{
  "best_metric": 0.36434125900268555,
  "best_model_checkpoint": "./vit-base-beans/checkpoint-3200",
  "epoch": 2.0,
  "eval_steps": 100,
  "global_step": 3228,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.006195786864931847,
      "grad_norm": 1.9642935991287231,
      "learning_rate": 0.00019938042131350682,
      "loss": 3.9212,
      "step": 10
    },
    {
      "epoch": 0.012391573729863693,
      "grad_norm": 1.8843648433685303,
      "learning_rate": 0.00019876084262701366,
      "loss": 4.1331,
      "step": 20
    },
    {
      "epoch": 0.01858736059479554,
      "grad_norm": 1.7067281007766724,
      "learning_rate": 0.00019814126394052047,
      "loss": 4.0633,
      "step": 30
    },
    {
      "epoch": 0.024783147459727387,
      "grad_norm": 1.7453334331512451,
      "learning_rate": 0.00019752168525402728,
      "loss": 3.9709,
      "step": 40
    },
    {
      "epoch": 0.030978934324659233,
      "grad_norm": 2.055915355682373,
      "learning_rate": 0.0001969021065675341,
      "loss": 3.8473,
      "step": 50
    },
    {
      "epoch": 0.03717472118959108,
      "grad_norm": 1.8063888549804688,
      "learning_rate": 0.00019628252788104092,
      "loss": 3.7734,
      "step": 60
    },
    {
      "epoch": 0.04337050805452292,
      "grad_norm": 1.8397654294967651,
      "learning_rate": 0.0001956629491945477,
      "loss": 3.8279,
      "step": 70
    },
    {
      "epoch": 0.04956629491945477,
      "grad_norm": 2.0568203926086426,
      "learning_rate": 0.00019504337050805452,
      "loss": 3.7613,
      "step": 80
    },
    {
      "epoch": 0.055762081784386616,
      "grad_norm": 2.3892436027526855,
      "learning_rate": 0.00019442379182156135,
      "loss": 3.6098,
      "step": 90
    },
    {
      "epoch": 0.061957868649318466,
      "grad_norm": 2.0522940158843994,
      "learning_rate": 0.00019380421313506816,
      "loss": 3.5924,
      "step": 100
    },
    {
      "epoch": 0.061957868649318466,
      "eval_accuracy": 0.1926889714993804,
      "eval_loss": 3.567479133605957,
      "eval_runtime": 134.6311,
      "eval_samples_per_second": 47.953,
      "eval_steps_per_second": 5.994,
      "step": 100
    },
    {
      "epoch": 0.06815365551425032,
      "grad_norm": 2.773609161376953,
      "learning_rate": 0.00019318463444857497,
      "loss": 3.5271,
      "step": 110
    },
    {
      "epoch": 0.07434944237918216,
      "grad_norm": 2.709200620651245,
      "learning_rate": 0.00019256505576208178,
      "loss": 3.5848,
      "step": 120
    },
    {
      "epoch": 0.080545229244114,
      "grad_norm": 2.2230889797210693,
      "learning_rate": 0.00019194547707558862,
      "loss": 3.419,
      "step": 130
    },
    {
      "epoch": 0.08674101610904585,
      "grad_norm": 2.003593683242798,
      "learning_rate": 0.00019132589838909543,
      "loss": 3.4474,
      "step": 140
    },
    {
      "epoch": 0.09293680297397769,
      "grad_norm": 2.4022834300994873,
      "learning_rate": 0.00019070631970260224,
      "loss": 3.3324,
      "step": 150
    },
    {
      "epoch": 0.09913258983890955,
      "grad_norm": 2.2967748641967773,
      "learning_rate": 0.00019008674101610905,
      "loss": 3.1463,
      "step": 160
    },
    {
      "epoch": 0.10532837670384139,
      "grad_norm": 2.498286724090576,
      "learning_rate": 0.0001894671623296159,
      "loss": 3.2931,
      "step": 170
    },
    {
      "epoch": 0.11152416356877323,
      "grad_norm": 2.418286085128784,
      "learning_rate": 0.0001888475836431227,
      "loss": 3.24,
      "step": 180
    },
    {
      "epoch": 0.11771995043370508,
      "grad_norm": 2.7124204635620117,
      "learning_rate": 0.0001882280049566295,
      "loss": 3.1876,
      "step": 190
    },
    {
      "epoch": 0.12391573729863693,
      "grad_norm": 2.0705409049987793,
      "learning_rate": 0.00018760842627013632,
      "loss": 3.0189,
      "step": 200
    },
    {
