
+      "learning_rate": 1.4369925574104344e-05,
+      "loss": 0.3586,
+      "step": 2915
+    },
+    {
+      "epoch": 1.684939411425274,
+      "grad_norm": 0.4010573684279247,
+      "learning_rate": 1.4347262232273671e-05,
+      "loss": 0.3578,
+      "step": 2920
+    },
+    {
+      "epoch": 1.6878245816503172,
+      "grad_norm": 0.3609006547431779,
+      "learning_rate": 1.4324571325022496e-05,
+      "loss": 0.3592,
+      "step": 2925
+    },
+    {
+      "epoch": 1.6907097518753607,
+      "grad_norm": 0.3682631162288877,
+      "learning_rate": 1.4301852996230889e-05,
+      "loss": 0.3611,
+      "step": 2930
+    },
+    {
+      "epoch": 1.693594922100404,
+      "grad_norm": 0.39391349541100773,
+      "learning_rate": 1.42791073899528e-05,
+      "loss": 0.3549,
+      "step": 2935
+    },
+    {
+      "epoch": 1.696480092325447,
+      "grad_norm": 0.44233833212347734,
+      "learning_rate": 1.4256334650415135e-05,
+      "loss": 0.3537,
+      "step": 2940
+    },
+    {
+      "epoch": 1.6993652625504905,
+      "grad_norm": 0.3869468573057741,
+      "learning_rate": 1.4233534922016859e-05,
+      "loss": 0.37,
+      "step": 2945
+    },
+    {
+      "epoch": 1.702250432775534,
+      "grad_norm": 0.37586357463502135,
+      "learning_rate": 1.4210708349328062e-05,
+      "loss": 0.3675,
+      "step": 2950
+    },
+    {
+      "epoch": 1.7051356030005769,
+      "grad_norm": 0.3903279762819204,
+      "learning_rate": 1.418785507708905e-05,
+      "loss": 0.369,
+      "step": 2955
+    },
+    {
+      "epoch": 1.7080207732256203,
+      "grad_norm": 0.3969910656440696,
+      "learning_rate": 1.4164975250209432e-05,
+      "loss": 0.3694,
+      "step": 2960
+    },
+    {
+      "epoch": 1.7109059434506637,
+      "grad_norm": 0.3556902520239883,
+      "learning_rate": 1.4142069013767192e-05,
+      "loss": 0.3504,
+      "step": 2965
+    },
+    {
+      "epoch": 1.713791113675707,
+      "grad_norm": 0.3583546127221291,
+      "learning_rate": 1.4119136513007776e-05,
+      "loss": 0.3681,
+      "step": 2970
+    },
+    {
+      "epoch": 1.71667628390075,
+      "grad_norm": 0.3968433520239081,
+      "learning_rate": 1.409617789334317e-05,
+      "loss": 0.3752,
+      "step": 2975
+    },
+    {
+      "epoch": 1.7195614541257935,
+      "grad_norm": 0.3898401819324146,
+      "learning_rate": 1.4073193300350973e-05,
+      "loss": 0.3698,
+      "step": 2980
+    },
+    {
+      "epoch": 1.7224466243508367,
+      "grad_norm": 0.3698983985858203,
+      "learning_rate": 1.405018287977348e-05,
+      "loss": 0.3521,
+      "step": 2985
+    },
+    {
+      "epoch": 1.72533179457588,
+      "grad_norm": 0.4816180292439013,
+      "learning_rate": 1.402714677751675e-05,
+      "loss": 0.3604,
+      "step": 2990
+    },
+    {
+      "epoch": 1.7282169648009233,
+      "grad_norm": 0.4104869915376292,
+      "learning_rate": 1.40040851396497e-05,
+      "loss": 0.3799,
+      "step": 2995
+    },
+    {
+      "epoch": 1.7311021350259665,
+      "grad_norm": 0.3917852021920299,
+      "learning_rate": 1.3980998112403146e-05,
+      "loss": 0.3584,
+      "step": 3000
+    },
+    {
+      "epoch": 1.7339873052510097,
+      "grad_norm": 0.3910325407989454,
+      "learning_rate": 1.3957885842168908e-05,
+      "loss": 0.3579,
+      "step": 3005
+    },
+    {
+      "epoch": 1.7368724754760532,
+      "grad_norm": 0.35526389545543485,
+      "learning_rate": 1.3934748475498867e-05,
+      "loss": 0.3588,
+      "step": 3010
+    },
+    {
+      "epoch": 1.7397576457010964,
+      "grad_norm": 0.39546625910444144,
+      "learning_rate": 1.3911586159104032e-05,
+      "loss": 0.3476,
+      "step": 3015
+    },
+    {
+      "epoch": 1.7426428159261396,
+      "grad_norm": 0.369055887907894,
+      "learning_rate": 1.3888399039853618e-05,
+      "loss": 0.352,
+      "step": 3020
+    },
+    {
+      "epoch": 1.745527986151183,
+      "grad_norm": 0.35801953491056676,
+      "learning_rate": 1.3865187264774114e-05,
+      "loss": 0.3554,
+      "step": 3025
+    },
+    {
+      "epoch": 1.7484131563762262,
+      "grad_norm": 0.3531901383810597,
+      "learning_rate": 1.3841950981048342e-05,
+      "loss": 0.3522,
+      "step": 3030
+    },
+    {
+      "epoch": 1.7512983266012694,
+      "grad_norm": 0.39607234863789387,
+      "learning_rate": 1.3818690336014543e-05,
+      "loss": 0.3762,
+      "step": 3035
+    },
+    {
+      "epoch": 1.7541834968263128,
+      "grad_norm": 0.38127586298657323,
+      "learning_rate": 1.3795405477165406e-05,
+      "loss": 0.3563,
+      "step": 3040
+    },
+    {
+      "epoch": 1.757068667051356,
+      "grad_norm": 0.43662892822050353,
+      "learning_rate": 1.3772096552147187e-05,
+      "loss": 0.352,
+      "step": 3045
+    },
+    {
+      "epoch": 1.7599538372763992,
+      "grad_norm": 0.3889378354940805,
+      "learning_rate": 1.3748763708758717e-05,
+      "loss": 0.3686,
+      "step": 3050
+    },
+    {
+      "epoch": 1.7628390075014426,
+      "grad_norm": 0.3972446044944282,
+      "learning_rate": 1.3725407094950506e-05,
+      "loss": 0.3621,
+      "step": 3055
+    },
+    {
+      "epoch": 1.7657241777264858,
+      "grad_norm": 0.40962998535223355,
+      "learning_rate": 1.3702026858823781e-05,
+      "loss": 0.3876,
+      "step": 3060
+    },
+    {
+      "epoch": 1.768609347951529,
+      "grad_norm": 0.35961773766923205,
+      "learning_rate": 1.3678623148629555e-05,
+      "loss": 0.3736,
+      "step": 3065
+    },
+    {
+      "epoch": 1.7714945181765724,
+      "grad_norm": 0.38471972852000913,
+      "learning_rate": 1.3655196112767695e-05,
+      "loss": 0.3491,
+      "step": 3070
+    },
+    {
+      "epoch": 1.7743796884016156,
+      "grad_norm": 0.36648434186274914,
+      "learning_rate": 1.363174589978596e-05,
+      "loss": 0.3593,
+      "step": 3075
+    },
+    {
+      "epoch": 1.7772648586266588,
+      "grad_norm": 0.3872125897463841,
+      "learning_rate": 1.3608272658379087e-05,
+      "loss": 0.3646,
+      "step": 3080
+    },
+    {
+      "epoch": 1.7801500288517023,
+      "grad_norm": 0.3740044921465578,
+      "learning_rate": 1.358477653738782e-05,
+      "loss": 0.3517,
+      "step": 3085
+    },
+    {
+      "epoch": 1.7830351990767457,
+      "grad_norm": 0.38429954155990775,
+      "learning_rate": 1.3561257685797992e-05,
+      "loss": 0.3655,
+      "step": 3090
+    },
+    {
+      "epoch": 1.7859203693017887,
+      "grad_norm": 0.3478740415125172,
+      "learning_rate": 1.3537716252739552e-05,
+      "loss": 0.3675,
+      "step": 3095
+    },
+    {
+      "epoch": 1.788805539526832,
+      "grad_norm": 0.3597910283031001,
+      "learning_rate": 1.3514152387485651e-05,
+      "loss": 0.3489,
+      "step": 3100
+    },
+    {
+      "epoch": 1.7916907097518755,
+      "grad_norm": 0.41193186738188275,
+      "learning_rate": 1.3490566239451672e-05,
+      "loss": 0.3818,
+      "step": 3105
+    },
+    {
+      "epoch": 1.7945758799769185,
+      "grad_norm": 0.33623886878783116,
+      "learning_rate": 1.3466957958194284e-05,
+      "loss": 0.3618,
+      "step": 3110
+    },
+    {
+      "epoch": 1.797461050201962,
+      "grad_norm": 0.35667893854991767,
+      "learning_rate": 1.344332769341052e-05,
+      "loss": 0.3558,
+      "step": 3115
+    },
+    {
+      "epoch": 1.8003462204270053,
+      "grad_norm": 0.39258025755310366,
+      "learning_rate": 1.3419675594936778e-05,
+      "loss": 0.3477,
+      "step": 3120
+    },
+    {
+      "epoch": 1.8032313906520485,
+      "grad_norm": 0.3782720017741446,
+      "learning_rate": 1.339600181274793e-05,
+      "loss": 0.3735,
+      "step": 3125
+    },
+    {
+      "epoch": 1.8061165608770917,
+      "grad_norm": 0.4035057736179016,
+      "learning_rate": 1.3372306496956324e-05,
+      "loss": 0.3845,
+      "step": 3130
+    },
+    {
+      "epoch": 1.8090017311021351,
+      "grad_norm": 0.3763574258043012,
+      "learning_rate": 1.3348589797810854e-05,
+      "loss": 0.3602,
+      "step": 3135
+    },
+    {
+      "epoch": 1.8118869013271783,
+      "grad_norm": 0.35992628716248726,
+      "learning_rate": 1.3324851865696e-05,
+      "loss": 0.3792,
+      "step": 3140
+    },
+    {
+      "epoch": 1.8147720715522215,
+      "grad_norm": 0.37804380856317177,
+      "learning_rate": 1.3301092851130886e-05,
+      "loss": 0.3657,
+      "step": 3145
+    },
+    {
+      "epoch": 1.817657241777265,
+      "grad_norm": 0.3598537111947173,
+      "learning_rate": 1.3277312904768314e-05,
+      "loss": 0.3721,
+      "step": 3150
+    },
+    {
+      "epoch": 1.8205424120023082,
+      "grad_norm": 0.34265734973552825,
+      "learning_rate": 1.32535121773938e-05,
+      "loss": 0.3673,
+      "step": 3155
+    },
+    {
+      "epoch": 1.8234275822273514,
+      "grad_norm": 0.40992539604640754,
+      "learning_rate": 1.3229690819924653e-05,
+      "loss": 0.3738,
+      "step": 3160
+    },
+    {
+      "epoch": 1.8263127524523948,
+      "grad_norm": 0.39170242560206425,
+      "learning_rate": 1.3205848983408968e-05,
+      "loss": 0.3725,
+      "step": 3165
+    },
+    {
+      "epoch": 1.829197922677438,
+      "grad_norm": 0.4135698812295932,
+      "learning_rate": 1.3181986819024717e-05,
+      "loss": 0.3765,
+      "step": 3170
+    },
+    {
+      "epoch": 1.8320830929024812,
+      "grad_norm": 0.39775337228906954,
+      "learning_rate": 1.3158104478078753e-05,
+      "loss": 0.3648,
+      "step": 3175
+    },
+    {
+      "epoch": 1.8349682631275246,
+      "grad_norm": 0.4025320364227362,
+      "learning_rate": 1.3134202112005876e-05,
+      "loss": 0.3679,
+      "step": 3180
+    },
+    {
+      "epoch": 1.8378534333525678,
+      "grad_norm": 0.39435133301085734,
+      "learning_rate": 1.311027987236786e-05,
+      "loss": 0.3766,
+      "step": 3185
+    },
+    {
+      "epoch": 1.840738603577611,
+      "grad_norm": 0.3791471283767141,
+      "learning_rate": 1.3086337910852487e-05,
+      "loss": 0.3867,
+      "step": 3190
+    },
+    {
+      "epoch": 1.8436237738026544,
+      "grad_norm": 0.37370201756820587,
+      "learning_rate": 1.306237637927261e-05,
+      "loss": 0.3669,
+      "step": 3195
+    },
+    {
+      "epoch": 1.8465089440276976,
+      "grad_norm": 0.38003239897050844,
+      "learning_rate": 1.3038395429565156e-05,
+      "loss": 0.3597,
+      "step": 3200
+    },
+    {
+      "epoch": 1.8493941142527408,
+      "grad_norm": 0.37745120795042336,
+      "learning_rate": 1.3014395213790187e-05,
+      "loss": 0.3666,
+      "step": 3205
+    },
+    {
+      "epoch": 1.8522792844777842,
+      "grad_norm": 0.369849417274224,
+      "learning_rate": 1.299037588412993e-05,
+      "loss": 0.3731,
+      "step": 3210
+    },
+    {
+      "epoch": 1.8551644547028274,
+      "grad_norm": 0.37108268900622904,
+      "learning_rate": 1.2966337592887807e-05,
+      "loss": 0.3619,
+      "step": 3215
+    },
+    {
+      "epoch": 1.8580496249278706,
+      "grad_norm": 0.3603718565169327,
+      "learning_rate": 1.2942280492487478e-05,
+      "loss": 0.3711,
+      "step": 3220
+    },
+    {
+      "epoch": 1.860934795152914,
+      "grad_norm": 0.38800493629769506,
+      "learning_rate": 1.291820473547186e-05,
+      "loss": 0.3672,
+      "step": 3225
+    },
+    {
+      "epoch": 1.8638199653779572,
+      "grad_norm": 0.3778795059973765,
+      "learning_rate": 1.2894110474502182e-05,
+      "loss": 0.3665,
+      "step": 3230
+    },
+    {
+      "epoch": 1.8667051356030004,
+      "grad_norm": 0.40908012961452245,
+      "learning_rate": 1.2869997862356994e-05,
+      "loss": 0.3639,
+      "step": 3235
+    },
+    {
+      "epoch": 1.8695903058280439,
+      "grad_norm": 0.36111567245322745,
+      "learning_rate": 1.2845867051931214e-05,
+      "loss": 0.3679,
+      "step": 3240
+    },
+    {
+      "epoch": 1.8724754760530873,
+      "grad_norm": 0.4126916975430262,
+      "learning_rate": 1.2821718196235147e-05,
+      "loss": 0.3577,
+      "step": 3245
+    },
+    {
+      "epoch": 1.8753606462781303,
+      "grad_norm": 0.3896170917861394,
+      "learning_rate": 1.2797551448393527e-05,
+      "loss": 0.3533,
+      "step": 3250
+    },
+    {
+      "epoch": 1.8782458165031737,
+      "grad_norm": 0.3918073404798387,
+      "learning_rate": 1.2773366961644537e-05,
+      "loss": 0.3693,
+      "step": 3255
+    },
+    {
+      "epoch": 1.881130986728217,
+      "grad_norm": 0.36560888725527363,
+      "learning_rate": 1.274916488933884e-05,
+      "loss": 0.3698,
+      "step": 3260
+    },
+    {
+      "epoch": 1.88401615695326,
+      "grad_norm": 0.35904154235729674,
+      "learning_rate": 1.2724945384938608e-05,
+      "loss": 0.3621,
+      "step": 3265
+    },
+    {
+      "epoch": 1.8869013271783035,
+      "grad_norm": 0.3712234770214718,
+      "learning_rate": 1.2700708602016545e-05,
+      "loss": 0.3596,
+      "step": 3270
+    },
+    {
+      "epoch": 1.889786497403347,
+      "grad_norm": 0.35515855306613603,
+      "learning_rate": 1.2676454694254918e-05,
+      "loss": 0.3436,
+      "step": 3275
+    },
+    {
+      "epoch": 1.8926716676283901,
+      "grad_norm": 0.38038321952993087,
+      "learning_rate": 1.265218381544458e-05,
+      "loss": 0.3598,
+      "step": 3280
+    },
+    {
+      "epoch": 1.8955568378534333,
+      "grad_norm": 0.38405790753936503,
+      "learning_rate": 1.2627896119483998e-05,
+      "loss": 0.3613,
+      "step": 3285
+    },
+    {
+      "epoch": 1.8984420080784767,
+      "grad_norm": 0.35411934799722744,
+      "learning_rate": 1.2603591760378266e-05,
+      "loss": 0.3621,
+      "step": 3290
+    },
+    {
+      "epoch": 1.90132717830352,
+      "grad_norm": 0.3833342471723807,
+      "learning_rate": 1.2579270892238147e-05,
+      "loss": 0.3688,
+      "step": 3295
+    },
+    {
+      "epoch": 1.9042123485285631,
+      "grad_norm": 0.3783988493179844,
+      "learning_rate": 1.2554933669279076e-05,
+      "loss": 0.3566,
+      "step": 3300
+    },
+    {
+      "epoch": 1.9070975187536066,
+      "grad_norm": 0.3849340817581559,
+      "learning_rate": 1.2530580245820202e-05,
+      "loss": 0.3664,
+      "step": 3305
+    },
+    {
+      "epoch": 1.9099826889786498,
+      "grad_norm": 0.36371725434046853,
+      "learning_rate": 1.250621077628339e-05,
+      "loss": 0.3622,
+      "step": 3310
+    },
+    {
+      "epoch": 1.912867859203693,
+      "grad_norm": 0.3741399204987393,
+      "learning_rate": 1.2481825415192255e-05,
+      "loss": 0.3381,
+      "step": 3315
+    },
+    {
+      "epoch": 1.9157530294287364,
+      "grad_norm": 0.4669077205458527,
+      "learning_rate": 1.245742431717118e-05,
+      "loss": 0.3823,
+      "step": 3320
+    },
+    {
+      "epoch": 1.9186381996537796,
+      "grad_norm": 0.35803705351389503,
+      "learning_rate": 1.2433007636944331e-05,
+      "loss": 0.351,
+      "step": 3325
+    },
+    {
+      "epoch": 1.9215233698788228,
+      "grad_norm": 0.3701949796075695,
+      "learning_rate": 1.240857552933468e-05,
+      "loss": 0.3458,
+      "step": 3330
+    },
+    {
+      "epoch": 1.9244085401038662,
+      "grad_norm": 0.3867704015823223,
+      "learning_rate": 1.2384128149263023e-05,
+      "loss": 0.372,
+      "step": 3335
+    },
+    {
+      "epoch": 1.9272937103289094,
+      "grad_norm": 0.34734943471086316,
+      "learning_rate": 1.2359665651746995e-05,
+      "loss": 0.3587,
+      "step": 3340
+    },
+    {
+      "epoch": 1.9301788805539526,
+      "grad_norm": 0.3456302062945545,
+      "learning_rate": 1.2335188191900087e-05,
+      "loss": 0.3639,
+      "step": 3345
+    },
+    {
+      "epoch": 1.933064050778996,
+      "grad_norm": 0.37432101569615994,
+      "learning_rate": 1.231069592493067e-05,
+      "loss": 0.3567,
+      "step": 3350
+    },
+    {
+      "epoch": 1.9359492210040392,
+      "grad_norm": 0.38495008542439035,
+      "learning_rate": 1.2286189006140998e-05,
+      "loss": 0.355,
+      "step": 3355
+    },
+    {
+      "epoch": 1.9388343912290824,
+      "grad_norm": 0.3603027711889298,
+      "learning_rate": 1.2261667590926232e-05,
+      "loss": 0.3643,
+      "step": 3360
+    },
+    {
+      "epoch": 1.9417195614541258,
+      "grad_norm": 0.3801200430528795,
+      "learning_rate": 1.223713183477346e-05,
+      "loss": 0.3547,
+      "step": 3365
+    },
+    {
+      "epoch": 1.944604731679169,
+      "grad_norm": 0.3801873136595747,
+      "learning_rate": 1.2212581893260694e-05,
+      "loss": 0.3674,
+      "step": 3370
+    },
+    {
+      "epoch": 1.9474899019042122,
+      "grad_norm": 0.3724561589479728,
+      "learning_rate": 1.2188017922055901e-05,
+      "loss": 0.3466,
+      "step": 3375
+    },
+    {
+      "epoch": 1.9503750721292556,
+      "grad_norm": 0.38382927060326005,
+      "learning_rate": 1.2163440076916005e-05,
+      "loss": 0.3736,
+      "step": 3380
+    },
+    {
+      "epoch": 1.953260242354299,
+      "grad_norm": 0.40285349364914064,
+      "learning_rate": 1.2138848513685902e-05,
+      "loss": 0.3518,
+      "step": 3385
+    },
+    {
+      "epoch": 1.956145412579342,
+      "grad_norm": 0.39374554584138266,
+      "learning_rate": 1.2114243388297478e-05,
+      "loss": 0.3553,
+      "step": 3390
+    },
+    {
+      "epoch": 1.9590305828043855,
+      "grad_norm": 0.3733836670088024,
+      "learning_rate": 1.2089624856768603e-05,
+      "loss": 0.3782,
+      "step": 3395
+    },
+    {
+      "epoch": 1.9619157530294289,
+      "grad_norm": 0.3512295476922515,
+      "learning_rate": 1.2064993075202172e-05,
+      "loss": 0.352,
+      "step": 3400
+    },
+    {
+      "epoch": 1.9648009232544719,
+      "grad_norm": 0.35755975140067886,
+      "learning_rate": 1.2040348199785074e-05,
+      "loss": 0.3614,
+      "step": 3405
+    },
+    {
+      "epoch": 1.9676860934795153,
+      "grad_norm": 0.3541512834606944,
+      "learning_rate": 1.2015690386787245e-05,
+      "loss": 0.359,
+      "step": 3410
+    },
+    {
+      "epoch": 1.9705712637045587,
+      "grad_norm": 0.3774486665711377,
+      "learning_rate": 1.1991019792560648e-05,
+      "loss": 0.3658,
+      "step": 3415
+    },
+    {
+      "epoch": 1.9734564339296017,