      "epoch": 0.12391573729863693,
      "eval_accuracy": 0.30467781908302355,
      "eval_loss": 3.0312585830688477,
      "eval_runtime": 106.8938,
      "eval_samples_per_second": 60.396,
      "eval_steps_per_second": 7.55,
      "step": 200
    },
    {
      "epoch": 0.13011152416356878,
      "grad_norm": 2.31803822517395,
      "learning_rate": 0.00018698884758364313,
      "loss": 2.8399,
      "step": 210
    },
    {
      "epoch": 0.13630731102850063,
      "grad_norm": 2.759551525115967,
      "learning_rate": 0.00018636926889714994,
      "loss": 2.9606,
      "step": 220
    },
    {
      "epoch": 0.14250309789343246,
      "grad_norm": 2.6668756008148193,
      "learning_rate": 0.00018574969021065675,
      "loss": 2.8822,
      "step": 230
    },
    {
      "epoch": 0.14869888475836432,
      "grad_norm": 2.4296159744262695,
      "learning_rate": 0.00018513011152416359,
      "loss": 3.0186,
      "step": 240
    },
    {
      "epoch": 0.15489467162329615,
      "grad_norm": 2.854280948638916,
      "learning_rate": 0.0001845105328376704,
      "loss": 2.85,
      "step": 250
    },
    {
      "epoch": 0.161090458488228,
      "grad_norm": 2.2501845359802246,
      "learning_rate": 0.0001838909541511772,
      "loss": 2.7842,
      "step": 260
    },
    {
      "epoch": 0.16728624535315986,
      "grad_norm": 2.2110307216644287,
      "learning_rate": 0.00018327137546468402,
      "loss": 2.6941,
      "step": 270
    },
    {
      "epoch": 0.1734820322180917,
      "grad_norm": 3.233548164367676,
      "learning_rate": 0.00018265179677819085,
      "loss": 2.7383,
      "step": 280
    },
    {
      "epoch": 0.17967781908302355,
      "grad_norm": 3.4657251834869385,
      "learning_rate": 0.00018203221809169766,
      "loss": 2.8092,
      "step": 290
    },
    {
      "epoch": 0.18587360594795538,
      "grad_norm": 2.545624256134033,
      "learning_rate": 0.00018141263940520447,
      "loss": 2.5541,
      "step": 300
    },
    {
      "epoch": 0.18587360594795538,
      "eval_accuracy": 0.3956009913258984,
      "eval_loss": 2.557483434677124,
      "eval_runtime": 106.1092,
      "eval_samples_per_second": 60.843,
      "eval_steps_per_second": 7.605,
      "step": 300
    },
    {
      "epoch": 0.19206939281288724,
      "grad_norm": 3.7214767932891846,
      "learning_rate": 0.00018079306071871128,
      "loss": 2.673,
      "step": 310
    },
    {
      "epoch": 0.1982651796778191,
      "grad_norm": 2.865368604660034,
      "learning_rate": 0.00018017348203221812,
      "loss": 2.4952,
      "step": 320
    },
    {
      "epoch": 0.20446096654275092,
      "grad_norm": 2.8324084281921387,
      "learning_rate": 0.0001795539033457249,
      "loss": 2.5799,
      "step": 330
    },
    {
      "epoch": 0.21065675340768278,
      "grad_norm": 2.980775833129883,
      "learning_rate": 0.00017893432465923171,
      "loss": 2.518,
      "step": 340
    },
    {
      "epoch": 0.21685254027261464,
      "grad_norm": 3.9543583393096924,
      "learning_rate": 0.00017831474597273855,
      "loss": 2.3632,
      "step": 350
    },
    {
      "epoch": 0.22304832713754646,
      "grad_norm": 6.1679368019104,
      "learning_rate": 0.00017769516728624536,
      "loss": 2.4168,
      "step": 360
    },
    {
      "epoch": 0.22924411400247832,
      "grad_norm": 3.2867679595947266,
      "learning_rate": 0.00017707558859975217,
      "loss": 2.4561,
      "step": 370
    },
    {
      "epoch": 0.23543990086741015,
      "grad_norm": 3.4070069789886475,
      "learning_rate": 0.00017645600991325898,
      "loss": 2.3094,
      "step": 380
    },
    {
      "epoch": 0.241635687732342,
      "grad_norm": 2.8205995559692383,
      "learning_rate": 0.00017583643122676582,
      "loss": 2.2237,
      "step": 390
    },
    {
      "epoch": 0.24783147459727387,
      "grad_norm": 3.1000003814697266,
      "learning_rate": 0.00017521685254027263,