+      "grad_norm": 0.38262821666740215,
+      "learning_rate": 1.1966336573538287e-05,
+      "loss": 0.3763,
+      "step": 3420
+    },
+    {
+      "epoch": 1.976341604154645,
+      "grad_norm": 0.3903065644133358,
+      "learning_rate": 1.1941640886233224e-05,
+      "loss": 0.3589,
+      "step": 3425
+    },
+    {
+      "epoch": 1.9792267743796885,
+      "grad_norm": 0.3579664105437974,
+      "learning_rate": 1.1916932887237571e-05,
+      "loss": 0.3594,
+      "step": 3430
+    },
+    {
+      "epoch": 1.9821119446047317,
+      "grad_norm": 0.385483963575225,
+      "learning_rate": 1.1892212733221523e-05,
+      "loss": 0.3738,
+      "step": 3435
+    },
+    {
+      "epoch": 1.984997114829775,
+      "grad_norm": 0.37602191182512,
+      "learning_rate": 1.1867480580932336e-05,
+      "loss": 0.3539,
+      "step": 3440
+    },
+    {
+      "epoch": 1.9878822850548183,
+      "grad_norm": 0.40221670992777864,
+      "learning_rate": 1.1842736587193349e-05,
+      "loss": 0.3618,
+      "step": 3445
+    },
+    {
+      "epoch": 1.9907674552798615,
+      "grad_norm": 0.3945792439685852,
+      "learning_rate": 1.1817980908902981e-05,
+      "loss": 0.3617,
+      "step": 3450
+    },
+    {
+      "epoch": 1.9936526255049047,
+      "grad_norm": 0.3638645756888353,
+      "learning_rate": 1.1793213703033755e-05,
+      "loss": 0.3621,
+      "step": 3455
+    },
+    {
+      "epoch": 1.9965377957299482,
+      "grad_norm": 0.383149019555126,
+      "learning_rate": 1.1768435126631276e-05,
+      "loss": 0.3596,
+      "step": 3460
+    },
+    {
+      "epoch": 1.9994229659549914,
+      "grad_norm": 0.380227166238423,
+      "learning_rate": 1.1743645336813248e-05,
+      "loss": 0.3638,
+      "step": 3465
+    },
+    {
+      "epoch": 2.0,
+      "eval_loss": 0.4331684410572052,
+      "eval_runtime": 0.5935,
+      "eval_samples_per_second": 129.744,
+      "eval_steps_per_second": 3.37,
+      "step": 3466
+    },
+    {
+      "epoch": 2.0023081361800346,
+      "grad_norm": 0.41074392868999404,
+      "learning_rate": 1.1718844490768488e-05,
+      "loss": 0.3104,
+      "step": 3470
+    },
+    {
+      "epoch": 2.005193306405078,
+      "grad_norm": 0.3928008512339745,
+      "learning_rate": 1.1694032745755906e-05,
+      "loss": 0.261,
+      "step": 3475
+    },
+    {
+      "epoch": 2.008078476630121,
+      "grad_norm": 0.3931528202345155,
+      "learning_rate": 1.1669210259103535e-05,
+      "loss": 0.287,
+      "step": 3480
+    },
+    {
+      "epoch": 2.0109636468551644,
+      "grad_norm": 0.3869283477561703,
+      "learning_rate": 1.164437718820751e-05,
+      "loss": 0.2683,
+      "step": 3485
+    },
+    {
+      "epoch": 2.013848817080208,
+      "grad_norm": 0.3991121860772099,
+      "learning_rate": 1.1619533690531076e-05,
+      "loss": 0.2695,
+      "step": 3490
+    },
+    {
+      "epoch": 2.016733987305251,
+      "grad_norm": 0.4025523043283219,
+      "learning_rate": 1.1594679923603606e-05,
+      "loss": 0.2662,
+      "step": 3495
+    },
+    {
+      "epoch": 2.019619157530294,
+      "grad_norm": 0.4441663382272395,
+      "learning_rate": 1.1569816045019575e-05,
+      "loss": 0.2663,
+      "step": 3500
+    },
+    {
+      "epoch": 2.0225043277553376,
+      "grad_norm": 0.4158755932657419,
+      "learning_rate": 1.1544942212437586e-05,
+      "loss": 0.2701,
+      "step": 3505
+    },
+    {
+      "epoch": 2.025389497980381,
+      "grad_norm": 0.4324043740872268,
+      "learning_rate": 1.152005858357935e-05,
+      "loss": 0.2751,
+      "step": 3510
+    },
+    {
+      "epoch": 2.028274668205424,
+      "grad_norm": 0.41576677704649545,
+      "learning_rate": 1.1495165316228704e-05,
+      "loss": 0.2837,
+      "step": 3515
+    },
+    {
+      "epoch": 2.0311598384304674,
+      "grad_norm": 0.4261462623114962,
+      "learning_rate": 1.1470262568230593e-05,
+      "loss": 0.2778,
+      "step": 3520
+    },
+    {
+      "epoch": 2.034045008655511,
+      "grad_norm": 0.4222222859454043,
+      "learning_rate": 1.1445350497490077e-05,
+      "loss": 0.2663,
+      "step": 3525
+    },
+    {
+      "epoch": 2.036930178880554,
+      "grad_norm": 0.4149492406466183,
+      "learning_rate": 1.1420429261971342e-05,
+      "loss": 0.2615,
+      "step": 3530
+    },
+    {
+      "epoch": 2.0398153491055973,
+      "grad_norm": 0.4325594095101094,
+      "learning_rate": 1.139549901969667e-05,
+      "loss": 0.276,
+      "step": 3535
+    },
+    {
+      "epoch": 2.0427005193306407,
+      "grad_norm": 0.4066105524550755,
+      "learning_rate": 1.137055992874547e-05,
+      "loss": 0.2749,
+      "step": 3540
+    },
+    {
+      "epoch": 2.0455856895556837,
+      "grad_norm": 0.4091251188163908,
+      "learning_rate": 1.1345612147253247e-05,
+      "loss": 0.2662,
+      "step": 3545
+    },
+    {
+      "epoch": 2.048470859780727,
+      "grad_norm": 0.3948683403145309,
+      "learning_rate": 1.1320655833410621e-05,
+      "loss": 0.2813,
+      "step": 3550
+    },
+    {
+      "epoch": 2.0513560300057705,
+      "grad_norm": 0.41481838952861994,
+      "learning_rate": 1.1295691145462306e-05,
+      "loss": 0.2863,
+      "step": 3555
+    },
+    {
+      "epoch": 2.0542412002308135,
+      "grad_norm": 0.4032363652268982,
+      "learning_rate": 1.1270718241706118e-05,
+      "loss": 0.2565,
+      "step": 3560
+    },
+    {
+      "epoch": 2.057126370455857,
+      "grad_norm": 0.48615644869007646,
+      "learning_rate": 1.1245737280491973e-05,
+      "loss": 0.2856,
+      "step": 3565
+    },
+    {
+      "epoch": 2.0600115406809003,
+      "grad_norm": 0.4219271094611073,
+      "learning_rate": 1.122074842022087e-05,
+      "loss": 0.2719,
+      "step": 3570
+    },
+    {
+      "epoch": 2.0628967109059433,
+      "grad_norm": 0.41257650256952855,
+      "learning_rate": 1.1195751819343906e-05,
+      "loss": 0.2745,
+      "step": 3575
+    },
+    {
+      "epoch": 2.0657818811309867,
+      "grad_norm": 0.428723721941346,
+      "learning_rate": 1.1170747636361245e-05,
+      "loss": 0.2779,
+      "step": 3580
+    },
+    {
+      "epoch": 2.06866705135603,
+      "grad_norm": 0.40354050541964437,
+      "learning_rate": 1.1145736029821148e-05,
+      "loss": 0.283,
+      "step": 3585
+    },
+    {
+      "epoch": 2.071552221581073,
+      "grad_norm": 0.43472195003969,
+      "learning_rate": 1.1120717158318928e-05,
+      "loss": 0.281,
+      "step": 3590
+    },
+    {
+      "epoch": 2.0744373918061165,
+      "grad_norm": 0.4123476736485213,
+      "learning_rate": 1.1095691180495973e-05,
+      "loss": 0.2798,
+      "step": 3595
+    },
+    {
+      "epoch": 2.07732256203116,
+      "grad_norm": 0.4150389828010794,
+      "learning_rate": 1.1070658255038733e-05,
+      "loss": 0.2871,
+      "step": 3600
+    },
+    {
+      "epoch": 2.080207732256203,
+      "grad_norm": 0.4712720490312681,
+      "learning_rate": 1.1045618540677712e-05,
+      "loss": 0.2815,
+      "step": 3605
+    },
+    {
+      "epoch": 2.0830929024812463,
+      "grad_norm": 0.5081431777602221,
+      "learning_rate": 1.1020572196186457e-05,
+      "loss": 0.263,
+      "step": 3610
+    },
+    {
+      "epoch": 2.0859780727062898,
+      "grad_norm": 0.4549015236527131,
+      "learning_rate": 1.0995519380380557e-05,
+      "loss": 0.2717,
+      "step": 3615
+    },
+    {
+      "epoch": 2.0888632429313327,
+      "grad_norm": 0.5033902066650326,
+      "learning_rate": 1.097046025211664e-05,
+      "loss": 0.2813,
+      "step": 3620
+    },
+    {
+      "epoch": 2.091748413156376,
+      "grad_norm": 0.40826808731777586,
+      "learning_rate": 1.0945394970291354e-05,
+      "loss": 0.2607,
+      "step": 3625
+    },
+    {
+      "epoch": 2.0946335833814196,
+      "grad_norm": 0.4593185041491404,
+      "learning_rate": 1.0920323693840368e-05,
+      "loss": 0.28,
+      "step": 3630
+    },
+    {
+      "epoch": 2.0975187536064626,
+      "grad_norm": 0.4374169481716498,
+      "learning_rate": 1.0895246581737366e-05,
+      "loss": 0.2968,
+      "step": 3635
+    },
+    {
+      "epoch": 2.100403923831506,
+      "grad_norm": 0.4281118948367996,
+      "learning_rate": 1.087016379299303e-05,
+      "loss": 0.2877,
+      "step": 3640
+    },
+    {
+      "epoch": 2.1032890940565494,
+      "grad_norm": 0.45375611915195996,
+      "learning_rate": 1.0845075486654037e-05,
+      "loss": 0.261,
+      "step": 3645
+    },
+    {
+      "epoch": 2.106174264281593,
+      "grad_norm": 0.4267355049796937,
+      "learning_rate": 1.0819981821802053e-05,
+      "loss": 0.2786,
+      "step": 3650
+    },
+    {
+      "epoch": 2.109059434506636,
+      "grad_norm": 0.4376492488121602,
+      "learning_rate": 1.0794882957552722e-05,
+      "loss": 0.2766,
+      "step": 3655
+    },
+    {
+      "epoch": 2.111944604731679,
+      "grad_norm": 0.45029828413170214,
+      "learning_rate": 1.0769779053054651e-05,
+      "loss": 0.283,
+      "step": 3660
+    },
+    {
+      "epoch": 2.1148297749567226,
+      "grad_norm": 0.43071600979778896,
+      "learning_rate": 1.0744670267488417e-05,
+      "loss": 0.2848,
+      "step": 3665
+    },
+    {
+      "epoch": 2.1177149451817656,
+      "grad_norm": 0.3970843751114558,
+      "learning_rate": 1.071955676006554e-05,
+      "loss": 0.2762,
+      "step": 3670
+    },
+    {
+      "epoch": 2.120600115406809,
+      "grad_norm": 0.38320271641559517,
+      "learning_rate": 1.0694438690027475e-05,
+      "loss": 0.2819,
+      "step": 3675
+    },
+    {
+      "epoch": 2.1234852856318525,
+      "grad_norm": 0.4529791521783765,
+      "learning_rate": 1.0669316216644623e-05,
+      "loss": 0.266,
+      "step": 3680
+    },
+    {
+      "epoch": 2.1263704558568954,
+      "grad_norm": 0.42862083344692714,
+      "learning_rate": 1.0644189499215292e-05,
+      "loss": 0.2765,
+      "step": 3685
+    },
+    {
+      "epoch": 2.129255626081939,
+      "grad_norm": 0.43384197829762733,
+      "learning_rate": 1.061905869706471e-05,
+      "loss": 0.3061,
+      "step": 3690
+    },
+    {
+      "epoch": 2.1321407963069823,
+      "grad_norm": 0.4224346600569041,
+      "learning_rate": 1.0593923969544e-05,
+      "loss": 0.3003,
+      "step": 3695
+    },
+    {
+      "epoch": 2.1350259665320253,
+      "grad_norm": 0.4332701906364038,
+      "learning_rate": 1.056878547602918e-05,
+      "loss": 0.2765,
+      "step": 3700
+    },
+    {
+      "epoch": 2.1379111367570687,
+      "grad_norm": 0.4393240656228905,
+      "learning_rate": 1.0543643375920145e-05,
+      "loss": 0.2879,
+      "step": 3705
+    },
+    {
+      "epoch": 2.140796306982112,
+      "grad_norm": 0.4036822061189937,
+      "learning_rate": 1.0518497828639658e-05,
+      "loss": 0.27,
+      "step": 3710
+    },
+    {
+      "epoch": 2.143681477207155,
+      "grad_norm": 0.427794972829589,
+      "learning_rate": 1.0493348993632337e-05,
+      "loss": 0.2749,
+      "step": 3715
+    },
+    {
+      "epoch": 2.1465666474321985,
+      "grad_norm": 0.39124739436748485,
+      "learning_rate": 1.046819703036366e-05,
+      "loss": 0.2718,
+      "step": 3720
+    },
+    {
+      "epoch": 2.149451817657242,
+      "grad_norm": 0.4099521410733268,
+      "learning_rate": 1.0443042098318927e-05,
+      "loss": 0.2733,
+      "step": 3725
+    },
+    {
+      "epoch": 2.152336987882285,
+      "grad_norm": 0.4225255601400179,
+      "learning_rate": 1.0417884357002268e-05,
+      "loss": 0.2773,
+      "step": 3730
+    },
+    {
+      "epoch": 2.1552221581073283,
+      "grad_norm": 0.43412449015519566,
+      "learning_rate": 1.039272396593563e-05,
+      "loss": 0.2544,
+      "step": 3735
+    },
+    {
+      "epoch": 2.1581073283323717,
+      "grad_norm": 0.414617071495182,
+      "learning_rate": 1.0367561084657752e-05,
+      "loss": 0.2672,
+      "step": 3740
+    },
+    {
+      "epoch": 2.1609924985574147,
+      "grad_norm": 0.4451465025534693,
+      "learning_rate": 1.0342395872723173e-05,
+      "loss": 0.2597,
+      "step": 3745
+    },
+    {
+      "epoch": 2.163877668782458,
+      "grad_norm": 0.4052440261347642,
+      "learning_rate": 1.03172284897012e-05,
+      "loss": 0.2758,
+      "step": 3750
+    },
+    {
+      "epoch": 2.1667628390075016,
+      "grad_norm": 0.4393651404130225,
+      "learning_rate": 1.0292059095174923e-05,
+      "loss": 0.281,
+      "step": 3755
+    },
+    {
+      "epoch": 2.1696480092325445,
+      "grad_norm": 0.43033068987713674,
+      "learning_rate": 1.026688784874017e-05,
+      "loss": 0.2839,
+      "step": 3760
+    },
+    {
+      "epoch": 2.172533179457588,
+      "grad_norm": 0.4187797584558774,
+      "learning_rate": 1.024171491000452e-05,
+      "loss": 0.2955,
+      "step": 3765
+    },
+    {
+      "epoch": 2.1754183496826314,
+      "grad_norm": 0.4306833753003216,
+      "learning_rate": 1.021654043858628e-05,
+      "loss": 0.268,
+      "step": 3770
+    },
+    {
+      "epoch": 2.1783035199076743,
+      "grad_norm": 0.46354342504920265,
+      "learning_rate": 1.0191364594113475e-05,
+      "loss": 0.2753,
+      "step": 3775
+    },
+    {
+      "epoch": 2.1811886901327178,
+      "grad_norm": 0.4667906762400645,
+      "learning_rate": 1.0166187536222844e-05,
+      "loss": 0.2826,
+      "step": 3780
+    },
+    {
+      "epoch": 2.184073860357761,
+      "grad_norm": 0.4268396541209257,
+      "learning_rate": 1.0141009424558803e-05,
+      "loss": 0.2594,
+      "step": 3785
+    },
+    {
+      "epoch": 2.1869590305828046,
+      "grad_norm": 0.42773226782347973,
+      "learning_rate": 1.0115830418772474e-05,
+      "loss": 0.2815,
+      "step": 3790
+    },
+    {
+      "epoch": 2.1898442008078476,
+      "grad_norm": 0.42670690759544905,
+      "learning_rate": 1.0090650678520624e-05,
+      "loss": 0.2884,
+      "step": 3795
+    },
+    {
+      "epoch": 2.192729371032891,
+      "grad_norm": 0.43803219635914387,
+      "learning_rate": 1.0065470363464694e-05,
+      "loss": 0.2789,
+      "step": 3800
+    },
+    {
+      "epoch": 2.1956145412579344,
+      "grad_norm": 0.4799532890386973,
+      "learning_rate": 1.0040289633269764e-05,
+      "loss": 0.2945,
+      "step": 3805
+    },
+    {
+      "epoch": 2.1984997114829774,
+      "grad_norm": 0.4207961621120341,
+      "learning_rate": 1.0015108647603545e-05,
+      "loss": 0.2665,
+      "step": 3810
+    },
+    {
+      "epoch": 2.201384881708021,
+      "grad_norm": 0.43772623733868893,
+      "learning_rate": 9.98992756613537e-06,
+      "loss": 0.2791,
+      "step": 3815
+    },
+    {
+      "epoch": 2.2042700519330642,
+      "grad_norm": 0.4246990837045649,
+      "learning_rate": 9.96474654853518e-06,
+      "loss": 0.2702,
+      "step": 3820
+    },
+    {
+      "epoch": 2.207155222158107,
+      "grad_norm": 0.4214383786492429,
+      "learning_rate": 9.93956575447251e-06,
+      "loss": 0.2724,
+      "step": 3825
+    },
+    {
+      "epoch": 2.2100403923831506,
+      "grad_norm": 0.42698570251672047,
+      "learning_rate": 9.914385343615476e-06,
+      "loss": 0.2714,
+      "step": 3830
+    },
+    {
+      "epoch": 2.212925562608194,
+      "grad_norm": 0.4504000608499869,
+      "learning_rate": 9.889205475629763e-06,
+      "loss": 0.2754,
+      "step": 3835
+    },
+    {
+      "epoch": 2.215810732833237,
+      "grad_norm": 0.4323598292397766,
+      "learning_rate": 9.864026310177621e-06,
+      "loss": 0.2624,
+      "step": 3840
+    },
+    {
+      "epoch": 2.2186959030582805,
+      "grad_norm": 0.40922887326049673,
+      "learning_rate": 9.83884800691684e-06,
+      "loss": 0.2633,
+      "step": 3845
+    },
+    {
+      "epoch": 2.221581073283324,
+      "grad_norm": 0.4357930415869159,
+      "learning_rate": 9.813670725499745e-06,
+      "loss": 0.2904,
+      "step": 3850
+    },
+    {
+      "epoch": 2.224466243508367,
+      "grad_norm": 0.47764591829975994,
+      "learning_rate": 9.788494625572177e-06,
+      "loss": 0.2685,
+      "step": 3855
+    },
+    {
+      "epoch": 2.2273514137334103,
+      "grad_norm": 0.4306543758497723,
+      "learning_rate": 9.763319866772486e-06,
+      "loss": 0.2757,
+      "step": 3860
+    },
+    {
+      "epoch": 2.2302365839584537,
+      "grad_norm": 0.451979701648703,
+      "learning_rate": 9.738146608730527e-06,
+      "loss": 0.281,
+      "step": 3865
+    },
+    {
+      "epoch": 2.2331217541834967,
+      "grad_norm": 0.39863721704621635,
+      "learning_rate": 9.712975011066633e-06,
+      "loss": 0.2651,
+      "step": 3870
+    },
+    {
+      "epoch": 2.23600692440854,
+      "grad_norm": 0.4431702895497509,
+      "learning_rate": 9.687805233390602e-06,
+      "loss": 0.2875,
+      "step": 3875
+    },
+    {
+      "epoch": 2.2388920946335835,
+      "grad_norm": 0.42463992002267936,
+      "learning_rate": 9.662637435300702e-06,
+      "loss": 0.2747,
+      "step": 3880
+    },
+    {
+      "epoch": 2.2417772648586265,
+      "grad_norm": 0.4421016875554418,
+      "learning_rate": 9.637471776382642e-06,
+      "loss": 0.3014,
+      "step": 3885
+    },
+    {
+      "epoch": 2.24466243508367,
+      "grad_norm": 0.4309775900999432,
+      "learning_rate": 9.612308416208573e-06,
+      "loss": 0.3009,
+      "step": 3890
+    },
+    {
+      "epoch": 2.2475476053087133,
+      "grad_norm": 0.38604226765745275,
+      "learning_rate": 9.587147514336068e-06,
+      "loss": 0.2784,
+      "step": 3895
+    },
+    {
+      "epoch": 2.2504327755337563,
+      "grad_norm": 0.4812969089447562,
+      "learning_rate": 9.561989230307106e-06,
+      "loss": 0.2768,
+      "step": 3900
+    },
+    {
+      "epoch": 2.2533179457587997,
+      "grad_norm": 0.42601904264725454,
+      "learning_rate": 9.536833723647073e-06,
+      "loss": 0.2797,
+      "step": 3905
+    },
+    {
+      "epoch": 2.256203115983843,
+      "grad_norm": 0.4357929404581439,
+      "learning_rate": 9.51168115386374e-06,
+      "loss": 0.2785,
+      "step": 3910
+    },
+    {
+      "epoch": 2.259088286208886,
+      "grad_norm": 0.41501335099687836,
+      "learning_rate": 9.486531680446266e-06,
+      "loss": 0.2674,
+      "step": 3915
+    },
+    {
+      "epoch": 2.2619734564339296,
+      "grad_norm": 0.4259900039138935,
+      "learning_rate": 9.46138546286416e-06,
+      "loss": 0.2659,
+      "step": 3920
+    },
+    {
+      "epoch": 2.264858626658973,
+      "grad_norm": 0.4619671751542806,
+      "learning_rate": 9.436242660566299e-06,
+      "loss": 0.266,
+      "step": 3925
+    },
+    {
+      "epoch": 2.267743796884016,
+      "grad_norm": 0.4284866042642329,
+      "learning_rate": 9.411103432979895e-06,
+      "loss": 0.277,
+      "step": 3930
+    },
+    {
+      "epoch": 2.2706289671090594,
+      "grad_norm": 0.4405127132258486,
+      "learning_rate": 9.385967939509503e-06,
+      "loss": 0.2639,
+      "step": 3935
+    },
+    {
+      "epoch": 2.273514137334103,
+      "grad_norm": 0.41142311165199374,
+      "learning_rate": 9.360836339535997e-06,
+      "loss": 0.2898,
+      "step": 3940
+    },
+    {
+      "epoch": 2.2763993075591458,
+      "grad_norm": 0.41387485994040385,
+      "learning_rate": 9.335708792415554e-06,
+      "loss": 0.2831,
+      "step": 3945
+    },
+    {
+      "epoch": 2.279284477784189,
+      "grad_norm": 0.4321573127781368,
+      "learning_rate": 9.310585457478665e-06,
+      "loss": 0.2667,
+      "step": 3950
+    },
+    {
+      "epoch": 2.2821696480092326,
+      "grad_norm": 0.4184357007573235,
+      "learning_rate": 9.285466494029103e-06,
+      "loss": 0.2725,
+      "step": 3955
+    },
+    {
+      "epoch": 2.2850548182342756,
+      "grad_norm": 0.4338603810054309,
+      "learning_rate": 9.260352061342928e-06,
+      "loss": 0.2788,
+      "step": 3960
+    },
+    {
+      "epoch": 2.287939988459319,
+      "grad_norm": 0.5970591619349905,
+      "learning_rate": 9.235242318667473e-06,
+      "loss": 0.2679,
+      "step": 3965
+    },
+    {
+      "epoch": 2.2908251586843624,
+      "grad_norm": 0.41404764849303466,
+      "learning_rate": 9.21013742522032e-06,
+      "loss": 0.2638,
+      "step": 3970
+    },
+    {
+      "epoch": 2.293710328909406,
+      "grad_norm": 0.44951250598474296,
+      "learning_rate": 9.185037540188319e-06,
+      "loss": 0.2674,
+      "step": 3975
+    },
+    {
+      "epoch": 2.296595499134449,
+      "grad_norm": 0.42037037135886823,
+      "learning_rate": 9.159942822726539e-06,
+      "loss": 0.2713,
+      "step": 3980
+    },
+    {
+      "epoch": 2.2994806693594922,
+      "grad_norm": 0.42502356412209286,
+      "learning_rate": 9.134853431957316e-06,
+      "loss": 0.2648,
+      "step": 3985
+    },
+    {
+      "epoch": 2.3023658395845357,
+      "grad_norm": 0.4216737901866948,
+      "learning_rate": 9.109769526969179e-06,
+      "loss": 0.2724,
+      "step": 3990
+    },
+    {
+      "epoch": 2.3052510098095786,
+      "grad_norm": 0.4167455267572832,
+      "learning_rate": 9.084691266815888e-06,
+      "loss": 0.2766,
+      "step": 3995
+    },
+    {
+      "epoch": 2.308136180034622,
+      "grad_norm": 0.45897408948064594,
+      "learning_rate": 9.059618810515405e-06,
+      "loss": 0.2781,
+      "step": 4000
+    },
+    {
+      "epoch": 2.3110213502596655,
+      "grad_norm": 0.42036299610150296,
+      "learning_rate": 9.034552317048895e-06,
+      "loss": 0.2707,
+      "step": 4005
+    },
+    {
+      "epoch": 2.3139065204847085,
+      "grad_norm": 0.4277783623427705,
+      "learning_rate": 9.009491945359711e-06,
+      "loss": 0.2832,
+      "step": 4010
+    },
+    {
+      "epoch": 2.316791690709752,
+      "grad_norm": 0.421315487309371,
+      "learning_rate": 8.984437854352384e-06,
+      "loss": 0.2754,
+      "step": 4015
+    },
+    {
+      "epoch": 2.3196768609347953,
+      "grad_norm": 0.4657256225278301,
+      "learning_rate": 8.959390202891625e-06,
+      "loss": 0.2912,
+      "step": 4020
+    },
+    {
+      "epoch": 2.3225620311598383,
+      "grad_norm": 0.4272698374442868,
+      "learning_rate": 8.93434914980131e-06,
+      "loss": 0.2662,
+      "step": 4025
+    },
+    {
+      "epoch": 2.3254472013848817,
+      "grad_norm": 0.4198417527704976,
+      "learning_rate": 8.909314853863483e-06,
+      "loss": 0.2727,
+      "step": 4030
+    },
+    {
+      "epoch": 2.328332371609925,
+      "grad_norm": 0.5124921390099292,
+      "learning_rate": 8.884287473817332e-06,
+      "loss": 0.2708,
+      "step": 4035
+    },
+    {
+      "epoch": 2.331217541834968,
+      "grad_norm": 0.42269744019953115,
+      "learning_rate": 8.859267168358198e-06,
+      "loss": 0.2745,
+      "step": 4040
+    },
+    {
+      "epoch": 2.3341027120600115,
+      "grad_norm": 0.42318242166255127,
+      "learning_rate": 8.834254096136561e-06,
+      "loss": 0.2949,
+      "step": 4045
+    },
+    {
+      "epoch": 2.336987882285055,
+      "grad_norm": 0.419851745283628,
+      "learning_rate": 8.809248415757028e-06,
+      "loss": 0.2795,
+      "step": 4050
+    },
+    {
+      "epoch": 2.339873052510098,
+      "grad_norm": 0.44233002428351004,
+      "learning_rate": 8.784250285777361e-06,
+      "loss": 0.2719,
+      "step": 4055
+    },
+    {
+      "epoch": 2.3427582227351413,
+      "grad_norm": 0.396083210912282,
+      "learning_rate": 8.759259864707415e-06,
+      "loss": 0.2697,
+      "step": 4060
+    },
+    {
+      "epoch": 2.3456433929601848,
+      "grad_norm": 0.47094149055802703,
+      "learning_rate": 8.73427731100818e-06,
+      "loss": 0.2724,
+      "step": 4065
+    },
+    {
+      "epoch": 2.3485285631852277,
+      "grad_norm": 0.42603022428645243,
+      "learning_rate": 8.70930278309076e-06,
+      "loss": 0.2654,
+      "step": 4070
+    },
+    {
+      "epoch": 2.351413733410271,
+      "grad_norm": 0.4379509456011344,
+      "learning_rate": 8.684336439315362e-06,
+      "loss": 0.2786,
+      "step": 4075
+    },
+    {
+      "epoch": 2.3542989036353146,
+      "grad_norm": 0.426972682822357,
+      "learning_rate": 8.659378437990305e-06,
+      "loss": 0.2852,
+      "step": 4080
+    },
+    {
+      "epoch": 2.357184073860358,
+      "grad_norm": 0.42210417306175224,
+      "learning_rate": 8.634428937371008e-06,
+      "loss": 0.2777,
+      "step": 4085
+    },
+    {
+      "epoch": 2.360069244085401,
+      "grad_norm": 0.4490344314229441,
+      "learning_rate": 8.609488095658987e-06,
+      "loss": 0.2732,
+      "step": 4090
+    },
+    {
+      "epoch": 2.3629544143104444,
+      "grad_norm": 0.44262847306870384,
+      "learning_rate": 8.584556071000845e-06,
+      "loss": 0.2724,
+      "step": 4095
+    },
+    {
+      "epoch": 2.365839584535488,
+      "grad_norm": 0.43681619099011737,
+      "learning_rate": 8.559633021487298e-06,
+      "loss": 0.291,
+      "step": 4100
+    },
+    {
+      "epoch": 2.368724754760531,
+      "grad_norm": 0.669470605388525,
+      "learning_rate": 8.53471910515213e-06,
+      "loss": 0.29,
+      "step": 4105
+    },
+    {
+      "epoch": 2.371609924985574,
+      "grad_norm": 0.4872795075157814,
+      "learning_rate": 8.509814479971226e-06,
+      "loss": 0.293,
+      "step": 4110
+    },
+    {
+      "epoch": 2.3744950952106176,
+      "grad_norm": 0.4330590098492678,
+      "learning_rate": 8.484919303861555e-06,
+      "loss": 0.2794,
+      "step": 4115
+    },
+    {
+      "epoch": 2.3773802654356606,
+      "grad_norm": 0.42534287206042853,
+      "learning_rate": 8.460033734680158e-06,
+      "loss": 0.2766,
+      "step": 4120
+    },
+    {
+      "epoch": 2.380265435660704,
+      "grad_norm": 0.43174048146055444,
+      "learning_rate": 8.43515793022318e-06,
+      "loss": 0.2797,
+      "step": 4125
+    },
+    {
+      "epoch": 2.3831506058857475,
+      "grad_norm": 0.4429583272831574,
+      "learning_rate": 8.410292048224838e-06,
+      "loss": 0.269,
+      "step": 4130
+    },
+    {
+      "epoch": 2.3860357761107904,
+      "grad_norm": 0.41491168931306777,
+      "learning_rate": 8.38543624635643e-06,
+      "loss": 0.2771,
+      "step": 4135
+    },
+    {
+      "epoch": 2.388920946335834,
+      "grad_norm": 0.4490843763025565,
+      "learning_rate": 8.360590682225345e-06,
+      "loss": 0.2775,
+      "step": 4140
+    },
+    {
+      "epoch": 2.3918061165608773,
+      "grad_norm": 0.4162339442858775,
+      "learning_rate": 8.335755513374044e-06,
+      "loss": 0.2725,
+      "step": 4145
+    },
+    {
+      "epoch": 2.3946912867859202,
+      "grad_norm": 0.3966357439538777,
+      "learning_rate": 8.310930897279088e-06,
+      "loss": 0.2824,
+      "step": 4150
+    },
+    {
+      "epoch": 2.3975764570109637,
+      "grad_norm": 0.43243463473106125,
+      "learning_rate": 8.286116991350113e-06,
+      "loss": 0.273,
+      "step": 4155
+    },
+    {
+      "epoch": 2.400461627236007,
+      "grad_norm": 0.40628353100856796,
+      "learning_rate": 8.261313952928853e-06,
+      "loss": 0.2697,
+      "step": 4160
+    },
+    {
+      "epoch": 2.40334679746105,
+      "grad_norm": 0.41821368197595044,
+      "learning_rate": 8.236521939288116e-06,
+      "loss": 0.2722,
+      "step": 4165
+    },
+    {
+      "epoch": 2.4062319676860935,
+      "grad_norm": 0.43315603245507267,
+      "learning_rate": 8.211741107630817e-06,
+      "loss": 0.2811,
+      "step": 4170
+    },
+    {
+      "epoch": 2.409117137911137,
+      "grad_norm": 0.4192983792279987,
+      "learning_rate": 8.186971615088967e-06,
+      "loss": 0.27,
+      "step": 4175
+    },
+    {
+      "epoch": 2.41200230813618,
+      "grad_norm": 0.4228869690317952,
+      "learning_rate": 8.162213618722679e-06,
+      "loss": 0.2936,
+      "step": 4180
+    },
+    {
+      "epoch": 2.4148874783612233,
+      "grad_norm": 0.4264461059873532,
+      "learning_rate": 8.137467275519156e-06,
+      "loss": 0.255,
+      "step": 4185
+    },
+    {
+      "epoch": 2.4177726485862667,
+      "grad_norm": 0.42183959270214744,
+      "learning_rate": 8.112732742391722e-06,
+      "loss": 0.265,
+      "step": 4190
+    },
+    {
+      "epoch": 2.4206578188113097,
+      "grad_norm": 0.440409389917108,
+      "learning_rate": 8.088010176178816e-06,
+      "loss": 0.2895,
+      "step": 4195
+    },
+    {
+      "epoch": 2.423542989036353,
+      "grad_norm": 0.4190921915499105,
+      "learning_rate": 8.063299733642991e-06,
+      "loss": 0.268,
+      "step": 4200
+    },
+    {
+      "epoch": 2.4264281592613965,
+      "grad_norm": 0.4370598207669186,
+      "learning_rate": 8.03860157146993e-06,
+      "loss": 0.2764,
+      "step": 4205
+    },
+    {
+      "epoch": 2.4293133294864395,
+      "grad_norm": 0.4607794338360012,
+      "learning_rate": 8.013915846267439e-06,
+      "loss": 0.2863,
+      "step": 4210
+    },
+    {
+      "epoch": 2.432198499711483,
+      "grad_norm": 0.4378000785386685,
+      "learning_rate": 7.98924271456447e-06,
+      "loss": 0.2747,
+      "step": 4215
+    },
+    {
+      "epoch": 2.4350836699365264,
+      "grad_norm": 0.4598949217946755,
+      "learning_rate": 7.964582332810122e-06,
+      "loss": 0.2756,
+      "step": 4220
+    },
+    {
+      "epoch": 2.4379688401615693,
+      "grad_norm": 0.4305567337108496,
+      "learning_rate": 7.939934857372646e-06,
+      "loss": 0.2734,
+      "step": 4225
+    },
+    {
+      "epoch": 2.4408540103866128,
+      "grad_norm": 0.42896614513275355,
+      "learning_rate": 7.915300444538458e-06,
+      "loss": 0.2704,
+      "step": 4230
+    },
+    {
+      "epoch": 2.443739180611656,
+      "grad_norm": 0.4104247108771218,
+      "learning_rate": 7.89067925051114e-06,
+      "loss": 0.2683,
+      "step": 4235
+    },
+    {
+      "epoch": 2.446624350836699,
+      "grad_norm": 0.42265110955709223,
+      "learning_rate": 7.866071431410458e-06,
+      "loss": 0.2834,
+      "step": 4240
+    },
+    {
+      "epoch": 2.4495095210617426,
+      "grad_norm": 0.4025030251156565,
+      "learning_rate": 7.841477143271374e-06,
+      "loss": 0.2719,
+      "step": 4245
+    },
+    {
+      "epoch": 2.452394691286786,
+      "grad_norm": 0.4347245013374062,
+      "learning_rate": 7.816896542043048e-06,
+      "loss": 0.2849,
+      "step": 4250
+    },
+    {
+      "epoch": 2.455279861511829,
+      "grad_norm": 0.3970293249450069,
+      "learning_rate": 7.792329783587853e-06,
+      "loss": 0.281,
+      "step": 4255
+    },
+    {
+      "epoch": 2.4581650317368724,
+      "grad_norm": 0.5089714627565546,
+      "learning_rate": 7.767777023680384e-06,
+      "loss": 0.2886,
+      "step": 4260
+    },
+    {
+      "epoch": 2.461050201961916,
+      "grad_norm": 0.42036661292632566,
+      "learning_rate": 7.743238418006476e-06,
+      "loss": 0.274,
+      "step": 4265
+    },
+    {
+      "epoch": 2.463935372186959,
+      "grad_norm": 0.41995706488380274,
+      "learning_rate": 7.71871412216222e-06,
+      "loss": 0.2834,
+      "step": 4270
+    },
+    {
+      "epoch": 2.466820542412002,
+      "grad_norm": 0.405500972649185,
+      "learning_rate": 7.694204291652962e-06,
+      "loss": 0.2723,
+      "step": 4275
+    },
+    {
+      "epoch": 2.4697057126370456,
+      "grad_norm": 0.47474988784176503,
+      "learning_rate": 7.669709081892324e-06,
+      "loss": 0.276,
+      "step": 4280
+    },
+    {
+      "epoch": 2.472590882862089,
+      "grad_norm": 0.4208990452736956,
+      "learning_rate": 7.645228648201225e-06,
+      "loss": 0.2686,
+      "step": 4285
+    },
+    {
+      "epoch": 2.475476053087132,
+      "grad_norm": 0.5125940516643855,
+      "learning_rate": 7.62076314580689e-06,
+      "loss": 0.2791,
+      "step": 4290
+    },
+    {
+      "epoch": 2.4783612233121755,
+      "grad_norm": 0.43397963005803125,
+      "learning_rate": 7.596312729841868e-06,
+      "loss": 0.2714,
+      "step": 4295
+    },
+    {
+      "epoch": 2.481246393537219,
+      "grad_norm": 0.43916382261638937,
+      "learning_rate": 7.571877555343039e-06,
+      "loss": 0.2953,
+      "step": 4300
+    },
+    {
+      "epoch": 2.484131563762262,
+      "grad_norm": 0.42968851926015045,
+      "learning_rate": 7.547457777250649e-06,
+      "loss": 0.2915,
+      "step": 4305
+    },
+    {
+      "epoch": 2.4870167339873053,
+      "grad_norm": 0.4334658833975296,
+      "learning_rate": 7.523053550407308e-06,
+      "loss": 0.2781,
+      "step": 4310
+    },
+    {
+      "epoch": 2.4899019042123487,
+      "grad_norm": 0.442518451619476,
+      "learning_rate": 7.498665029557028e-06,
+      "loss": 0.2795,
+      "step": 4315
+    },
+    {
+      "epoch": 2.4927870744373917,
+      "grad_norm": 0.4381590289442612,
+      "learning_rate": 7.474292369344226e-06,
+      "loss": 0.2641,
+      "step": 4320
+    },
+    {
+      "epoch": 2.495672244662435,
+      "grad_norm": 0.44001237345403327,
+      "learning_rate": 7.449935724312745e-06,
+      "loss": 0.2688,
+      "step": 4325
+    },
+    {
+      "epoch": 2.4985574148874785,
+      "grad_norm": 0.43261375909491934,
+      "learning_rate": 7.42559524890488e-06,
+      "loss": 0.2791,
+      "step": 4330
+    },
+    {
+      "epoch": 2.5014425851125215,
+      "grad_norm": 0.41821164862411353,
+      "learning_rate": 7.401271097460401e-06,
+      "loss": 0.2753,
+      "step": 4335
+    },
+    {
+      "epoch": 2.504327755337565,
+      "grad_norm": 0.4284156665016015,
+      "learning_rate": 7.376963424215568e-06,
+      "loss": 0.2916,
+      "step": 4340
+    },
+    {
+      "epoch": 2.5072129255626083,
+      "grad_norm": 0.42595494161348285,
+      "learning_rate": 7.352672383302153e-06,
+      "loss": 0.2587,
+      "step": 4345
+    },
+    {
+      "epoch": 2.5100980957876513,
+      "grad_norm": 0.4234233587645987,
+      "learning_rate": 7.328398128746463e-06,
+      "loss": 0.2723,
+      "step": 4350
+    },
+    {
+      "epoch": 2.5129832660126947,
+      "grad_norm": 0.4206786678316223,
+      "learning_rate": 7.304140814468365e-06,
+      "loss": 0.2737,
+      "step": 4355
+    },
+    {
+      "epoch": 2.515868436237738,
+      "grad_norm": 0.4421492786532555,
+      "learning_rate": 7.27990059428032e-06,
+      "loss": 0.2871,
+      "step": 4360
+    },
+    {
+      "epoch": 2.5187536064627816,
+      "grad_norm": 0.4071794979099346,
+      "learning_rate": 7.255677621886388e-06,
+      "loss": 0.2669,
+      "step": 4365
+    },
+    {
+      "epoch": 2.5216387766878245,
+      "grad_norm": 0.4085099532665915,
+      "learning_rate": 7.23147205088126e-06,
+      "loss": 0.2749,
+      "step": 4370
+    },
+    {
+      "epoch": 2.524523946912868,
+      "grad_norm": 0.4540968719545943,
+      "learning_rate": 7.207284034749294e-06,
+      "loss": 0.2784,
+      "step": 4375
+    },
+    {
+      "epoch": 2.5274091171379114,
+      "grad_norm": 0.43868950738012735,
+      "learning_rate": 7.183113726863529e-06,
+      "loss": 0.2817,
+      "step": 4380
+    },
+    {
+      "epoch": 2.5302942873629544,
+      "grad_norm": 0.5603667780494723,
+      "learning_rate": 7.15896128048473e-06,
+      "loss": 0.2779,
+      "step": 4385
+    },
+    {
+      "epoch": 2.533179457587998,
+      "grad_norm": 0.4643684780575487,
+      "learning_rate": 7.134826848760392e-06,
+      "loss": 0.2704,
+      "step": 4390
+    },
+    {
+      "epoch": 2.536064627813041,
+      "grad_norm": 0.4311580857729655,
+      "learning_rate": 7.110710584723785e-06,
+      "loss": 0.2773,
+      "step": 4395
+    },
+    {
+      "epoch": 2.538949798038084,
+      "grad_norm": 0.41665610893096905,
+      "learning_rate": 7.086612641292985e-06,
+      "loss": 0.2697,
+      "step": 4400
+    },
+    {
+      "epoch": 2.5418349682631276,
+      "grad_norm": 0.4448150891591054,
+      "learning_rate": 7.062533171269887e-06,
+      "loss": 0.2679,
+      "step": 4405
+    },
+    {
+      "epoch": 2.544720138488171,
+      "grad_norm": 0.4485899811392423,
+      "learning_rate": 7.038472327339268e-06,
+      "loss": 0.2624,
+      "step": 4410
+    },
+    {
+      "epoch": 2.547605308713214,
+      "grad_norm": 0.4694083142624704,
+      "learning_rate": 7.0144302620677804e-06,
+      "loss": 0.2671,
+      "step": 4415
+    },
+    {
+      "epoch": 2.5504904789382574,
+      "grad_norm": 0.44085510653638454,
+      "learning_rate": 6.990407127903015e-06,
+      "loss": 0.2741,
+      "step": 4420
+    },
+    {
+      "epoch": 2.553375649163301,
+      "grad_norm": 0.4331688809785418,
+      "learning_rate": 6.966403077172521e-06,
+      "loss": 0.2784,