      "loss": 2.114,
      "step": 400
    },
    {
      "epoch": 0.24783147459727387,
      "eval_accuracy": 0.45709417596034696,
      "eval_loss": 2.233164072036743,
      "eval_runtime": 106.5812,
      "eval_samples_per_second": 60.574,
      "eval_steps_per_second": 7.572,
      "step": 400
    },
    {
      "epoch": 0.2540272614622057,
      "grad_norm": 3.512111186981201,
      "learning_rate": 0.00017459727385377944,
      "loss": 2.1117,
      "step": 410
    },
    {
      "epoch": 0.26022304832713755,
      "grad_norm": 4.639825344085693,
      "learning_rate": 0.00017397769516728625,
      "loss": 2.0458,
      "step": 420
    },
    {
      "epoch": 0.2664188351920694,
      "grad_norm": 3.5039663314819336,
      "learning_rate": 0.00017335811648079309,
      "loss": 2.1726,
      "step": 430
    },
    {
      "epoch": 0.27261462205700127,
      "grad_norm": 3.6471774578094482,
      "learning_rate": 0.0001727385377942999,
      "loss": 2.1414,
      "step": 440
    },
    {
      "epoch": 0.2788104089219331,
      "grad_norm": 4.325891971588135,
      "learning_rate": 0.0001721189591078067,
      "loss": 2.0651,
      "step": 450
    },
    {
      "epoch": 0.2850061957868649,
      "grad_norm": 2.3341152667999268,
      "learning_rate": 0.00017149938042131352,
      "loss": 2.0514,
      "step": 460
    },
    {
      "epoch": 0.29120198265179675,
      "grad_norm": 3.562957525253296,
      "learning_rate": 0.00017087980173482033,
      "loss": 1.8995,
      "step": 470
    },
    {
      "epoch": 0.29739776951672864,
      "grad_norm": 4.375968933105469,
      "learning_rate": 0.00017026022304832714,
      "loss": 1.981,
      "step": 480
    },
    {
      "epoch": 0.30359355638166047,
      "grad_norm": 2.5124247074127197,
      "learning_rate": 0.00016964064436183395,
      "loss": 1.8624,
      "step": 490
    },
    {
      "epoch": 0.3097893432465923,
      "grad_norm": 4.64390230178833,
      "learning_rate": 0.00016902106567534078,
      "loss": 1.9624,
      "step": 500
    },
    {
      "epoch": 0.3097893432465923,
      "eval_accuracy": 0.559634448574969,
      "eval_loss": 1.945489764213562,
      "eval_runtime": 107.2688,
      "eval_samples_per_second": 60.185,
      "eval_steps_per_second": 7.523,
      "step": 500
    },
    {
      "epoch": 0.3159851301115242,
      "grad_norm": 3.2166125774383545,
      "learning_rate": 0.0001684014869888476,
      "loss": 1.8621,
      "step": 510
    },
    {
      "epoch": 0.322180916976456,
      "grad_norm": 3.2079391479492188,
      "learning_rate": 0.0001677819083023544,
      "loss": 2.048,
      "step": 520
    },
    {
      "epoch": 0.32837670384138784,
      "grad_norm": 4.8925652503967285,
      "learning_rate": 0.0001671623296158612,
      "loss": 1.9486,
      "step": 530
    },
    {
      "epoch": 0.3345724907063197,
      "grad_norm": 3.3474984169006348,
      "learning_rate": 0.00016654275092936805,
      "loss": 1.9194,
      "step": 540
    },
    {
      "epoch": 0.34076827757125155,
      "grad_norm": 3.6939406394958496,
      "learning_rate": 0.00016592317224287486,
      "loss": 1.8698,
      "step": 550
    },
    {
      "epoch": 0.3469640644361834,
      "grad_norm": 3.176316022872925,
      "learning_rate": 0.00016530359355638167,
      "loss": 1.7831,
      "step": 560
    },
    {
      "epoch": 0.35315985130111527,
      "grad_norm": 3.4785337448120117,
      "learning_rate": 0.00016468401486988848,
      "loss": 1.7184,
      "step": 570
    },
    {
      "epoch": 0.3593556381660471,
      "grad_norm": 2.730374336242676,
      "learning_rate": 0.00016406443618339532,
      "loss": 1.6898,
      "step": 580
    },
    {
      "epoch": 0.3655514250309789,
      "grad_norm": 3.3291196823120117,
      "learning_rate": 0.0001634448574969021,
      "loss": 1.6437,
      "step": 590
    },
    {
      "epoch": 0.37174721189591076,
      "grad_norm": 4.465322017669678,