+      "step": 4425
+    },
+    {
+      "epoch": 2.556260819388344,
+      "grad_norm": 0.42326748215143545,
+      "learning_rate": 6.942418262082833e-06,
+      "loss": 0.2803,
+      "step": 4430
+    },
+    {
+      "epoch": 2.5591459896133872,
+      "grad_norm": 0.4579702510573031,
+      "learning_rate": 6.91845283471853e-06,
+      "loss": 0.2629,
+      "step": 4435
+    },
+    {
+      "epoch": 2.5620311598384307,
+      "grad_norm": 0.41574752352195643,
+      "learning_rate": 6.894506947041246e-06,
+      "loss": 0.2606,
+      "step": 4440
+    },
+    {
+      "epoch": 2.5649163300634736,
+      "grad_norm": 0.4365002129279704,
+      "learning_rate": 6.870580750888722e-06,
+      "loss": 0.2724,
+      "step": 4445
+    },
+    {
+      "epoch": 2.567801500288517,
+      "grad_norm": 0.4399394455174853,
+      "learning_rate": 6.846674397973825e-06,
+      "loss": 0.2649,
+      "step": 4450
+    },
+    {
+      "epoch": 2.5706866705135605,
+      "grad_norm": 0.447371975699018,
+      "learning_rate": 6.822788039883621e-06,
+      "loss": 0.2684,
+      "step": 4455
+    },
+    {
+      "epoch": 2.5735718407386035,
+      "grad_norm": 0.3961735893626953,
+      "learning_rate": 6.7989218280783686e-06,
+      "loss": 0.2587,
+      "step": 4460
+    },
+    {
+      "epoch": 2.576457010963647,
+      "grad_norm": 0.44985375940450545,
+      "learning_rate": 6.775075913890597e-06,
+      "loss": 0.2613,
+      "step": 4465
+    },
+    {
+      "epoch": 2.5793421811886903,
+      "grad_norm": 0.4488817173284117,
+      "learning_rate": 6.751250448524123e-06,
+      "loss": 0.2814,
+      "step": 4470
+    },
+    {
+      "epoch": 2.5822273514137333,
+      "grad_norm": 0.5139233580575341,
+      "learning_rate": 6.727445583053096e-06,
+      "loss": 0.2707,
+      "step": 4475
+    },
+    {
+      "epoch": 2.5851125216387767,
+      "grad_norm": 0.4742477912373411,
+      "learning_rate": 6.703661468421065e-06,
+      "loss": 0.2971,
+      "step": 4480
+    },
+    {
+      "epoch": 2.58799769186382,
+      "grad_norm": 0.4354163078529058,
+      "learning_rate": 6.679898255439974e-06,
+      "loss": 0.2796,
+      "step": 4485
+    },
+    {
+      "epoch": 2.590882862088863,
+      "grad_norm": 0.42132555668081595,
+      "learning_rate": 6.656156094789253e-06,
+      "loss": 0.2844,
+      "step": 4490
+    },
+    {
+      "epoch": 2.5937680323139065,
+      "grad_norm": 0.43721072545918765,
+      "learning_rate": 6.632435137014823e-06,
+      "loss": 0.2778,
+      "step": 4495
+    },
+    {
+      "epoch": 2.59665320253895,
+      "grad_norm": 0.405910083382628,
+      "learning_rate": 6.608735532528178e-06,
+      "loss": 0.2977,
+      "step": 4500
+    },
+    {
+      "epoch": 2.599538372763993,
+      "grad_norm": 0.41260341077774265,
+      "learning_rate": 6.585057431605406e-06,
+      "loss": 0.2686,
+      "step": 4505
+    },
+    {
+      "epoch": 2.6024235429890363,
+      "grad_norm": 0.44802578482097516,
+      "learning_rate": 6.561400984386243e-06,
+      "loss": 0.2809,
+      "step": 4510
+    },
+    {
+      "epoch": 2.6053087132140798,
+      "grad_norm": 0.4321051913017041,
+      "learning_rate": 6.537766340873125e-06,
+      "loss": 0.2842,
+      "step": 4515
+    },
+    {
+      "epoch": 2.6081938834391227,
+      "grad_norm": 0.4775646961550856,
+      "learning_rate": 6.514153650930228e-06,
+      "loss": 0.2702,
+      "step": 4520
+    },
+    {
+      "epoch": 2.611079053664166,
+      "grad_norm": 0.4330851918875865,
+      "learning_rate": 6.4905630642825245e-06,
+      "loss": 0.261,
+      "step": 4525
+    },
+    {
+      "epoch": 2.6139642238892096,
+      "grad_norm": 0.46362229239200453,
+      "learning_rate": 6.466994730514842e-06,
+      "loss": 0.2815,
+      "step": 4530
+    },
+    {
+      "epoch": 2.6168493941142525,
+      "grad_norm": 0.41520002516181664,
+      "learning_rate": 6.443448799070895e-06,
+      "loss": 0.2786,
+      "step": 4535
+    },
+    {
+      "epoch": 2.619734564339296,
+      "grad_norm": 0.45005348838742215,
+      "learning_rate": 6.419925419252353e-06,
+      "loss": 0.2758,
+      "step": 4540
+    },
+    {
+      "epoch": 2.6226197345643394,
+      "grad_norm": 0.56725726849741,
+      "learning_rate": 6.39642474021788e-06,
+      "loss": 0.2641,
+      "step": 4545
+    },
+    {
+      "epoch": 2.6255049047893824,
+      "grad_norm": 0.49336912135631245,
+      "learning_rate": 6.3729469109822115e-06,
+      "loss": 0.2803,
+      "step": 4550
+    },
+    {
+      "epoch": 2.628390075014426,
+      "grad_norm": 0.4407248177296093,
+      "learning_rate": 6.349492080415187e-06,
+      "loss": 0.2849,
+      "step": 4555
+    },
+    {
+      "epoch": 2.631275245239469,
+      "grad_norm": 0.41465469511645586,
+      "learning_rate": 6.326060397240814e-06,
+      "loss": 0.275,
+      "step": 4560
+    },
+    {
+      "epoch": 2.634160415464512,
+      "grad_norm": 0.4463897975975723,
+      "learning_rate": 6.3026520100363276e-06,
+      "loss": 0.2763,
+      "step": 4565
+    },
+    {
+      "epoch": 2.6370455856895556,
+      "grad_norm": 0.4179248454832217,
+      "learning_rate": 6.279267067231242e-06,
+      "loss": 0.2759,
+      "step": 4570
+    },
+    {
+      "epoch": 2.639930755914599,
+      "grad_norm": 0.4482183921648813,
+      "learning_rate": 6.255905717106421e-06,
+      "loss": 0.2809,
+      "step": 4575
+    },
+    {
+      "epoch": 2.642815926139642,
+      "grad_norm": 0.45808207591686173,
+      "learning_rate": 6.232568107793127e-06,
+      "loss": 0.2891,
+      "step": 4580
+    },
+    {
+      "epoch": 2.6457010963646854,
+      "grad_norm": 0.43079637132281384,
+      "learning_rate": 6.209254387272085e-06,
+      "loss": 0.2846,
+      "step": 4585
+    },
+    {
+      "epoch": 2.648586266589729,
+      "grad_norm": 0.43896911522538273,
+      "learning_rate": 6.185964703372538e-06,
+      "loss": 0.2709,
+      "step": 4590
+    },
+    {
+      "epoch": 2.651471436814772,
+      "grad_norm": 0.41937040243151874,
+      "learning_rate": 6.162699203771324e-06,
+      "loss": 0.2698,
+      "step": 4595
+    },
+    {
+      "epoch": 2.6543566070398152,
+      "grad_norm": 0.436960743544877,
+      "learning_rate": 6.139458035991932e-06,
+      "loss": 0.2571,
+      "step": 4600
+    },
+    {
+      "epoch": 2.6572417772648587,
+      "grad_norm": 0.43802501221211554,
+      "learning_rate": 6.116241347403564e-06,
+      "loss": 0.2845,
+      "step": 4605
+    },
+    {
+      "epoch": 2.6601269474899016,
+      "grad_norm": 0.4586510827028596,
+      "learning_rate": 6.093049285220198e-06,
+      "loss": 0.2715,
+      "step": 4610
+    },
+    {
+      "epoch": 2.663012117714945,
+      "grad_norm": 0.4206625935899696,
+      "learning_rate": 6.069881996499664e-06,
+      "loss": 0.2725,
+      "step": 4615
+    },
+    {
+      "epoch": 2.6658972879399885,
+      "grad_norm": 0.4183330676839759,
+      "learning_rate": 6.046739628142705e-06,
+      "loss": 0.2732,
+      "step": 4620
+    },
+    {
+      "epoch": 2.668782458165032,
+      "grad_norm": 0.4380908491163701,
+      "learning_rate": 6.023622326892051e-06,
+      "loss": 0.2763,
+      "step": 4625
+    },
+    {
+      "epoch": 2.671667628390075,
+      "grad_norm": 0.44598178634221547,
+      "learning_rate": 6.000530239331481e-06,
+      "loss": 0.2794,
+      "step": 4630
+    },
+    {
+      "epoch": 2.6745527986151183,
+      "grad_norm": 0.43851972482947554,
+      "learning_rate": 5.977463511884898e-06,
+      "loss": 0.2679,
+      "step": 4635
+    },
+    {
+      "epoch": 2.6774379688401617,
+      "grad_norm": 0.4173251500861273,
+      "learning_rate": 5.954422290815396e-06,
+      "loss": 0.2757,
+      "step": 4640
+    },
+    {
+      "epoch": 2.6803231390652047,
+      "grad_norm": 0.42797667265485484,
+      "learning_rate": 5.931406722224344e-06,
+      "loss": 0.2673,
+      "step": 4645
+    },
+    {
+      "epoch": 2.683208309290248,
+      "grad_norm": 0.41916940809600545,
+      "learning_rate": 5.908416952050453e-06,
+      "loss": 0.2587,
+      "step": 4650
+    },
+    {
+      "epoch": 2.6860934795152915,
+      "grad_norm": 0.40342454442056763,
+      "learning_rate": 5.885453126068836e-06,
+      "loss": 0.2575,
+      "step": 4655
+    },
+    {
+      "epoch": 2.6889786497403345,
+      "grad_norm": 0.42304042640556994,
+      "learning_rate": 5.862515389890111e-06,
+      "loss": 0.2743,
+      "step": 4660
+    },
+    {
+      "epoch": 2.691863819965378,
+      "grad_norm": 0.4459071254483481,
+      "learning_rate": 5.839603888959455e-06,
+      "loss": 0.2764,
+      "step": 4665
+    },
+    {
+      "epoch": 2.6947489901904214,
+      "grad_norm": 0.41434533982836086,
+      "learning_rate": 5.8167187685557045e-06,
+      "loss": 0.2631,
+      "step": 4670
+    },
+    {
+      "epoch": 2.6976341604154648,
+      "grad_norm": 0.44266182668844456,
+      "learning_rate": 5.7938601737904e-06,
+      "loss": 0.2709,
+      "step": 4675
+    },
+    {
+      "epoch": 2.7005193306405078,
+      "grad_norm": 0.4225234227671778,
+      "learning_rate": 5.771028249606906e-06,
+      "loss": 0.2736,
+      "step": 4680
+    },
+    {
+      "epoch": 2.703404500865551,
+      "grad_norm": 0.43317051916217664,
+      "learning_rate": 5.74822314077946e-06,
+      "loss": 0.2812,
+      "step": 4685
+    },
+    {
+      "epoch": 2.7062896710905946,
+      "grad_norm": 0.4224513310713421,
+      "learning_rate": 5.725444991912266e-06,
+      "loss": 0.2657,
+      "step": 4690
+    },
+    {
+      "epoch": 2.7091748413156376,
+      "grad_norm": 0.4314258773445562,
+      "learning_rate": 5.702693947438589e-06,
+      "loss": 0.271,
+      "step": 4695
+    },
+    {
+      "epoch": 2.712060011540681,
+      "grad_norm": 0.4422321485260852,
+      "learning_rate": 5.679970151619823e-06,
+      "loss": 0.2711,
+      "step": 4700
+    },
+    {
+      "epoch": 2.7149451817657244,
+      "grad_norm": 0.44035975780616426,
+      "learning_rate": 5.6572737485445804e-06,
+      "loss": 0.2874,
+      "step": 4705
+    },
+    {
+      "epoch": 2.7178303519907674,
+      "grad_norm": 0.44924301832440483,
+      "learning_rate": 5.634604882127771e-06,
+      "loss": 0.2703,
+      "step": 4710
+    },
+    {
+      "epoch": 2.720715522215811,
+      "grad_norm": 0.4107307992041668,
+      "learning_rate": 5.611963696109726e-06,
+      "loss": 0.2729,
+      "step": 4715
+    },
+    {
+      "epoch": 2.7236006924408542,
+      "grad_norm": 0.4372366602675938,
+      "learning_rate": 5.58935033405523e-06,
+      "loss": 0.276,
+      "step": 4720
+    },
+    {
+      "epoch": 2.726485862665897,
+      "grad_norm": 0.45334235025453323,
+      "learning_rate": 5.5667649393526535e-06,
+      "loss": 0.2801,
+      "step": 4725
+    },
+    {
+      "epoch": 2.7293710328909406,
+      "grad_norm": 0.4201853295737695,
+      "learning_rate": 5.544207655213021e-06,
+      "loss": 0.2831,
+      "step": 4730
+    },
+    {
+      "epoch": 2.732256203115984,
+      "grad_norm": 0.44159486930994,
+      "learning_rate": 5.521678624669121e-06,
+      "loss": 0.2644,
+      "step": 4735
+    },
+    {
+      "epoch": 2.735141373341027,
+      "grad_norm": 0.45477978259519153,
+      "learning_rate": 5.499177990574591e-06,
+      "loss": 0.2856,
+      "step": 4740
+    },
+    {
+      "epoch": 2.7380265435660704,
+      "grad_norm": 0.4329706475468657,
+      "learning_rate": 5.4767058956029955e-06,
+      "loss": 0.2757,
+      "step": 4745
+    },
+    {
+      "epoch": 2.740911713791114,
+      "grad_norm": 0.46147021358014734,
+      "learning_rate": 5.454262482246954e-06,
+      "loss": 0.2816,
+      "step": 4750
+    },
+    {
+      "epoch": 2.743796884016157,
+      "grad_norm": 0.4661777717367274,
+      "learning_rate": 5.431847892817208e-06,
+      "loss": 0.2715,
+      "step": 4755
+    },
+    {
+      "epoch": 2.7466820542412003,
+      "grad_norm": 0.42606918880864486,
+      "learning_rate": 5.40946226944173e-06,
+      "loss": 0.2785,
+      "step": 4760
+    },
+    {
+      "epoch": 2.7495672244662437,
+      "grad_norm": 0.44527079904346323,
+      "learning_rate": 5.387105754064829e-06,
+      "loss": 0.2776,
+      "step": 4765
+    },
+    {
+      "epoch": 2.7524523946912867,
+      "grad_norm": 0.4377246196123211,
+      "learning_rate": 5.364778488446247e-06,
+      "loss": 0.2753,
+      "step": 4770
+    },
+    {
+      "epoch": 2.75533756491633,
+      "grad_norm": 0.44386684339819055,
+      "learning_rate": 5.342480614160247e-06,
+      "loss": 0.2859,
+      "step": 4775
+    },
+    {
+      "epoch": 2.7582227351413735,
+      "grad_norm": 0.4235638405047958,
+      "learning_rate": 5.320212272594725e-06,
+      "loss": 0.2811,
+      "step": 4780
+    },
+    {
+      "epoch": 2.7611079053664165,
+      "grad_norm": 0.45140518863698265,
+      "learning_rate": 5.297973604950321e-06,
+      "loss": 0.2704,
+      "step": 4785
+    },
+    {
+      "epoch": 2.76399307559146,
+      "grad_norm": 0.4079388286223738,
+      "learning_rate": 5.275764752239523e-06,
+      "loss": 0.2625,
+      "step": 4790
+    },
+    {
+      "epoch": 2.7668782458165033,
+      "grad_norm": 0.41869006995290337,
+      "learning_rate": 5.2535858552857474e-06,
+      "loss": 0.2603,
+      "step": 4795
+    },
+    {
+      "epoch": 2.7697634160415463,
+      "grad_norm": 0.4333081525779454,
+      "learning_rate": 5.231437054722477e-06,
+      "loss": 0.2908,
+      "step": 4800
+    },
+    {
+      "epoch": 2.7726485862665897,
+      "grad_norm": 0.46442418499425026,
+      "learning_rate": 5.209318490992355e-06,
+      "loss": 0.2711,
+      "step": 4805
+    },
+    {
+      "epoch": 2.775533756491633,
+      "grad_norm": 0.42235720398392995,
+      "learning_rate": 5.187230304346304e-06,
+      "loss": 0.2874,
+      "step": 4810
+    },
+    {
+      "epoch": 2.778418926716676,
+      "grad_norm": 0.42546995948748306,
+      "learning_rate": 5.1651726348426146e-06,
+      "loss": 0.2845,
+      "step": 4815
+    },
+    {
+      "epoch": 2.7813040969417195,
+      "grad_norm": 0.4458920282497016,
+      "learning_rate": 5.143145622346089e-06,
+      "loss": 0.2656,
+      "step": 4820
+    },
+    {
+      "epoch": 2.784189267166763,
+      "grad_norm": 0.4140358163182573,
+      "learning_rate": 5.121149406527123e-06,
+      "loss": 0.2674,
+      "step": 4825
+    },
+    {
+      "epoch": 2.787074437391806,
+      "grad_norm": 0.4300310193339041,
+      "learning_rate": 5.099184126860837e-06,
+      "loss": 0.266,
+      "step": 4830
+    },
+    {
+      "epoch": 2.7899596076168494,
+      "grad_norm": 0.8717864481697682,
+      "learning_rate": 5.077249922626197e-06,
+      "loss": 0.2794,
+      "step": 4835
+    },
+    {
+      "epoch": 2.792844777841893,
+      "grad_norm": 0.43179815522266113,
+      "learning_rate": 5.055346932905125e-06,
+      "loss": 0.2821,
+      "step": 4840
+    },
+    {
+      "epoch": 2.7957299480669358,
+      "grad_norm": 0.438127032880629,
+      "learning_rate": 5.033475296581606e-06,
+      "loss": 0.2749,
+      "step": 4845
+    },
+    {
+      "epoch": 2.798615118291979,
+      "grad_norm": 0.42678461004761115,
+      "learning_rate": 5.011635152340816e-06,
+      "loss": 0.2724,
+      "step": 4850
+    },
+    {
+      "epoch": 2.8015002885170226,
+      "grad_norm": 0.4403224341518811,
+      "learning_rate": 4.9898266386682534e-06,
+      "loss": 0.2848,
+      "step": 4855
+    },
+    {
+      "epoch": 2.8043854587420656,
+      "grad_norm": 0.4551147345010523,
+      "learning_rate": 4.968049893848854e-06,
+      "loss": 0.2637,
+      "step": 4860
+    },
+    {
+      "epoch": 2.807270628967109,
+      "grad_norm": 0.43423817098138445,
+      "learning_rate": 4.9463050559660985e-06,
+      "loss": 0.268,
+      "step": 4865
+    },
+    {
+      "epoch": 2.8101557991921524,
+      "grad_norm": 0.42455505455856984,
+      "learning_rate": 4.924592262901154e-06,
+      "loss": 0.2681,
+      "step": 4870
+    },
+    {
+      "epoch": 2.8130409694171954,
+      "grad_norm": 0.4530758791142782,
+      "learning_rate": 4.9029116523320035e-06,
+      "loss": 0.268,
+      "step": 4875
+    },
+    {
+      "epoch": 2.815926139642239,
+      "grad_norm": 0.4249843125062271,
+      "learning_rate": 4.881263361732552e-06,
+      "loss": 0.2753,
+      "step": 4880
+    },
+    {
+      "epoch": 2.8188113098672822,
+      "grad_norm": 0.4472656624417374,
+      "learning_rate": 4.859647528371784e-06,
+      "loss": 0.283,
+      "step": 4885
+    },
+    {
+      "epoch": 2.821696480092325,
+      "grad_norm": 0.42578297771011603,
+      "learning_rate": 4.838064289312862e-06,
+      "loss": 0.2898,
+      "step": 4890
+    },
+    {
+      "epoch": 2.8245816503173686,
+      "grad_norm": 0.4436993117564527,
+      "learning_rate": 4.816513781412282e-06,
+      "loss": 0.2701,
+      "step": 4895
+    },
+    {
+      "epoch": 2.827466820542412,
+      "grad_norm": 0.4621541772278322,
+      "learning_rate": 4.7949961413189885e-06,
+      "loss": 0.2771,
+      "step": 4900
+    },
+    {
+      "epoch": 2.830351990767455,
+      "grad_norm": 0.4266188005044788,
+      "learning_rate": 4.77351150547352e-06,
+      "loss": 0.2803,
+      "step": 4905
+    },
+    {
+      "epoch": 2.8332371609924984,
+      "grad_norm": 0.45663638676525325,
+      "learning_rate": 4.752060010107145e-06,
+      "loss": 0.2826,
+      "step": 4910
+    },
+    {
+      "epoch": 2.836122331217542,
+      "grad_norm": 0.42586386429655265,
+      "learning_rate": 4.730641791240981e-06,
+      "loss": 0.2633,
+      "step": 4915
+    },
+    {
+      "epoch": 2.839007501442585,
+      "grad_norm": 0.42216429186560445,
+      "learning_rate": 4.709256984685147e-06,
+      "loss": 0.2835,
+      "step": 4920
+    },
+    {
+      "epoch": 2.8418926716676283,
+      "grad_norm": 0.45151376072343974,
+      "learning_rate": 4.687905726037902e-06,
+      "loss": 0.2716,
+      "step": 4925
+    },
+    {
+      "epoch": 2.8447778418926717,
+      "grad_norm": 0.40227119407723355,
+      "learning_rate": 4.666588150684786e-06,
+      "loss": 0.2744,
+      "step": 4930
+    },
+    {
+      "epoch": 2.847663012117715,
+      "grad_norm": 0.41851949935896066,
+      "learning_rate": 4.6453043937977496e-06,
+      "loss": 0.2771,
+      "step": 4935
+    },
+    {
+      "epoch": 2.850548182342758,
+      "grad_norm": 0.40933746114854325,
+      "learning_rate": 4.624054590334304e-06,
+      "loss": 0.2689,
+      "step": 4940
+    },
+    {
+      "epoch": 2.8534333525678015,
+      "grad_norm": 0.47564242491449826,
+      "learning_rate": 4.6028388750366756e-06,
+      "loss": 0.2809,
+      "step": 4945
+    },
+    {
+      "epoch": 2.856318522792845,
+      "grad_norm": 0.43828857688753076,
+      "learning_rate": 4.581657382430932e-06,
+      "loss": 0.2853,
+      "step": 4950
+    },
+    {
+      "epoch": 2.859203693017888,
+      "grad_norm": 0.43574357681542614,
+      "learning_rate": 4.56051024682615e-06,
+      "loss": 0.2756,
+      "step": 4955
+    },
+    {
+      "epoch": 2.8620888632429313,
+      "grad_norm": 0.42537161903700743,
+      "learning_rate": 4.53939760231354e-06,
+      "loss": 0.2874,
+      "step": 4960
+    },
+    {
+      "epoch": 2.8649740334679747,
+      "grad_norm": 0.4597603651479367,
+      "learning_rate": 4.518319582765623e-06,
+      "loss": 0.2892,
+      "step": 4965
+    },
+    {
+      "epoch": 2.8678592036930177,
+      "grad_norm": 0.4446590811574243,
+      "learning_rate": 4.497276321835357e-06,
+      "loss": 0.2712,
+      "step": 4970
+    },
+    {
+      "epoch": 2.870744373918061,
+      "grad_norm": 0.46299522122029557,
+      "learning_rate": 4.476267952955295e-06,
+      "loss": 0.3009,
+      "step": 4975
+    },
+    {
+      "epoch": 2.8736295441431046,
+      "grad_norm": 0.4385843033290836,
+      "learning_rate": 4.455294609336771e-06,
+      "loss": 0.2755,
+      "step": 4980
+    },
+    {
+      "epoch": 2.876514714368148,
+      "grad_norm": 0.44649836304692625,
+      "learning_rate": 4.434356423968999e-06,
+      "loss": 0.2837,
+      "step": 4985
+    },
+    {
+      "epoch": 2.879399884593191,
+      "grad_norm": 0.4656079138932178,
+      "learning_rate": 4.413453529618274e-06,
+      "loss": 0.2896,
+      "step": 4990
+    },
+    {
+      "epoch": 2.8822850548182344,
+      "grad_norm": 0.40090614598820945,
+      "learning_rate": 4.3925860588271015e-06,
+      "loss": 0.3039,
+      "step": 4995
+    },
+    {
+      "epoch": 2.885170225043278,
+      "grad_norm": 0.429166127365773,
+      "learning_rate": 4.371754143913398e-06,
+      "loss": 0.2732,
+      "step": 5000
+    },
+    {
+      "epoch": 2.888055395268321,
+      "grad_norm": 0.4028565372934939,
+      "learning_rate": 4.350957916969598e-06,
+      "loss": 0.2629,
+      "step": 5005
+    },
+    {
+      "epoch": 2.890940565493364,
+      "grad_norm": 0.4626974059523127,
+      "learning_rate": 4.330197509861851e-06,
+      "loss": 0.2748,
+      "step": 5010
+    },
+    {
+      "epoch": 2.8938257357184076,
+      "grad_norm": 0.4168915007613455,
+      "learning_rate": 4.3094730542291875e-06,
+      "loss": 0.2851,
+      "step": 5015
+    },
+    {
+      "epoch": 2.8967109059434506,
+      "grad_norm": 0.4404850147613879,
+      "learning_rate": 4.288784681482657e-06,
+      "loss": 0.2633,
+      "step": 5020
+    },
+    {
+      "epoch": 2.899596076168494,
+      "grad_norm": 0.4368263858822652,
+      "learning_rate": 4.268132522804532e-06,
+      "loss": 0.2795,
+      "step": 5025
+    },
+    {
+      "epoch": 2.9024812463935374,
+      "grad_norm": 0.4518971041077726,
+      "learning_rate": 4.247516709147437e-06,
+      "loss": 0.2779,
+      "step": 5030
+    },
+    {
+      "epoch": 2.9053664166185804,
+      "grad_norm": 0.42168757805620943,
+      "learning_rate": 4.2269373712335535e-06,
+      "loss": 0.2781,
+      "step": 5035
+    },
+    {
+      "epoch": 2.908251586843624,
+      "grad_norm": 0.44301611235467,
+      "learning_rate": 4.206394639553766e-06,
+      "loss": 0.2844,
+      "step": 5040
+    },
+    {
+      "epoch": 2.9111367570686673,
+      "grad_norm": 0.43608207255196907,
+      "learning_rate": 4.185888644366841e-06,
+      "loss": 0.2764,
+      "step": 5045
+    },
+    {
+      "epoch": 2.9140219272937102,
+      "grad_norm": 0.43614919217568904,
+      "learning_rate": 4.165419515698613e-06,
+      "loss": 0.2686,
+      "step": 5050
+    },
+    {
+      "epoch": 2.9169070975187537,
+      "grad_norm": 0.45163629978863207,
+      "learning_rate": 4.14498738334115e-06,
+      "loss": 0.2685,
+      "step": 5055
+    },
+    {
+      "epoch": 2.919792267743797,
+      "grad_norm": 0.42566284160748424,
+      "learning_rate": 4.1245923768519235e-06,
+      "loss": 0.2624,
+      "step": 5060
+    },
+    {
+      "epoch": 2.92267743796884,
+      "grad_norm": 0.42580290639250135,
+      "learning_rate": 4.104234625552991e-06,
+      "loss": 0.2656,
+      "step": 5065
+    },
+    {
+      "epoch": 2.9255626081938835,
+      "grad_norm": 0.4508732247385924,
+      "learning_rate": 4.083914258530202e-06,
+      "loss": 0.2829,
+      "step": 5070
+    },
+    {
+      "epoch": 2.928447778418927,
+      "grad_norm": 0.4230373908678959,
+      "learning_rate": 4.063631404632336e-06,
+      "loss": 0.283,
+      "step": 5075
+    },
+    {
+      "epoch": 2.93133294864397,
+      "grad_norm": 0.4519630975638044,
+      "learning_rate": 4.043386192470309e-06,
+      "loss": 0.299,
+      "step": 5080
+    },
+    {
+      "epoch": 2.9342181188690133,
+      "grad_norm": 0.46105467021109736,
+      "learning_rate": 4.0231787504163684e-06,
+      "loss": 0.2878,
+      "step": 5085
+    },
+    {
+      "epoch": 2.9371032890940567,
+      "grad_norm": 0.4511914211925812,
+      "learning_rate": 4.00300920660325e-06,
+      "loss": 0.2778,
+      "step": 5090
+    },
+    {
+      "epoch": 2.9399884593190997,
+      "grad_norm": 0.44702115253367053,
+      "learning_rate": 3.982877688923396e-06,
+      "loss": 0.2662,
+      "step": 5095
+    },
+    {
+      "epoch": 2.942873629544143,
+      "grad_norm": 0.427417788950817,
+      "learning_rate": 3.962784325028119e-06,
+      "loss": 0.2713,
+      "step": 5100
+    },
+    {
+      "epoch": 2.9457587997691865,
+      "grad_norm": 0.43897355082883444,
+      "learning_rate": 3.942729242326814e-06,
+      "loss": 0.2743,
+      "step": 5105
+    },
+    {
+      "epoch": 2.9486439699942295,
+      "grad_norm": 0.4670512082769263,
+      "learning_rate": 3.9227125679861286e-06,
+      "loss": 0.284,
+      "step": 5110
+    },
+    {
+      "epoch": 2.951529140219273,
+      "grad_norm": 0.4664395393065826,
+      "learning_rate": 3.902734428929172e-06,
+      "loss": 0.2769,
+      "step": 5115
+    },
+    {
+      "epoch": 2.9544143104443163,
+      "grad_norm": 0.4670672506794604,
+      "learning_rate": 3.882794951834711e-06,
+      "loss": 0.2754,
+      "step": 5120
+    },
+    {
+      "epoch": 2.9572994806693593,
+      "grad_norm": 0.43499585946337327,
+      "learning_rate": 3.862894263136361e-06,
+      "loss": 0.2688,
+      "step": 5125
+    },
+    {
+      "epoch": 2.9601846508944027,
+      "grad_norm": 0.4471133717151604,
+      "learning_rate": 3.8430324890217805e-06,
+      "loss": 0.2685,
+      "step": 5130
+    },
+    {
+      "epoch": 2.963069821119446,
+      "grad_norm": 0.4382618072040372,
+      "learning_rate": 3.823209755431873e-06,
+      "loss": 0.2809,
+      "step": 5135
+    },
+    {
+      "epoch": 2.965954991344489,
+      "grad_norm": 0.4241285275096586,
+      "learning_rate": 3.8034261880600034e-06,
+      "loss": 0.27,
+      "step": 5140
+    },
+    {
+      "epoch": 2.9688401615695326,
+      "grad_norm": 0.4091644110971949,
+      "learning_rate": 3.78368191235118e-06,
+      "loss": 0.2799,
+      "step": 5145
+    },
+    {
+      "epoch": 2.971725331794576,
+      "grad_norm": 0.43102734170478396,
+      "learning_rate": 3.76397705350127e-06,
+      "loss": 0.2797,
+      "step": 5150
+    },
+    {
+      "epoch": 2.974610502019619,
+      "grad_norm": 0.4370911603217546,
+      "learning_rate": 3.7443117364561964e-06,
+      "loss": 0.2701,
+      "step": 5155
+    },
+    {
+      "epoch": 2.9774956722446624,
+      "grad_norm": 0.42916373087001025,
+      "learning_rate": 3.7246860859111655e-06,
+      "loss": 0.2734,
+      "step": 5160
+    },
+    {
+      "epoch": 2.980380842469706,
+      "grad_norm": 0.4275858080889107,
+      "learning_rate": 3.705100226309858e-06,
+      "loss": 0.2677,
+      "step": 5165
+    },
+    {
+      "epoch": 2.983266012694749,
+      "grad_norm": 0.4310089026075201,
+      "learning_rate": 3.6855542818436397e-06,
+      "loss": 0.2726,
+      "step": 5170
+    },
+    {
+      "epoch": 2.986151182919792,
+      "grad_norm": 0.44889673891857773,
+      "learning_rate": 3.6660483764507916e-06,
+      "loss": 0.265,
+      "step": 5175
+    },
+    {
+      "epoch": 2.9890363531448356,
+      "grad_norm": 0.4597978279604376,
+      "learning_rate": 3.6465826338157007e-06,
+      "loss": 0.2857,
+      "step": 5180
+    },
+    {
+      "epoch": 2.9919215233698786,
+      "grad_norm": 0.4672195167157573,
+      "learning_rate": 3.6271571773680893e-06,
+      "loss": 0.2765,
+      "step": 5185
+    },
+    {
+      "epoch": 2.994806693594922,
+      "grad_norm": 0.4409412220124302,
+      "learning_rate": 3.6077721302822355e-06,
+      "loss": 0.2826,
+      "step": 5190
+    },
+    {
+      "epoch": 2.9976918638199654,
+      "grad_norm": 0.4448322757550867,
+      "learning_rate": 3.588427615476189e-06,
+      "loss": 0.2951,
+      "step": 5195
+    },
+    {
+      "epoch": 3.0,
+      "eval_loss": 0.4704112410545349,
+      "eval_runtime": 0.5938,
+      "eval_samples_per_second": 129.667,
+      "eval_steps_per_second": 3.368,
+      "step": 5199
+    },
+    {
+      "epoch": 3.000577034045009,
+      "grad_norm": 0.478834624205184,
+      "learning_rate": 3.5691237556109794e-06,
+      "loss": 0.2548,
+      "step": 5200
+    },
+    {
+      "epoch": 3.003462204270052,
+      "grad_norm": 0.45264093142712225,
+      "learning_rate": 3.549860673089852e-06,
+      "loss": 0.2178,
+      "step": 5205
+    },
+    {
+      "epoch": 3.0063473744950953,
+      "grad_norm": 0.42695857914733454,
+      "learning_rate": 3.5306384900574943e-06,
+      "loss": 0.2145,
+      "step": 5210
+    },
+    {
+      "epoch": 3.0092325447201387,
+      "grad_norm": 0.4252664702307335,
+      "learning_rate": 3.5114573283992536e-06,
+      "loss": 0.2017,
+      "step": 5215
+    },
+    {
+      "epoch": 3.0121177149451817,
+      "grad_norm": 0.4362013593591915,
+      "learning_rate": 3.492317309740362e-06,
+      "loss": 0.202,
+      "step": 5220
+    },
+    {
+      "epoch": 3.015002885170225,
+      "grad_norm": 0.7918296216376329,
+      "learning_rate": 3.473218555445166e-06,
+      "loss": 0.2023,
+      "step": 5225
+    },
+    {
+      "epoch": 3.0178880553952685,
+      "grad_norm": 0.4425689050623567,
+      "learning_rate": 3.4541611866163748e-06,
+      "loss": 0.2156,
+      "step": 5230
+    },
+    {
+      "epoch": 3.0207732256203115,
+      "grad_norm": 0.434243728200696,
+      "learning_rate": 3.435145324094259e-06,
+      "loss": 0.1967,
+      "step": 5235
+    },
+    {
+      "epoch": 3.023658395845355,
+      "grad_norm": 0.48550762220985233,
+      "learning_rate": 3.4161710884559186e-06,
+      "loss": 0.228,
+      "step": 5240
+    },
+    {
+      "epoch": 3.0265435660703983,
+      "grad_norm": 0.4451374888019577,
+      "learning_rate": 3.3972386000144975e-06,
+      "loss": 0.1994,
+      "step": 5245
+    },
+    {
+      "epoch": 3.0294287362954413,
+      "grad_norm": 0.4479982505677569,
+      "learning_rate": 3.378347978818425e-06,
+      "loss": 0.2064,
+      "step": 5250
+    },
+    {
+      "epoch": 3.0323139065204847,
+      "grad_norm": 0.48420281819894295,
+      "learning_rate": 3.359499344650651e-06,
+      "loss": 0.2104,
+      "step": 5255
+    },
+    {
+      "epoch": 3.035199076745528,
+      "grad_norm": 0.4674463960460507,
+      "learning_rate": 3.3406928170278997e-06,
+      "loss": 0.2107,
+      "step": 5260
+    },
+    {
+      "epoch": 3.038084246970571,
+      "grad_norm": 0.42110517886246845,
+      "learning_rate": 3.3219285151999037e-06,
+      "loss": 0.1936,
+      "step": 5265
+    },
+    {
+      "epoch": 3.0409694171956145,
+      "grad_norm": 0.4325902687189473,
+      "learning_rate": 3.3032065581486393e-06,
+      "loss": 0.2107,
+      "step": 5270
+    },
+    {
+      "epoch": 3.043854587420658,
+      "grad_norm": 0.45814670201275987,
+      "learning_rate": 3.2845270645875838e-06,
+      "loss": 0.2073,
+      "step": 5275
+    },
+    {
+      "epoch": 3.046739757645701,
+      "grad_norm": 0.4673351294536787,
+      "learning_rate": 3.26589015296096e-06,
+      "loss": 0.2141,
+      "step": 5280
+    },
+    {
+      "epoch": 3.0496249278707444,
+      "grad_norm": 0.43997040330941384,
+      "learning_rate": 3.2472959414429883e-06,
+      "loss": 0.2108,
+      "step": 5285
+    },
+    {
+      "epoch": 3.0525100980957878,
+      "grad_norm": 0.4862585478605358,
+      "learning_rate": 3.228744547937125e-06,
+      "loss": 0.2089,
+      "step": 5290
+    },
+    {
+      "epoch": 3.0553952683208307,
+      "grad_norm": 0.4501337212448104,
+      "learning_rate": 3.2102360900753237e-06,
+      "loss": 0.216,
+      "step": 5295
+    },
+    {
+      "epoch": 3.058280438545874,
+      "grad_norm": 0.43785695040034356,
+      "learning_rate": 3.1917706852173003e-06,
+      "loss": 0.1999,
+      "step": 5300
+    },
+    {
+      "epoch": 3.0611656087709176,
+      "grad_norm": 0.45799695752548214,
+      "learning_rate": 3.173348450449759e-06,
+      "loss": 0.2034,
+      "step": 5305
+    },
+    {
+      "epoch": 3.0640507789959606,
+      "grad_norm": 0.4896057098908631,
+      "learning_rate": 3.1549695025856873e-06,
+      "loss": 0.2138,
+      "step": 5310
+    },
+    {
+      "epoch": 3.066935949221004,
+      "grad_norm": 0.43978051760808934,
+      "learning_rate": 3.1366339581635785e-06,
+      "loss": 0.211,
+      "step": 5315
+    },
+    {
+      "epoch": 3.0698211194460474,
+      "grad_norm": 0.5408651817364915,
+      "learning_rate": 3.118341933446726e-06,
+      "loss": 0.2189,
+      "step": 5320
+    },
+    {
+      "epoch": 3.0727062896710904,
+      "grad_norm": 0.47167963172935967,
+      "learning_rate": 3.100093544422459e-06,
+      "loss": 0.2087,
+      "step": 5325
+    },
+    {
+      "epoch": 3.075591459896134,
+      "grad_norm": 0.45222235059141447,
+      "learning_rate": 3.0818889068014167e-06,
+      "loss": 0.2086,
+      "step": 5330
+    },
+    {
+      "epoch": 3.0784766301211772,
+      "grad_norm": 0.46762858344658953,
+      "learning_rate": 3.0637281360168348e-06,
+      "loss": 0.2058,
+      "step": 5335
+    },
+    {
+      "epoch": 3.0813618003462206,
+      "grad_norm": 0.4982055068360655,
+      "learning_rate": 3.045611347223776e-06,
+      "loss": 0.2168,
+      "step": 5340
+    },
+    {
+      "epoch": 3.0842469705712636,
+      "grad_norm": 0.44825046777882716,
+      "learning_rate": 3.0275386552984232e-06,
+      "loss": 0.2089,
+      "step": 5345
+    },
+    {
+      "epoch": 3.087132140796307,
+      "grad_norm": 0.4780597142010933,
+      "learning_rate": 3.00951017483735e-06,
+      "loss": 0.2,
+      "step": 5350
+    },
+    {
+      "epoch": 3.0900173110213505,
+      "grad_norm": 0.4749892463930495,
+      "learning_rate": 2.991526020156795e-06,
+      "loss": 0.2082,
+      "step": 5355
+    },
+    {
+      "epoch": 3.0929024812463934,
+      "grad_norm": 0.5316835135478458,
+      "learning_rate": 2.9735863052919256e-06,
+      "loss": 0.2117,
+      "step": 5360
+    },
+    {
+      "epoch": 3.095787651471437,
+      "grad_norm": 0.48592697259547013,
+      "learning_rate": 2.955691143996119e-06,
+      "loss": 0.2189,
+      "step": 5365
+    },
+    {
+      "epoch": 3.0986728216964803,
+      "grad_norm": 0.478325002856811,
+      "learning_rate": 2.9378406497402577e-06,
+      "loss": 0.2217,
+      "step": 5370
+    },
+    {
+      "epoch": 3.1015579919215233,
+      "grad_norm": 0.47784013307657136,
+      "learning_rate": 2.9200349357119805e-06,
+      "loss": 0.2047,
+      "step": 5375
+    },
+    {
+      "epoch": 3.1044431621465667,
+      "grad_norm": 0.4366920553845523,
+      "learning_rate": 2.902274114814995e-06,
+      "loss": 0.2052,
+      "step": 5380
+    },
+    {
+      "epoch": 3.10732833237161,
+      "grad_norm": 0.48726638763350977,
+      "learning_rate": 2.884558299668333e-06,
+      "loss": 0.2135,
+      "step": 5385
+    },
+    {
+      "epoch": 3.110213502596653,
+      "grad_norm": 0.4352029364038837,
+      "learning_rate": 2.866887602605667e-06,
+      "loss": 0.2146,
+      "step": 5390
+    },
+    {
+      "epoch": 3.1130986728216965,
+      "grad_norm": 0.4618849547883958,
+      "learning_rate": 2.8492621356745677e-06,
+      "loss": 0.2007,
+      "step": 5395
+    },
+    {
+      "epoch": 3.11598384304674,
+      "grad_norm": 0.5680788070054426,
+      "learning_rate": 2.831682010635811e-06,
+      "loss": 0.2254,
+      "step": 5400
+    },
+    {
+      "epoch": 3.118869013271783,
+      "grad_norm": 0.4478067921317212,
+      "learning_rate": 2.8141473389626706e-06,
+      "loss": 0.2025,
+      "step": 5405
+    },
+    {
+      "epoch": 3.1217541834968263,
+      "grad_norm": 0.4383072651557812,
+      "learning_rate": 2.7966582318402046e-06,
+      "loss": 0.2038,
+      "step": 5410
+    },
+    {
+      "epoch": 3.1246393537218697,
+      "grad_norm": 0.4469251744492116,
+      "learning_rate": 2.7792148001645505e-06,
+      "loss": 0.2034,
+      "step": 5415
+    },
+    {
+      "epoch": 3.1275245239469127,
+      "grad_norm": 0.45796076894000676,
+      "learning_rate": 2.7618171545422156e-06,
+      "loss": 0.2237,
+      "step": 5420
+    },
+    {
+      "epoch": 3.130409694171956,
+      "grad_norm": 0.4732594923700523,
+      "learning_rate": 2.744465405289406e-06,
+      "loss": 0.2115,
+      "step": 5425
+    },
+    {
+      "epoch": 3.1332948643969996,
+      "grad_norm": 0.47883717903500816,
+      "learning_rate": 2.7271596624312845e-06,
+      "loss": 0.2176,
+      "step": 5430
+    },
+    {
+      "epoch": 3.1361800346220425,
+      "grad_norm": 0.47490798721155425,
+      "learning_rate": 2.709900035701297e-06,
+      "loss": 0.1962,
+      "step": 5435
+    },
+    {
+      "epoch": 3.139065204847086,
+      "grad_norm": 0.47556110953605957,
+      "learning_rate": 2.6926866345404846e-06,
+      "loss": 0.2122,
+      "step": 5440
+    },
+    {
+      "epoch": 3.1419503750721294,
+      "grad_norm": 0.4674071776337381,
+      "learning_rate": 2.6755195680967607e-06,
+      "loss": 0.1992,
+      "step": 5445
+    },
+    {
+      "epoch": 3.1448355452971724,
+      "grad_norm": 0.474411421746897,
+      "learning_rate": 2.658398945224253e-06,
+      "loss": 0.1924,
+      "step": 5450
+    },
+    {
+      "epoch": 3.1477207155222158,
+      "grad_norm": 0.47606141844860833,
+      "learning_rate": 2.6413248744825837e-06,
+      "loss": 0.2076,
+      "step": 5455
+    },
+    {
+      "epoch": 3.150605885747259,
+      "grad_norm": 0.45905433780073146,
+      "learning_rate": 2.624297464136204e-06,
+      "loss": 0.2142,
+      "step": 5460
+    },
+    {
+      "epoch": 3.153491055972302,
+      "grad_norm": 0.536638485210775,
+      "learning_rate": 2.6073168221536906e-06,
+      "loss": 0.219,
+      "step": 5465
+    },
+    {
+      "epoch": 3.1563762261973456,
+      "grad_norm": 0.45678787981866464,
+      "learning_rate": 2.5903830562070675e-06,
+      "loss": 0.2036,
+      "step": 5470
+    },
+    {
+      "epoch": 3.159261396422389,
+      "grad_norm": 0.4496673360192395,
+      "learning_rate": 2.57349627367113e-06,
+      "loss": 0.2092,
+      "step": 5475
+    },
+    {
+      "epoch": 3.162146566647432,
+      "grad_norm": 0.4612630830131145,
+      "learning_rate": 2.5566565816227585e-06,
+      "loss": 0.204,
+      "step": 5480
+    },
+    {
+      "epoch": 3.1650317368724754,
+      "grad_norm": 0.45145251818392435,
+      "learning_rate": 2.5398640868402304e-06,
+      "loss": 0.2056,
+      "step": 5485
+    },
+    {
+      "epoch": 3.167916907097519,
+      "grad_norm": 0.49129615711916425,
+      "learning_rate": 2.523118895802553e-06,
+      "loss": 0.2069,
+      "step": 5490
+    },
+    {
+      "epoch": 3.170802077322562,
+      "grad_norm": 0.5139622757706684,
+      "learning_rate": 2.506421114688794e-06,
+      "loss": 0.221,
+      "step": 5495
+    },
+    {
+      "epoch": 3.1736872475476052,
+      "grad_norm": 0.48052040604024665,
+      "learning_rate": 2.4897708493773997e-06,
+      "loss": 0.1973,
+      "step": 5500
+    },
+    {
+      "epoch": 3.1765724177726486,
+      "grad_norm": 0.49706524838406957,
+      "learning_rate": 2.4731682054455174e-06,
+      "loss": 0.2053,
+      "step": 5505
+    },
+    {
+      "epoch": 3.179457587997692,
+      "grad_norm": 0.5133679294908896,
+      "learning_rate": 2.456613288168337e-06,
+      "loss": 0.2035,
+      "step": 5510
+    },
+    {
+      "epoch": 3.182342758222735,
+      "grad_norm": 0.483475787236041,
+      "learning_rate": 2.440106202518423e-06,
+      "loss": 0.2082,
+      "step": 5515
+    },
+    {
+      "epoch": 3.1852279284477785,
+      "grad_norm": 0.4815903117086154,
+      "learning_rate": 2.423647053165046e-06,
+      "loss": 0.2046,
+      "step": 5520
+    },
+    {
+      "epoch": 3.188113098672822,
+      "grad_norm": 0.4753159403205944,
+      "learning_rate": 2.4072359444735117e-06,
+      "loss": 0.2116,
+      "step": 5525
+    },
+    {
+      "epoch": 3.190998268897865,
+      "grad_norm": 0.5207748356365552,
+      "learning_rate": 2.390872980504516e-06,
+      "loss": 0.2078,
+      "step": 5530
+    },
+    {
+      "epoch": 3.1938834391229083,
+      "grad_norm": 0.500796915089432,
+      "learning_rate": 2.374558265013469e-06,
+      "loss": 0.2013,
+      "step": 5535
+    },
+    {
+      "epoch": 3.1967686093479517,
+      "grad_norm": 0.5192772858747745,
+      "learning_rate": 2.358291901449842e-06,
+      "loss": 0.2003,
+      "step": 5540
+    },
+    {
+      "epoch": 3.1996537795729947,
+      "grad_norm": 0.4809730211239497,
+      "learning_rate": 2.342073992956517e-06,
+      "loss": 0.2098,
+      "step": 5545
+    },
+    {
+      "epoch": 3.202538949798038,
+      "grad_norm": 0.4627046426816422,
+      "learning_rate": 2.3259046423691333e-06,
+      "loss": 0.2032,
+      "step": 5550
+    },
+    {
+      "epoch": 3.2054241200230815,
+      "grad_norm": 0.5339618244022528,
+      "learning_rate": 2.309783952215421e-06,
+      "loss": 0.2075,
+      "step": 5555
+    },
+    {
+      "epoch": 3.2083092902481245,
+      "grad_norm": 0.5160587961491079,
+      "learning_rate": 2.293712024714564e-06,
+      "loss": 0.2147,
+      "step": 5560
+    },
+    {
+      "epoch": 3.211194460473168,
+      "grad_norm": 0.5541552433226901,
+      "learning_rate": 2.2776889617765495e-06,
+      "loss": 0.2367,
+      "step": 5565
+    },
+    {
+      "epoch": 3.2140796306982113,
+      "grad_norm": 0.45574118435347694,
+      "learning_rate": 2.2617148650015275e-06,
+      "loss": 0.2104,
+      "step": 5570
+    },
+    {
+      "epoch": 3.2169648009232543,
+      "grad_norm": 0.4714574839149813,
+      "learning_rate": 2.2457898356791496e-06,
+      "loss": 0.2154,
+      "step": 5575
+    },
+    {
+      "epoch": 3.2198499711482977,
+      "grad_norm": 0.5065703016927723,
+      "learning_rate": 2.2299139747879385e-06,
+      "loss": 0.2045,
+      "step": 5580
+    },
+    {
+      "epoch": 3.222735141373341,
+      "grad_norm": 0.46572520384835114,
+      "learning_rate": 2.214087382994653e-06,
+      "loss": 0.2368,
+      "step": 5585
+    },
+    {
+      "epoch": 3.225620311598384,
+      "grad_norm": 0.4780383032384353,
+      "learning_rate": 2.198310160653636e-06,
+      "loss": 0.2041,
+      "step": 5590
+    },
+    {
+      "epoch": 3.2285054818234276,
+      "grad_norm": 0.48268579681924456,
+      "learning_rate": 2.1825824078061898e-06,
+      "loss": 0.2057,
+      "step": 5595
+    },
+    {
+      "epoch": 3.231390652048471,
+      "grad_norm": 0.48646053407264,
+      "learning_rate": 2.1669042241799378e-06,
+      "loss": 0.199,
+      "step": 5600
+    },
+    {
+      "epoch": 3.234275822273514,
+      "grad_norm": 0.49576817208299834,
+      "learning_rate": 2.1512757091881874e-06,
+      "loss": 0.1971,
+      "step": 5605
+    },
+    {
+      "epoch": 3.2371609924985574,
+      "grad_norm": 0.4841025834202475,
+      "learning_rate": 2.1356969619293023e-06,
+      "loss": 0.2227,
+      "step": 5610
+    },
+    {
+      "epoch": 3.240046162723601,
+      "grad_norm": 0.5512557293571464,
+      "learning_rate": 2.1201680811860815e-06,
+      "loss": 0.1993,
+      "step": 5615
+    },
+    {
+      "epoch": 3.2429313329486438,
+      "grad_norm": 0.49443124657159826,
+      "learning_rate": 2.1046891654251266e-06,
+      "loss": 0.207,
+      "step": 5620
+    },
+    {
+      "epoch": 3.245816503173687,
+      "grad_norm": 0.502803434696454,
+      "learning_rate": 2.089260312796213e-06,
+      "loss": 0.2037,
+      "step": 5625
+    },
+    {
+      "epoch": 3.2487016733987306,
+      "grad_norm": 0.4761464874997464,
+      "learning_rate": 2.0738816211316716e-06,
+      "loss": 0.1969,
+      "step": 5630
+    },
+    {
+      "epoch": 3.251586843623774,
+      "grad_norm": 0.5250059943014331,
+      "learning_rate": 2.0585531879457743e-06,
+      "loss": 0.1998,
+      "step": 5635
+    },
+    {
+      "epoch": 3.254472013848817,
+      "grad_norm": 0.4534253902795417,
+      "learning_rate": 2.043275110434113e-06,
+      "loss": 0.2057,
+      "step": 5640
+    },
+    {
+      "epoch": 3.2573571840738604,
+      "grad_norm": 0.456942792070371,
+      "learning_rate": 2.0280474854729727e-06,
+      "loss": 0.2014,
+      "step": 5645
+    },
+    {
+      "epoch": 3.260242354298904,
+      "grad_norm": 0.49132441725377424,
+      "learning_rate": 2.0128704096187258e-06,
+      "loss": 0.214,
+      "step": 5650
+    },
+    {
+      "epoch": 3.263127524523947,
+      "grad_norm": 0.43559057805097934,
+      "learning_rate": 1.9977439791072296e-06,
+      "loss": 0.2025,
+      "step": 5655
+    },
+    {
+      "epoch": 3.2660126947489903,
+      "grad_norm": 0.5032263108222483,
+      "learning_rate": 1.9826682898531923e-06,
+      "loss": 0.2036,
+      "step": 5660
+    },
+    {
+      "epoch": 3.2688978649740337,
+      "grad_norm": 0.5166100007288694,
+      "learning_rate": 1.967643437449591e-06,
+      "loss": 0.2039,
+      "step": 5665
+    },
+    {
+      "epoch": 3.2717830351990767,
+      "grad_norm": 0.5625840783677031,
+      "learning_rate": 1.9526695171670428e-06,
+      "loss": 0.2186,
+      "step": 5670
+    },
+    {
+      "epoch": 3.27466820542412,
+      "grad_norm": 0.4771204637584591,
+      "learning_rate": 1.937746623953218e-06,
+      "loss": 0.2105,
+      "step": 5675
+    },
+    {
+      "epoch": 3.2775533756491635,
+      "grad_norm": 0.5120534992728352,
+      "learning_rate": 1.9228748524322283e-06,
+      "loss": 0.1962,
+      "step": 5680
+    },
+    {
+      "epoch": 3.2804385458742065,
+      "grad_norm": 0.5018775730767316,
+      "learning_rate": 1.908054296904023e-06,
+      "loss": 0.2109,
+      "step": 5685
+    },
+    {
+      "epoch": 3.28332371609925,
+      "grad_norm": 0.4876912669877195,
+      "learning_rate": 1.8932850513438139e-06,
+      "loss": 0.2126,
+      "step": 5690
+    },
+    {
+      "epoch": 3.2862088863242933,
+      "grad_norm": 0.4691715888051167,
+      "learning_rate": 1.8785672094014484e-06,
+      "loss": 0.2016,
+      "step": 5695
+    },
+    {
+      "epoch": 3.2890940565493363,
+      "grad_norm": 0.48546733091545335,
+      "learning_rate": 1.8639008644008317e-06,
+      "loss": 0.2139,
+      "step": 5700
+    },
+    {
+      "epoch": 3.2919792267743797,
+      "grad_norm": 0.48053507834963044,
+      "learning_rate": 1.84928610933934e-06,
+      "loss": 0.2069,
+      "step": 5705
+    },
+    {
+      "epoch": 3.294864396999423,
+      "grad_norm": 0.4978213968888964,
+      "learning_rate": 1.8347230368872227e-06,
+      "loss": 0.2125,
+      "step": 5710
+    },
+    {
+      "epoch": 3.297749567224466,
+      "grad_norm": 0.4619698699264187,
+      "learning_rate": 1.8202117393870122e-06,
+      "loss": 0.2048,
+      "step": 5715
+    },
+    {
+      "epoch": 3.3006347374495095,
+      "grad_norm": 0.4439799217375983,
+      "learning_rate": 1.805752308852945e-06,
+      "loss": 0.2076,
+      "step": 5720
+    },
+    {
+      "epoch": 3.303519907674553,
+      "grad_norm": 0.5050476605960446,
+      "learning_rate": 1.7913448369703801e-06,
+      "loss": 0.2148,
+      "step": 5725
+    },
+    {
+      "epoch": 3.306405077899596,
+      "grad_norm": 0.4449280284195996,
+      "learning_rate": 1.776989415095206e-06,
+      "loss": 0.1961,
+      "step": 5730
+    },
+    {
+      "epoch": 3.3092902481246393,
+      "grad_norm": 0.5023493469028406,
+      "learning_rate": 1.7626861342532764e-06,
+      "loss": 0.2108,
+      "step": 5735
+    },
+    {
+      "epoch": 3.3121754183496828,
+      "grad_norm": 0.4899997511129661,
+      "learning_rate": 1.748435085139818e-06,
+      "loss": 0.2041,
+      "step": 5740
+    },
+    {
+      "epoch": 3.3150605885747257,
+      "grad_norm": 0.620987392864056,
+      "learning_rate": 1.7342363581188716e-06,
+      "loss": 0.1967,
+      "step": 5745
+    },
+    {
+      "epoch": 3.317945758799769,
+      "grad_norm": 0.48947716401563934,
+      "learning_rate": 1.720090043222704e-06,
+      "loss": 0.2093,
+      "step": 5750
+    },
+    {
+      "epoch": 3.3208309290248126,
+      "grad_norm": 0.49193608057425164,
+      "learning_rate": 1.705996230151239e-06,
+      "loss": 0.2143,
+      "step": 5755
+    },
+    {
+      "epoch": 3.3237160992498556,
+      "grad_norm": 0.5410390908943622,
+      "learning_rate": 1.6919550082715108e-06,
+      "loss": 0.2101,
+      "step": 5760
+    },
+    {
+      "epoch": 3.326601269474899,
+      "grad_norm": 0.44748685099184915,
+      "learning_rate": 1.6779664666170626e-06,
+      "loss": 0.2154,
+      "step": 5765
+    },
+    {
+      "epoch": 3.3294864396999424,
+      "grad_norm": 0.49202824127233535,
+      "learning_rate": 1.6640306938874052e-06,
+      "loss": 0.2084,
+      "step": 5770
+    },
+    {
+      "epoch": 3.3323716099249854,
+      "grad_norm": 0.5129313111006824,
+      "learning_rate": 1.650147778447444e-06,
+      "loss": 0.2043,
+      "step": 5775
+    },
+    {
+      "epoch": 3.335256780150029,
+      "grad_norm": 0.5011687713242771,
+      "learning_rate": 1.6363178083269381e-06,
+      "loss": 0.2137,
+      "step": 5780
+    },
+    {
+      "epoch": 3.338141950375072,
+      "grad_norm": 0.5125645791853729,
+      "learning_rate": 1.6225408712199097e-06,
+      "loss": 0.213,
+      "step": 5785
+    },
+    {
+      "epoch": 3.341027120600115,
+      "grad_norm": 0.496426542842325,
+      "learning_rate": 1.608817054484113e-06,
+      "loss": 0.2036,
+      "step": 5790
+    },
+    {
+      "epoch": 3.3439122908251586,
+      "grad_norm": 0.573443010809008,
+      "learning_rate": 1.5951464451404775e-06,
+      "loss": 0.1975,
+      "step": 5795
+    },
+    {
+      "epoch": 3.346797461050202,
+      "grad_norm": 0.4777011846082858,
+      "learning_rate": 1.5815291298725387e-06,
+      "loss": 0.2278,
+      "step": 5800
+    },
+    {
+      "epoch": 3.349682631275245,
+      "grad_norm": 0.4994285452680278,
+      "learning_rate": 1.5679651950259178e-06,
+      "loss": 0.2019,
+      "step": 5805
+    },
+    {
+      "epoch": 3.3525678015002884,
+      "grad_norm": 0.47305655838061333,
+      "learning_rate": 1.5544547266077425e-06,
+      "loss": 0.2121,
+      "step": 5810
+    },
+    {
+      "epoch": 3.355452971725332,
+      "grad_norm": 0.5117964943614841,
+      "learning_rate": 1.54099781028613e-06,
+      "loss": 0.2117,
+      "step": 5815
+    },
+    {
+      "epoch": 3.358338141950375,
+      "grad_norm": 0.47969243336386125,
+      "learning_rate": 1.52759453138962e-06,
+      "loss": 0.2103,
+      "step": 5820
+    },
+    {
+      "epoch": 3.3612233121754183,
+      "grad_norm": 0.46863801053654247,
+      "learning_rate": 1.514244974906649e-06,
+      "loss": 0.2162,
+      "step": 5825
+    },
+    {
+      "epoch": 3.3641084824004617,
+      "grad_norm": 0.4937087047605848,
+      "learning_rate": 1.5009492254850056e-06,
+      "loss": 0.2074,
+      "step": 5830
+    },
+    {
+      "epoch": 3.366993652625505,
+      "grad_norm": 0.5349150058728949,
+      "learning_rate": 1.4877073674313004e-06,
+      "loss": 0.2092,
+      "step": 5835
+    },
+    {
+      "epoch": 3.369878822850548,
+      "grad_norm": 0.509590831524668,
+      "learning_rate": 1.4745194847104184e-06,
+      "loss": 0.2258,
+      "step": 5840
+    },
+    {
+      "epoch": 3.3727639930755915,
+      "grad_norm": 0.5411363420436703,
+      "learning_rate": 1.461385660944994e-06,
+      "loss": 0.2026,
+      "step": 5845
+    },
+    {
+      "epoch": 3.375649163300635,
+      "grad_norm": 0.4869231603073319,
+      "learning_rate": 1.4483059794148869e-06,
+      "loss": 0.212,
+      "step": 5850
+    },
+    {
+      "epoch": 3.378534333525678,
+      "grad_norm": 0.5066221127114735,
+      "learning_rate": 1.4352805230566458e-06,
+      "loss": 0.2066,
+      "step": 5855
+    },
+    {
+      "epoch": 3.3814195037507213,
+      "grad_norm": 0.49848266645770745,
+      "learning_rate": 1.4223093744629802e-06,
+      "loss": 0.2039,
+      "step": 5860
+    },
+    {
+      "epoch": 3.3843046739757647,
+      "grad_norm": 0.46674023509056656,
+      "learning_rate": 1.409392615882248e-06,
+      "loss": 0.2056,
+      "step": 5865
+    },
+    {
+      "epoch": 3.3871898442008077,
+      "grad_norm": 0.47153344056400004,
+      "learning_rate": 1.3965303292179211e-06,
+      "loss": 0.2035,
+      "step": 5870
+    },
+    {
+      "epoch": 3.390075014425851,
+      "grad_norm": 0.5737791203551417,
+      "learning_rate": 1.3837225960280777e-06,
+      "loss": 0.2092,
+      "step": 5875
+    },
+    {
+      "epoch": 3.3929601846508946,
+      "grad_norm": 0.5430927208050875,
+      "learning_rate": 1.370969497524872e-06,
+      "loss": 0.2102,
+      "step": 5880
+    },
+    {
+      "epoch": 3.3958453548759375,
+      "grad_norm": 0.4648636735503783,
+      "learning_rate": 1.3582711145740378e-06,
+      "loss": 0.2036,
+      "step": 5885
+    },
+    {
+      "epoch": 3.398730525100981,
+      "grad_norm": 0.4809453402273537,
+      "learning_rate": 1.3456275276943543e-06,
+      "loss": 0.206,
+      "step": 5890
+    },
+    {
+      "epoch": 3.4016156953260244,
+      "grad_norm": 0.5512582459841553,
+      "learning_rate": 1.3330388170571496e-06,
+      "loss": 0.2057,
+      "step": 5895
+    },
+    {
+      "epoch": 3.4045008655510673,
+      "grad_norm": 0.5125832192345321,
+      "learning_rate": 1.3205050624857895e-06,
+      "loss": 0.2065,
+      "step": 5900
+    },
+    {
+      "epoch": 3.4073860357761108,
+      "grad_norm": 0.5558586586402942,
+      "learning_rate": 1.3080263434551743e-06,
+      "loss": 0.2166,
+      "step": 5905
+    },
+    {
+      "epoch": 3.410271206001154,
+      "grad_norm": 0.48236963618343165,
+      "learning_rate": 1.295602739091224e-06,
+      "loss": 0.2054,
+      "step": 5910
+    },
+    {
+      "epoch": 3.413156376226197,
+      "grad_norm": 0.47170396426714695,
+      "learning_rate": 1.2832343281703852e-06,
+      "loss": 0.2102,
+      "step": 5915
+    },
+    {
+      "epoch": 3.4160415464512406,
+      "grad_norm": 0.4953986618513792,
+      "learning_rate": 1.2709211891191331e-06,
+      "loss": 0.2262,
+      "step": 5920
+    },
+    {
+      "epoch": 3.418926716676284,
+      "grad_norm": 0.5094102635524805,
+      "learning_rate": 1.2586634000134735e-06,
+      "loss": 0.2052,
+      "step": 5925
+    },
+    {
+      "epoch": 3.4218118869013274,
+      "grad_norm": 0.46465406913045626,
+      "learning_rate": 1.2464610385784381e-06,
+      "loss": 0.2166,
+      "step": 5930
+    },
+    {
+      "epoch": 3.4246970571263704,
+      "grad_norm": 0.537941105319963,
+      "learning_rate": 1.2343141821876015e-06,
+      "loss": 0.2052,
+      "step": 5935
+    },
+    {
+      "epoch": 3.427582227351414,
+      "grad_norm": 0.4708366082196696,
+      "learning_rate": 1.2222229078625935e-06,
+      "loss": 0.2045,
+      "step": 5940
+    },
+    {
+      "epoch": 3.4304673975764572,
+      "grad_norm": 0.4910988413622275,
+      "learning_rate": 1.2101872922725976e-06,
+      "loss": 0.2131,
+      "step": 5945
+    },
+    {
+      "epoch": 3.4333525678015,
+      "grad_norm": 0.4726816646220706,
+      "learning_rate": 1.198207411733877e-06,
+      "loss": 0.2126,
+      "step": 5950
+    },
+    {
+      "epoch": 3.4362377380265436,
+      "grad_norm": 0.499819420315185,
+      "learning_rate": 1.1862833422092878e-06,
+      "loss": 0.2053,
+      "step": 5955
+    },
+    {
+      "epoch": 3.439122908251587,
+      "grad_norm": 0.45811978260654357,
+      "learning_rate": 1.1744151593077925e-06,
+      "loss": 0.2166,
+      "step": 5960
+    },
+    {
+      "epoch": 3.44200807847663,
+      "grad_norm": 0.49132678636515253,
+      "learning_rate": 1.1626029382839798e-06,
+      "loss": 0.2093,
+      "step": 5965
+    },
+    {
+      "epoch": 3.4448932487016735,
+      "grad_norm": 0.5178705009636367,
+      "learning_rate": 1.1508467540376e-06,
+      "loss": 0.1978,
+      "step": 5970
+    },
+    {
+      "epoch": 3.447778418926717,
+      "grad_norm": 0.4771433582042318,
+      "learning_rate": 1.1391466811130813e-06,
+      "loss": 0.2055,
+      "step": 5975
+    },
+    {
+      "epoch": 3.45066358915176,
+      "grad_norm": 0.4849667450545781,
+      "learning_rate": 1.127502793699049e-06,
+      "loss": 0.2114,
+      "step": 5980
+    },
+    {
+      "epoch": 3.4535487593768033,
+      "grad_norm": 0.5323744822003759,
+      "learning_rate": 1.1159151656278677e-06,
+      "loss": 0.2246,
+      "step": 5985
+    },
+    {
+      "epoch": 3.4564339296018467,
+      "grad_norm": 0.517295087718039,
+      "learning_rate": 1.104383870375172e-06,
+      "loss": 0.2033,
+      "step": 5990
+    },
+    {
+      "epoch": 3.4593190998268897,
+      "grad_norm": 0.49882873868199235,
+      "learning_rate": 1.0929089810593974e-06,
+      "loss": 0.2072,
+      "step": 5995
+    },
+    {
+      "epoch": 3.462204270051933,
+      "grad_norm": 0.5007022327120706,
+      "learning_rate": 1.0814905704413093e-06,
+      "loss": 0.2257,
+      "step": 6000
+    },
+    {
+      "epoch": 3.4650894402769765,
+      "grad_norm": 0.46291584199269026,
+      "learning_rate": 1.070128710923556e-06,
+      "loss": 0.2086,
+      "step": 6005
+    },
+    {
+      "epoch": 3.4679746105020195,
+      "grad_norm": 0.5269534068567228,
+      "learning_rate": 1.0588234745502012e-06,
+      "loss": 0.2112,
+      "step": 6010
+    },
+    {
+      "epoch": 3.470859780727063,
+      "grad_norm": 0.4910407597306869,
+      "learning_rate": 1.047574933006268e-06,
+      "loss": 0.2177,
+      "step": 6015
+    },
+    {
+      "epoch": 3.4737449509521063,
+      "grad_norm": 0.47260522012547096,
+      "learning_rate": 1.0363831576172834e-06,
+      "loss": 0.2001,
+      "step": 6020
+    },
+    {
+      "epoch": 3.4766301211771493,
+      "grad_norm": 0.46807651107457027,
+      "learning_rate": 1.0252482193488345e-06,
+      "loss": 0.1936,
+      "step": 6025
+    },
+    {
+      "epoch": 3.4795152914021927,
+      "grad_norm": 0.4496096381304417,
+      "learning_rate": 1.0141701888061018e-06,
+      "loss": 0.2051,
+      "step": 6030
+    },
+    {
+      "epoch": 3.482400461627236,
+      "grad_norm": 0.48460781075602494,
+      "learning_rate": 1.0031491362334289e-06,
+      "loss": 0.2025,
+      "step": 6035
+    },
+    {
+      "epoch": 3.485285631852279,
+      "grad_norm": 0.5182489885112946,
+      "learning_rate": 9.92185131513862e-07,
+      "loss": 0.2167,
+      "step": 6040
+    },
+    {
+      "epoch": 3.4881708020773226,
+      "grad_norm": 0.48427730291008736,
+      "learning_rate": 9.812782441687263e-07,
+      "loss": 0.2024,
+      "step": 6045
+    },
+    {
+      "epoch": 3.491055972302366,
+      "grad_norm": 0.5074747506126092,
+      "learning_rate": 9.704285433571637e-07,
+      "loss": 0.2143,
+      "step": 6050
+    },
+    {
+      "epoch": 3.493941142527409,
+      "grad_norm": 0.4838143459849086,
+      "learning_rate": 9.596360978757025e-07,
+      "loss": 0.2139,
+      "step": 6055
+    },
+    {
+      "epoch": 3.4968263127524524,
+      "grad_norm": 0.47554491034770197,
+      "learning_rate": 9.489009761578294e-07,
+      "loss": 0.206,
+      "step": 6060
+    },
+    {
+      "epoch": 3.499711482977496,
+      "grad_norm": 0.5171439253070426,
+      "learning_rate": 9.382232462735463e-07,
+      "loss": 0.2044,
+      "step": 6065
+    },
+    {
+      "epoch": 3.5025966532025388,
+      "grad_norm": 0.5115632720888774,
+      "learning_rate": 9.276029759289384e-07,
+      "loss": 0.2195,
+      "step": 6070
+    },
+    {
+      "epoch": 3.505481823427582,
+      "grad_norm": 0.5128873018700219,
+      "learning_rate": 9.17040232465748e-07,
+      "loss": 0.208,
+      "step": 6075
+    },
+    {
+      "epoch": 3.5083669936526256,
+      "grad_norm": 0.47047254608971883,
+      "learning_rate": 9.065350828609509e-07,
+      "loss": 0.2166,
+      "step": 6080
+    },
+    {
+      "epoch": 3.5112521638776686,
+      "grad_norm": 0.47331390143331864,
+      "learning_rate": 8.960875937263214e-07,
+      "loss": 0.1926,
+      "step": 6085
+    },
+    {
+      "epoch": 3.514137334102712,
+      "grad_norm": 0.5526849091013493,
+      "learning_rate": 8.856978313080244e-07,
+      "loss": 0.208,
+      "step": 6090
+    },
+    {
+      "epoch": 3.5170225043277554,
+      "grad_norm": 0.5056598583710792,
+      "learning_rate": 8.753658614861782e-07,
+      "loss": 0.202,
+      "step": 6095
+    },
+    {
+      "epoch": 3.5199076745527984,
+      "grad_norm": 0.45972967026892514,
+      "learning_rate": 8.650917497744548e-07,
+      "loss": 0.2052,
+      "step": 6100
+    },
+    {
+      "epoch": 3.522792844777842,
+      "grad_norm": 0.9905171948522371,
+      "learning_rate": 8.548755613196491e-07,
+      "loss": 0.1975,
+      "step": 6105
+    },
+    {
+      "epoch": 3.5256780150028852,
+      "grad_norm": 0.487725697200175,
+      "learning_rate": 8.447173609012693e-07,
+      "loss": 0.2003,
+      "step": 6110
+    },
+    {
+      "epoch": 3.5285631852279282,
+      "grad_norm": 0.436842348667491,
+      "learning_rate": 8.346172129311436e-07,
+      "loss": 0.2002,
+      "step": 6115
+    },
+    {
+      "epoch": 3.5314483554529716,
+      "grad_norm": 0.5324360594518225,
+      "learning_rate": 8.245751814529812e-07,
+      "loss": 0.2042,
+      "step": 6120
+    },
+    {
+      "epoch": 3.534333525678015,
+      "grad_norm": 0.49501211496626846,
+      "learning_rate": 8.145913301419894e-07,
+      "loss": 0.2046,
+      "step": 6125
+    },
+    {
+      "epoch": 3.537218695903058,
+      "grad_norm": 0.5636032763069877,
+      "learning_rate": 8.046657223044618e-07,
+      "loss": 0.2065,
+      "step": 6130
+    },
+    {
+      "epoch": 3.5401038661281015,
+      "grad_norm": 0.47764396959998434,
+      "learning_rate": 7.947984208773796e-07,
+      "loss": 0.201,
+      "step": 6135
+    },
+    {
+      "epoch": 3.542989036353145,
+      "grad_norm": 0.47110105270453406,
+      "learning_rate": 7.849894884280085e-07,
+      "loss": 0.1939,
+      "step": 6140
+    },
+    {
+      "epoch": 3.545874206578188,
+      "grad_norm": 0.5089791473431078,
+      "learning_rate": 7.752389871535027e-07,
+      "loss": 0.2237,
+      "step": 6145
+    },
+    {
+      "epoch": 3.5487593768032313,
+      "grad_norm": 0.4693514320901827,
+      "learning_rate": 7.65546978880517e-07,
+      "loss": 0.1918,
+      "step": 6150
+    },
+    {
+      "epoch": 3.5516445470282747,
+      "grad_norm": 0.5339146402486223,
+      "learning_rate": 7.55913525064802e-07,
+      "loss": 0.2076,
+      "step": 6155
+    },
+    {
+      "epoch": 3.5545297172533177,
+      "grad_norm": 0.510340675728564,
+      "learning_rate": 7.463386867908318e-07,
+      "loss": 0.1947,
+      "step": 6160
+    },
+    {
+      "epoch": 3.557414887478361,
+      "grad_norm": 0.5063108850560543,
+      "learning_rate": 7.368225247713978e-07,
+      "loss": 0.2293,
+      "step": 6165
+    },
+    {
+      "epoch": 3.5603000577034045,
+      "grad_norm": 0.5151597458874256,
+      "learning_rate": 7.273650993472414e-07,
+      "loss": 0.2027,
+      "step": 6170
+    },
+    {
+      "epoch": 3.563185227928448,
+      "grad_norm": 0.6646624866313264,
+      "learning_rate": 7.179664704866551e-07,
+      "loss": 0.1955,
+      "step": 6175
+    },
+    {
+      "epoch": 3.566070398153491,
+      "grad_norm": 0.4766136250823664,
+      "learning_rate": 7.086266977851141e-07,
+      "loss": 0.2012,
+      "step": 6180
+    },
+    {
+      "epoch": 3.5689555683785343,
+      "grad_norm": 0.4840543281228228,
+      "learning_rate": 6.993458404648945e-07,
+      "loss": 0.2069,
+      "step": 6185
+    },
+    {
+      "epoch": 3.5718407386035778,
+      "grad_norm": 0.49635401265932044,
+      "learning_rate": 6.901239573746987e-07,
+      "loss": 0.2054,
+      "step": 6190
+    },
+    {
+      "epoch": 3.5747259088286207,
+      "grad_norm": 0.5178071858503266,
+      "learning_rate": 6.809611069892785e-07,
+      "loss": 0.2114,
+      "step": 6195
+    },
+    {
+      "epoch": 3.577611079053664,
+      "grad_norm": 0.4805032184640311,
+      "learning_rate": 6.718573474090673e-07,
+      "loss": 0.2088,
+      "step": 6200
+    },
+    {
+      "epoch": 3.5804962492787076,
+      "grad_norm": 0.5312256937793656,
+      "learning_rate": 6.628127363598125e-07,
+      "loss": 0.218,
+      "step": 6205
+    },
+    {
+      "epoch": 3.583381419503751,
+      "grad_norm": 0.5038683897704207,
+      "learning_rate": 6.538273311922105e-07,
+      "loss": 0.2036,
+      "step": 6210
+    },
+    {
+      "epoch": 3.586266589728794,
+      "grad_norm": 0.48973488364704354,
+      "learning_rate": 6.449011888815359e-07,
+      "loss": 0.2059,
+      "step": 6215
+    },
+    {
+      "epoch": 3.5891517599538374,
+      "grad_norm": 0.46502727902053664,
+      "learning_rate": 6.360343660272905e-07,
+      "loss": 0.2032,
+      "step": 6220
+    },
+    {
+      "epoch": 3.592036930178881,
+      "grad_norm": 0.4693993644882787,
+      "learning_rate": 6.272269188528346e-07,
+      "loss": 0.1993,
+      "step": 6225
+    },
+    {
+      "epoch": 3.594922100403924,
+      "grad_norm": 0.4759028304516473,
+      "learning_rate": 6.184789032050408e-07,
+      "loss": 0.1937,
+      "step": 6230
+    },
+    {
+      "epoch": 3.597807270628967,
+      "grad_norm": 0.5255955515623099,
+      "learning_rate": 6.097903745539258e-07,
+      "loss": 0.2051,
+      "step": 6235
+    },
+    {
+      "epoch": 3.6006924408540106,
+      "grad_norm": 0.48476442425010263,
+      "learning_rate": 6.011613879923161e-07,
+      "loss": 0.1946,
+      "step": 6240
+    },
+    {
+      "epoch": 3.6035776110790536,
+      "grad_norm": 0.5048238073660125,
+      "learning_rate": 5.925919982354844e-07,
+      "loss": 0.2122,
+      "step": 6245
+    },
+    {
+      "epoch": 3.606462781304097,
+      "grad_norm": 0.5137209328897905,
+      "learning_rate": 5.840822596208073e-07,
+      "loss": 0.2078,
+      "step": 6250
+    },
+    {
+      "epoch": 3.6093479515291405,
+      "grad_norm": 0.46932455507505844,
+      "learning_rate": 5.756322261074232e-07,
+      "loss": 0.2105,
+      "step": 6255
+    },
+    {
+      "epoch": 3.6122331217541834,
+      "grad_norm": 0.47438897533582397,
+      "learning_rate": 5.672419512758909e-07,
+      "loss": 0.2046,
+      "step": 6260
+    },
+    {
+      "epoch": 3.615118291979227,
+      "grad_norm": 0.49163449448043284,
+      "learning_rate": 5.589114883278423e-07,
+      "loss": 0.2106,
+      "step": 6265
+    },
+    {
+      "epoch": 3.6180034622042703,
+      "grad_norm": 0.4948571949148204,
+      "learning_rate": 5.506408900856508e-07,
+      "loss": 0.2019,
+      "step": 6270
+    },
+    {
+      "epoch": 3.6208886324293132,
+      "grad_norm": 0.5239387280536255,
+      "learning_rate": 5.424302089920974e-07,
+      "loss": 0.2152,
+      "step": 6275
+    },
+    {
+      "epoch": 3.6237738026543567,
+      "grad_norm": 0.4645502539433645,
+      "learning_rate": 5.342794971100374e-07,
+      "loss": 0.2013,
+      "step": 6280
+    },
+    {
+      "epoch": 3.6266589728794,
+      "grad_norm": 0.5644959774991677,
+      "learning_rate": 5.261888061220643e-07,
+      "loss": 0.2022,
+      "step": 6285
+    },
+    {
+      "epoch": 3.629544143104443,
+      "grad_norm": 0.48442583058736516,
+      "learning_rate": 5.181581873301944e-07,
+      "loss": 0.2036,
+      "step": 6290
+    },
+    {
+      "epoch": 3.6324293133294865,
+      "grad_norm": 0.515528269609854,
+      "learning_rate": 5.101876916555271e-07,
+      "loss": 0.204,
+      "step": 6295
+    },
+    {
+      "epoch": 3.63531448355453,
+      "grad_norm": 0.5338898797051952,
+      "learning_rate": 5.022773696379313e-07,
+      "loss": 0.2033,
+      "step": 6300
+    },
+    {
+      "epoch": 3.638199653779573,
+      "grad_norm": 0.4794106672832799,
+      "learning_rate": 4.944272714357223e-07,
+      "loss": 0.2061,
+      "step": 6305
+    },
+    {
+      "epoch": 3.6410848240046163,
+      "grad_norm": 0.49326404127777695,
+      "learning_rate": 4.866374468253487e-07,
+      "loss": 0.1928,
+      "step": 6310
+    },
+    {
+      "epoch": 3.6439699942296597,
+      "grad_norm": 0.5247706559589024,
+      "learning_rate": 4.789079452010648e-07,
+      "loss": 0.2058,
+      "step": 6315
+    },
+    {
+      "epoch": 3.6468551644547027,
+      "grad_norm": 0.4781901107121829,
+      "learning_rate": 4.7123881557462834e-07,
+      "loss": 0.2045,
+      "step": 6320
+    },
+    {
+      "epoch": 3.649740334679746,
+      "grad_norm": 0.4743243681177699,
+      "learning_rate": 4.6363010657498597e-07,
+      "loss": 0.2118,
+      "step": 6325
+    },
+    {
+      "epoch": 3.6526255049047895,
+      "grad_norm": 0.5180788089096849,
+      "learning_rate": 4.560818664479671e-07,
+      "loss": 0.1981,
+      "step": 6330
+    },
+    {
+      "epoch": 3.6555106751298325,
+      "grad_norm": 0.479204470483921,
+      "learning_rate": 4.485941430559726e-07,
+      "loss": 0.2242,
+      "step": 6335
+    },
+    {
+      "epoch": 3.658395845354876,
+      "grad_norm": 0.49576155614359946,
+      "learning_rate": 4.411669838776733e-07,
+      "loss": 0.1973,
+      "step": 6340
+    },
+    {
+      "epoch": 3.6612810155799194,
+      "grad_norm": 0.48614229875693404,
+      "learning_rate": 4.338004360077164e-07,
+      "loss": 0.1984,
+      "step": 6345
+    },
+    {
+      "epoch": 3.6641661858049623,
+      "grad_norm": 0.46848988610190684,
+      "learning_rate": 4.2649454615641737e-07,
+      "loss": 0.2052,
+      "step": 6350
+    },
+    {
+      "epoch": 3.6670513560300058,
+      "grad_norm": 0.5277331889272244,
+      "learning_rate": 4.1924936064946633e-07,
+      "loss": 0.1966,
+      "step": 6355
+    },
+    {
+      "epoch": 3.669936526255049,
+      "grad_norm": 0.48392147766138405,
+      "learning_rate": 4.120649254276321e-07,
+      "loss": 0.2009,
+      "step": 6360
+    },
+    {
+      "epoch": 3.672821696480092,
+      "grad_norm": 0.48495642382601156,
+      "learning_rate": 4.04941286046483e-07,
+      "loss": 0.2158,
+      "step": 6365
+    },
+    {
+      "epoch": 3.6757068667051356,
+      "grad_norm": 0.49853651987090997,
+      "learning_rate": 3.9787848767607997e-07,
+      "loss": 0.2097,
+      "step": 6370
+    },
+    {
+      "epoch": 3.678592036930179,
+      "grad_norm": 0.49229974849216307,
+      "learning_rate": 3.908765751007038e-07,
+      "loss": 0.2174,
+      "step": 6375
+    },
+    {
+      "epoch": 3.681477207155222,
+      "grad_norm": 0.49794648252416773,
+      "learning_rate": 3.839355927185662e-07,
+      "loss": 0.2119,
+      "step": 6380
+    },
+    {
+      "epoch": 3.6843623773802654,
+      "grad_norm": 0.7006970650191741,
+      "learning_rate": 3.770555845415258e-07,
+      "loss": 0.2038,
+      "step": 6385
+    },
+    {
+      "epoch": 3.687247547605309,
+      "grad_norm": 0.5062300259095932,
+      "learning_rate": 3.7023659419481275e-07,
+      "loss": 0.2026,
+      "step": 6390
+    },
+    {
+      "epoch": 3.690132717830352,
+      "grad_norm": 0.49987127156630407,
+      "learning_rate": 3.6347866491675323e-07,
+      "loss": 0.2016,
+      "step": 6395
+    },
+    {
+      "epoch": 3.693017888055395,
+      "grad_norm": 0.5019423482974231,
+      "learning_rate": 3.5678183955849323e-07,
+      "loss": 0.2123,
+      "step": 6400
+    },
+    {
+      "epoch": 3.6959030582804386,
+      "grad_norm": 0.49420552946041524,
+      "learning_rate": 3.5014616058372306e-07,
+      "loss": 0.2066,
+      "step": 6405
+    },
+    {
+      "epoch": 3.6987882285054816,
+      "grad_norm": 0.483613460902034,
+      "learning_rate": 3.435716700684133e-07,
+      "loss": 0.2142,
+      "step": 6410
+    },
+    {
+      "epoch": 3.701673398730525,
+      "grad_norm": 0.8317874119600014,
+      "learning_rate": 3.3705840970054916e-07,
+      "loss": 0.1973,
+      "step": 6415
+    },
+    {
+      "epoch": 3.7045585689555685,
+      "grad_norm": 0.4583040079124961,
+      "learning_rate": 3.3060642077986313e-07,
+      "loss": 0.2027,
+      "step": 6420
+    },
+    {
+      "epoch": 3.7074437391806114,
+      "grad_norm": 0.4698006325080622,
+      "learning_rate": 3.242157442175686e-07,
+      "loss": 0.1926,
+      "step": 6425
+    },
+    {
+      "epoch": 3.710328909405655,
+      "grad_norm": 0.4935876897031322,
+      "learning_rate": 3.1788642053610984e-07,
+      "loss": 0.1982,
+      "step": 6430
+    },
+    {
+      "epoch": 3.7132140796306983,
+      "grad_norm": 0.4871065386845346,
+      "learning_rate": 3.1161848986890117e-07,
+      "loss": 0.201,
+      "step": 6435
+    },
+    {
+      "epoch": 3.7160992498557412,
+      "grad_norm": 0.5134006486764362,
+      "learning_rate": 3.054119919600673e-07,
+      "loss": 0.2103,
+      "step": 6440
+    },
+    {
+      "epoch": 3.7189844200807847,
+      "grad_norm": 0.5247799897213314,
+      "learning_rate": 2.9926696616420227e-07,
+      "loss": 0.2158,
+      "step": 6445
+    },
+    {
+      "epoch": 3.721869590305828,
+      "grad_norm": 0.45147630286318213,
+      "learning_rate": 2.9318345144610627e-07,
+      "loss": 0.2007,
+      "step": 6450
+    },
+    {
+      "epoch": 3.724754760530871,
+      "grad_norm": 0.480778845212167,
+      "learning_rate": 2.8716148638055166e-07,
+      "loss": 0.216,
+      "step": 6455
+    },
+    {
+      "epoch": 3.7276399307559145,
+      "grad_norm": 0.49023422691624635,
+      "learning_rate": 2.8120110915202945e-07,
+      "loss": 0.2131,
+      "step": 6460
+    },
+    {
+      "epoch": 3.730525100980958,
+      "grad_norm": 0.5078622584490374,
+      "learning_rate": 2.7530235755450886e-07,
+      "loss": 0.2085,
+      "step": 6465
+    },
+    {
+      "epoch": 3.733410271206001,
+      "grad_norm": 0.527238182528253,
+      "learning_rate": 2.6946526899120693e-07,
+      "loss": 0.2076,
+      "step": 6470
+    },
+    {
+      "epoch": 3.7362954414310443,
+      "grad_norm": 0.4829671940741909,
+      "learning_rate": 2.6368988047433373e-07,
+      "loss": 0.1978,
+      "step": 6475
+    },
+    {
+      "epoch": 3.7391806116560877,
+      "grad_norm": 0.5127022353864252,
+      "learning_rate": 2.5797622862487104e-07,
+      "loss": 0.2249,
+      "step": 6480
+    },
+    {
+      "epoch": 3.742065781881131,
+      "grad_norm": 0.4378947388730543,
+      "learning_rate": 2.523243496723382e-07,
+      "loss": 0.1999,
+      "step": 6485
+    },
+    {
+      "epoch": 3.744950952106174,
+      "grad_norm": 0.4730083265761743,
+      "learning_rate": 2.467342794545613e-07,
+      "loss": 0.2036,
+      "step": 6490
+    },
+    {
+      "epoch": 3.7478361223312175,
+      "grad_norm": 0.5083579379173573,
+      "learning_rate": 2.412060534174421e-07,
+      "loss": 0.2134,
+      "step": 6495
+    },
+    {
+      "epoch": 3.750721292556261,
+      "grad_norm": 0.5014960077674261,
+      "learning_rate": 2.35739706614736e-07,
+      "loss": 0.2066,
+      "step": 6500
+    },
+    {
+      "epoch": 3.753606462781304,
+      "grad_norm": 0.505196404447519,
+      "learning_rate": 2.3033527370783459e-07,
+      "loss": 0.2118,
+      "step": 6505
+    },
+    {
+      "epoch": 3.7564916330063474,
+      "grad_norm": 0.4720849386750261,
+      "learning_rate": 2.2499278896553678e-07,
+      "loss": 0.2176,
+      "step": 6510
+    },
+    {
+      "epoch": 3.759376803231391,
+      "grad_norm": 0.49946759486920755,
+      "learning_rate": 2.1971228626384233e-07,
+      "loss": 0.2073,
+      "step": 6515
+    },
+    {
+      "epoch": 3.762261973456434,
+      "grad_norm": 0.4836314820421101,
+      "learning_rate": 2.1449379908572431e-07,
+      "loss": 0.2079,
+      "step": 6520
+    },
+    {
+      "epoch": 3.765147143681477,
+      "grad_norm": 0.4848667800466061,
+      "learning_rate": 2.0933736052092812e-07,
+      "loss": 0.2244,
+      "step": 6525
+    },
+    {
+      "epoch": 3.7680323139065206,
+      "grad_norm": 0.4735713481776197,
+      "learning_rate": 2.0424300326575497e-07,
+      "loss": 0.2018,
+      "step": 6530
+    },
+    {
+      "epoch": 3.770917484131564,
+      "grad_norm": 0.5176491575721175,
+      "learning_rate": 1.992107596228554e-07,
+      "loss": 0.2109,
+      "step": 6535
+    },
+    {
+      "epoch": 3.773802654356607,
+      "grad_norm": 0.49048701673925643,
+      "learning_rate": 1.942406615010295e-07,
+      "loss": 0.2048,
+      "step": 6540
+    },
+    {
+      "epoch": 3.7766878245816504,
+      "grad_norm": 0.4951411517869068,
+      "learning_rate": 1.8933274041501915e-07,
+      "loss": 0.1982,
+      "step": 6545
+    },
+    {
+      "epoch": 3.779572994806694,
+      "grad_norm": 0.4615884792028935,
+      "learning_rate": 1.8448702748530723e-07,
+      "loss": 0.2105,
+      "step": 6550
+    },
+    {
+      "epoch": 3.782458165031737,
+      "grad_norm": 0.503114785095133,
+      "learning_rate": 1.7970355343792434e-07,
+      "loss": 0.1995,
+      "step": 6555
+    },
+    {
+      "epoch": 3.7853433352567802,
+      "grad_norm": 0.4531549414892555,
+      "learning_rate": 1.7498234860425567e-07,
+      "loss": 0.1954,
+      "step": 6560
+    },
+    {
+      "epoch": 3.7882285054818237,
+      "grad_norm": 0.4476621701726881,
+      "learning_rate": 1.7032344292084112e-07,
+      "loss": 0.207,
+      "step": 6565
+    },
+    {
+      "epoch": 3.7911136757068666,
+      "grad_norm": 0.5301338039777519,
+      "learning_rate": 1.6572686592919107e-07,
+      "loss": 0.2047,
+      "step": 6570
+    },
+    {
+      "epoch": 3.79399884593191,
+      "grad_norm": 0.5215830928997757,
+      "learning_rate": 1.6119264677559977e-07,
+      "loss": 0.2164,
+      "step": 6575
+    },
+    {
+      "epoch": 3.7968840161569535,
+      "grad_norm": 0.4581052464950232,
+      "learning_rate": 1.5672081421095552e-07,
+      "loss": 0.1954,
+      "step": 6580
+    },
+    {
+      "epoch": 3.7997691863819965,
+      "grad_norm": 0.4658151305069496,
+      "learning_rate": 1.5231139659056426e-07,
+      "loss": 0.2065,
+      "step": 6585
+    },
+    {
+      "epoch": 3.80265435660704,
+      "grad_norm": 0.5087191746251587,
+      "learning_rate": 1.4796442187396397e-07,
+      "loss": 0.2027,
+      "step": 6590
+    },
+    {
+      "epoch": 3.8055395268320833,
+      "grad_norm": 0.5505757142185078,
+      "learning_rate": 1.4367991762475497e-07,
+      "loss": 0.2095,
+      "step": 6595
+    },
+    {
+      "epoch": 3.8084246970571263,
+      "grad_norm": 0.5414562907573153,
+      "learning_rate": 1.394579110104144e-07,
+      "loss": 0.2143,
+      "step": 6600
+    },
+    {
+      "epoch": 3.8113098672821697,
+      "grad_norm": 0.46524927624850704,
+      "learning_rate": 1.352984288021353e-07,
+      "loss": 0.1978,
+      "step": 6605
+    },
+    {
+      "epoch": 3.814195037507213,
+      "grad_norm": 0.4628457981686407,
+      "learning_rate": 1.3120149737464893e-07,
+      "loss": 0.201,
+      "step": 6610
+    },
+    {
+      "epoch": 3.817080207732256,
+      "grad_norm": 0.4966243391627818,
+      "learning_rate": 1.2716714270606057e-07,
+      "loss": 0.2002,
+      "step": 6615
+    },
+    {
+      "epoch": 3.8199653779572995,
+      "grad_norm": 0.46771609095809985,
+      "learning_rate": 1.2319539037768614e-07,
+      "loss": 0.2009,
+      "step": 6620
+    },
+    {
+      "epoch": 3.822850548182343,
+      "grad_norm": 0.47874741702870294,
+      "learning_rate": 1.192862655738858e-07,
+      "loss": 0.2091,
+      "step": 6625
+    },
+    {
+      "epoch": 3.825735718407386,
+      "grad_norm": 0.48329839687108117,
+      "learning_rate": 1.1543979308190844e-07,
+      "loss": 0.2005,
+      "step": 6630
+    },
+    {
+      "epoch": 3.8286208886324293,
+      "grad_norm": 0.4891511461415484,
+      "learning_rate": 1.1165599729173193e-07,
+      "loss": 0.2088,
+      "step": 6635
+    },
+    {
+      "epoch": 3.8315060588574728,
+      "grad_norm": 0.52591217962733,
+      "learning_rate": 1.0793490219591085e-07,
+      "loss": 0.2048,
+      "step": 6640
+    },
+    {
+      "epoch": 3.8343912290825157,
+      "grad_norm": 0.49626455120247676,
+      "learning_rate": 1.0427653138942118e-07,
+      "loss": 0.2033,
+      "step": 6645
+    },
+    {
+      "epoch": 3.837276399307559,
+      "grad_norm": 0.4673741618418967,
+      "learning_rate": 1.0068090806951369e-07,
+      "loss": 0.221,
+      "step": 6650
+    },
+    {
+      "epoch": 3.8401615695326026,
+      "grad_norm": 0.47678344134659506,
+      "learning_rate": 9.714805503556413e-08,
+      "loss": 0.2276,
+      "step": 6655
+    },
+    {
+      "epoch": 3.8430467397576455,
+      "grad_norm": 1.143709435388536,
+      "learning_rate": 9.367799468893102e-08,
+      "loss": 0.1927,
+      "step": 6660
+    },
+    {
+      "epoch": 3.845931909982689,
+      "grad_norm": 0.46684620921540665,
+      "learning_rate": 9.027074903281364e-08,
+      "loss": 0.1986,
+      "step": 6665
+    },
+    {
+      "epoch": 3.8488170802077324,
+      "grad_norm": 0.5307872034727703,
+      "learning_rate": 8.692633967210872e-08,
+      "loss": 0.2188,
+      "step": 6670
+    },
+    {
+      "epoch": 3.8517022504327754,
+      "grad_norm": 0.4735022804301449,
+      "learning_rate": 8.364478781327956e-08,
+      "loss": 0.2141,
+      "step": 6675
+    },
+    {
+      "epoch": 3.854587420657819,
+      "grad_norm": 0.45935591835994294,
+      "learning_rate": 8.042611426421598e-08,
+      "loss": 0.2084,
+      "step": 6680
+    },
+    {
+      "epoch": 3.857472590882862,
+      "grad_norm": 0.515095531011111,
+      "learning_rate": 7.727033943410678e-08,
+      "loss": 0.2099,
+      "step": 6685
+    },
+    {
+      "epoch": 3.860357761107905,
+      "grad_norm": 0.48719846805820766,
+      "learning_rate": 7.417748333330533e-08,
+      "loss": 0.2108,
+      "step": 6690
+    },
+    {
+      "epoch": 3.8632429313329486,
+      "grad_norm": 0.4535374141566299,
+      "learning_rate": 7.114756557320745e-08,
+      "loss": 0.2101,
+      "step": 6695
+    },
+    {
+      "epoch": 3.866128101557992,
+      "grad_norm": 0.48820441041592805,
+      "learning_rate": 6.818060536612381e-08,
+      "loss": 0.1999,
+      "step": 6700
+    },
+    {
+      "epoch": 3.869013271783035,
+      "grad_norm": 0.4752993276059485,
+      "learning_rate": 6.527662152516323e-08,
+      "loss": 0.1855,
+      "step": 6705
+    },
+    {
+      "epoch": 3.8718984420080784,
+      "grad_norm": 0.4986229869294503,
+      "learning_rate": 6.243563246410622e-08,
+      "loss": 0.2105,
+      "step": 6710
+    },
+    {
+      "epoch": 3.874783612233122,
+      "grad_norm": 0.49185252551299946,
+      "learning_rate": 5.965765619728947e-08,
+      "loss": 0.2239,
+      "step": 6715
+    },
+    {
+      "epoch": 3.877668782458165,
+      "grad_norm": 0.5713339630980684,
+      "learning_rate": 5.694271033950038e-08,
+      "loss": 0.2193,
+      "step": 6720
+    },
+    {
+      "epoch": 3.8805539526832082,
+      "grad_norm": 0.5344135726437659,
+      "learning_rate": 5.429081210585274e-08,
+      "loss": 0.2163,
+      "step": 6725
+    },
+    {
+      "epoch": 3.8834391229082517,
+      "grad_norm": 0.46373171517071093,
+      "learning_rate": 5.170197831168677e-08,
+      "loss": 0.2072,
+      "step": 6730
+    },
+    {
+      "epoch": 3.8863242931332946,
+      "grad_norm": 0.5108545524577744,
+      "learning_rate": 4.917622537245925e-08,
+      "loss": 0.1981,
+      "step": 6735
+    },
+    {
+      "epoch": 3.889209463358338,
+      "grad_norm": 0.5134285782488883,
+      "learning_rate": 4.671356930363691e-08,
+      "loss": 0.2115,
+      "step": 6740
+    },
+    {
+      "epoch": 3.8920946335833815,
+      "grad_norm": 0.5112672014805328,
+      "learning_rate": 4.431402572060095e-08,
+      "loss": 0.2067,
+      "step": 6745
+    },
+    {
+      "epoch": 3.8949798038084245,
+      "grad_norm": 0.5093768129247295,
+      "learning_rate": 4.197760983854271e-08,
+      "loss": 0.2094,
+      "step": 6750
+    },
+    {
+      "epoch": 3.897864974033468,
+      "grad_norm": 0.49789509988732483,
+      "learning_rate": 3.970433647236926e-08,
+      "loss": 0.2109,
+      "step": 6755
+    },
+    {
+      "epoch": 3.9007501442585113,
+      "grad_norm": 0.45027822200576423,
+      "learning_rate": 3.749422003661018e-08,
+      "loss": 0.1992,
+      "step": 6760
+    },
+    {
+      "epoch": 3.9036353144835543,
+      "grad_norm": 0.4956800653304501,
+      "learning_rate": 3.5347274545322055e-08,
+      "loss": 0.2255,
+      "step": 6765
+    },
+    {
+      "epoch": 3.9065204847085977,
+      "grad_norm": 0.5131612404624672,
+      "learning_rate": 3.3263513612006305e-08,
+      "loss": 0.2123,
+      "step": 6770
+    },
+    {
+      "epoch": 3.909405654933641,
+      "grad_norm": 0.4801277473215345,
+      "learning_rate": 3.124295044951931e-08,
+      "loss": 0.2064,
+      "step": 6775
+    },
+    {
+      "epoch": 3.912290825158684,
+      "grad_norm": 0.47685091887520487,
+      "learning_rate": 2.928559786998575e-08,
+      "loss": 0.2026,
+      "step": 6780
+    },
+    {
+      "epoch": 3.9151759953837275,
+      "grad_norm": 0.4948070608182228,
+      "learning_rate": 2.7391468284722056e-08,
+      "loss": 0.2015,
+      "step": 6785
+    },
+    {
+      "epoch": 3.918061165608771,
+      "grad_norm": 0.49666575468745433,
+      "learning_rate": 2.5560573704157543e-08,
+      "loss": 0.2127,
+      "step": 6790
+    },
+    {
+      "epoch": 3.9209463358338144,
+      "grad_norm": 0.517584053580229,
+      "learning_rate": 2.379292573775338e-08,
+      "loss": 0.2103,
+      "step": 6795
+    },
+    {
+      "epoch": 3.9238315060588573,
+      "grad_norm": 0.4715466573891517,
+      "learning_rate": 2.2088535593933756e-08,
+      "loss": 0.2034,
+      "step": 6800
+    },
+    {
+      "epoch": 3.9267166762839008,
+      "grad_norm": 0.48184559815395395,
+      "learning_rate": 2.0447414080014826e-08,
+      "loss": 0.2163,
+      "step": 6805
+    },
+    {
+      "epoch": 3.929601846508944,
+      "grad_norm": 0.47730934367308564,
+      "learning_rate": 1.8869571602132543e-08,
+      "loss": 0.2055,
+      "step": 6810
+    },
+    {
+      "epoch": 3.932487016733987,
+      "grad_norm": 0.509248139581449,
+      "learning_rate": 1.735501816517937e-08,
+      "loss": 0.2001,
+      "step": 6815
+    },
+    {
+      "epoch": 3.9353721869590306,
+      "grad_norm": 0.5087728221195549,
+      "learning_rate": 1.59037633727388e-08,
+      "loss": 0.2079,
+      "step": 6820
+    },
+    {
+      "epoch": 3.938257357184074,
+      "grad_norm": 0.4986147495940189,
+      "learning_rate": 1.4515816427029816e-08,
+      "loss": 0.2173,
+      "step": 6825
+    },
+    {
+      "epoch": 3.9411425274091174,
+      "grad_norm": 0.5024656322502186,
+      "learning_rate": 1.3191186128841404e-08,
+      "loss": 0.1988,
+      "step": 6830
+    },
+    {
+      "epoch": 3.9440276976341604,
+      "grad_norm": 0.49534516720622374,
+      "learning_rate": 1.192988087748037e-08,
+      "loss": 0.2081,
+      "step": 6835
+    },
+    {
+      "epoch": 3.946912867859204,
+      "grad_norm": 0.479255083905943,
+      "learning_rate": 1.073190867071805e-08,
+      "loss": 0.2017,
+      "step": 6840
+    },
+    {
+      "epoch": 3.9497980380842472,
+      "grad_norm": 0.4632454498993253,
+      "learning_rate": 9.597277104739233e-09,
+      "loss": 0.2044,
+      "step": 6845
+    },
+    {
+      "epoch": 3.95268320830929,
+      "grad_norm": 0.5211397474253632,
+      "learning_rate": 8.525993374095543e-09,
+      "loss": 0.2058,
+      "step": 6850
+    },
+    {
+      "epoch": 3.9555683785343336,
+      "grad_norm": 0.48574173728886655,
+      "learning_rate": 7.518064271654357e-09,
+      "loss": 0.215,
+      "step": 6855
+    },
+    {
+      "epoch": 3.958453548759377,
+      "grad_norm": 0.4668619180516313,
+      "learning_rate": 6.573496188565509e-09,
+      "loss": 0.1986,
+      "step": 6860
+    },
+    {
+      "epoch": 3.96133871898442,
+      "grad_norm": 0.47768436258947716,
+      "learning_rate": 5.6922951142079904e-09,
+      "loss": 0.1948,
+      "step": 6865
+    },
+    {
+      "epoch": 3.9642238892094634,
+      "grad_norm": 0.5138599831276915,
+      "learning_rate": 4.874466636164421e-09,
+      "loss": 0.2113,
+      "step": 6870
+    },
+    {
+      "epoch": 3.967109059434507,
+      "grad_norm": 0.5000185830706366,
+      "learning_rate": 4.120015940177746e-09,
+      "loss": 0.2256,
+      "step": 6875
+    },
+    {
+      "epoch": 3.96999422965955,
+      "grad_norm": 0.4866192841385196,
+      "learning_rate": 3.4289478101201536e-09,
+      "loss": 0.2063,
+      "step": 6880
+    },
+    {
+      "epoch": 3.9728793998845933,
+      "grad_norm": 0.47456211919715674,
+      "learning_rate": 2.8012666279653155e-09,
+      "loss": 0.2031,
+      "step": 6885
+    },
+    {
+      "epoch": 3.9757645701096367,
+      "grad_norm": 0.4973916489751191,
+      "learning_rate": 2.2369763737573046e-09,
+      "loss": 0.2013,
+      "step": 6890
+    },
+    {
+      "epoch": 3.9786497403346797,
+      "grad_norm": 0.50633061407573,
+      "learning_rate": 1.7360806255861673e-09,
+      "loss": 0.2108,
+      "step": 6895
+    },
+    {
+      "epoch": 3.981534910559723,
+      "grad_norm": 0.4761645591299464,
+      "learning_rate": 1.29858255956683e-09,
+      "loss": 0.2145,
+      "step": 6900
+    },
+    {
+      "epoch": 3.9844200807847665,
+      "grad_norm": 0.4667942094809914,
+      "learning_rate": 9.244849498168951e-10,
+      "loss": 0.1998,
+      "step": 6905
+    },
+    {
+      "epoch": 3.9873052510098095,
+      "grad_norm": 0.4758585527520601,
+      "learning_rate": 6.137901684399871e-10,
+      "loss": 0.2103,
+      "step": 6910
+    },
+    {
+      "epoch": 3.990190421234853,
+      "grad_norm": 0.5058399856877976,
+      "learning_rate": 3.665001855113204e-10,
+      "loss": 0.2037,
+      "step": 6915
+    },
+    {
+      "epoch": 3.9930755914598963,
+      "grad_norm": 0.4840062336803948,
+      "learning_rate": 1.8261656906437553e-10,
+      "loss": 0.2114,
+      "step": 6920
+    },
+    {
+      "epoch": 3.9959607616849393,
+      "grad_norm": 0.49250096214422046,
+      "learning_rate": 6.21404850809082e-11,
+      "loss": 0.2046,
+      "step": 6925
+    },
+    {
+      "epoch": 3.9988459319099827,
+      "grad_norm": 0.521991079354229,
+      "learning_rate": 5.0726974842874035e-12,
+      "loss": 0.2225,
+      "step": 6930
+    },
+    {
+      "epoch": 4.0,
+      "eval_loss": 0.5301879048347473,
+      "eval_runtime": 0.5871,
+      "eval_samples_per_second": 131.147,
+      "eval_steps_per_second": 3.406,
+      "step": 6932
+    },
+    {
+      "epoch": 4.0,
+      "step": 6932,
+      "total_flos": 340525408321536.0,
+      "train_loss": 0.3252074327502578,
+      "train_runtime": 5343.4474,
+      "train_samples_per_second": 41.505,
+      "train_steps_per_second": 1.297
+    }
+  ],
+  "logging_steps": 5,
+  "max_steps": 6932,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 4,
+  "save_steps": 694,
+  "total_flos": 340525408321536.0,
+  "train_batch_size": 4,
+  "trial_name": null,
+  "trial_params": null
+}