      "learning_rate": 0.0001628252788104089,
      "loss": 1.6749,
      "step": 600
    },
    {
      "epoch": 0.37174721189591076,
      "eval_accuracy": 0.5786864931846345,
      "eval_loss": 1.7369675636291504,
      "eval_runtime": 106.0837,
      "eval_samples_per_second": 60.858,
      "eval_steps_per_second": 7.607,
      "step": 600
    },
    {
      "epoch": 0.37794299876084264,
      "grad_norm": 4.176516056060791,
      "learning_rate": 0.00016220570012391575,
      "loss": 1.6242,
      "step": 610
    },
    {
      "epoch": 0.38413878562577447,
      "grad_norm": 4.0990495681762695,
      "learning_rate": 0.00016158612143742256,
      "loss": 1.774,
      "step": 620
    },
    {
      "epoch": 0.3903345724907063,
      "grad_norm": 5.12111234664917,
      "learning_rate": 0.00016096654275092937,
      "loss": 1.7558,
      "step": 630
    },
    {
      "epoch": 0.3965303593556382,
      "grad_norm": 4.448328018188477,
      "learning_rate": 0.00016034696406443618,
      "loss": 1.6905,
      "step": 640
    },
    {
      "epoch": 0.40272614622057,
      "grad_norm": 4.468796253204346,
      "learning_rate": 0.00015972738537794301,
      "loss": 1.5202,
      "step": 650
    },
    {
      "epoch": 0.40892193308550184,
      "grad_norm": 4.141849517822266,
      "learning_rate": 0.00015910780669144982,
      "loss": 1.5273,
      "step": 660
    },
    {
      "epoch": 0.41511771995043373,
      "grad_norm": 4.007165431976318,
      "learning_rate": 0.00015848822800495664,
      "loss": 1.5843,
      "step": 670
    },
    {
      "epoch": 0.42131350681536556,
      "grad_norm": 3.8219428062438965,
      "learning_rate": 0.00015786864931846345,
      "loss": 1.6011,
      "step": 680
    },
    {
      "epoch": 0.4275092936802974,
      "grad_norm": 4.7010345458984375,
      "learning_rate": 0.00015724907063197028,
      "loss": 1.6389,
      "step": 690
    },
    {
      "epoch": 0.43370508054522927,
      "grad_norm": 3.79860258102417,
      "learning_rate": 0.0001566294919454771,
      "loss": 1.5852,
      "step": 700
    },
    {
      "epoch": 0.43370508054522927,
      "eval_accuracy": 0.6438971499380421,
      "eval_loss": 1.4947177171707153,
      "eval_runtime": 106.0976,
      "eval_samples_per_second": 60.85,
      "eval_steps_per_second": 7.606,
      "step": 700
    },
    {
      "epoch": 0.4399008674101611,
      "grad_norm": 3.5789403915405273,
      "learning_rate": 0.0001560099132589839,
      "loss": 1.526,
      "step": 710
    },
    {
      "epoch": 0.44609665427509293,
      "grad_norm": 4.297870635986328,
      "learning_rate": 0.0001553903345724907,
      "loss": 1.5452,
      "step": 720
    },
    {
      "epoch": 0.45229244114002476,
      "grad_norm": 5.054442882537842,
      "learning_rate": 0.00015477075588599752,
      "loss": 1.3529,
      "step": 730
    },
    {
      "epoch": 0.45848822800495664,
      "grad_norm": 5.724175930023193,
      "learning_rate": 0.00015415117719950433,
      "loss": 1.4825,
      "step": 740
    },
    {
      "epoch": 0.4646840148698885,
      "grad_norm": 3.8260886669158936,
      "learning_rate": 0.00015353159851301114,
      "loss": 1.4692,
      "step": 750
    },
    {
      "epoch": 0.4708798017348203,
      "grad_norm": 3.232948064804077,
      "learning_rate": 0.00015291201982651798,
      "loss": 1.43,
      "step": 760
    },
    {
      "epoch": 0.4770755885997522,
      "grad_norm": 6.948119163513184,
      "learning_rate": 0.0001522924411400248,
      "loss": 1.5213,
      "step": 770
    },
    {
      "epoch": 0.483271375464684,
      "grad_norm": 5.678015232086182,
      "learning_rate": 0.0001516728624535316,
      "loss": 1.2681,
      "step": 780
    },
    {
      "epoch": 0.48946716232961585,
      "grad_norm": 3.9260432720184326,
      "learning_rate": 0.0001510532837670384,
      "loss": 1.4108,
      "step": 790
    },
    {
      "epoch": 0.49566294919454773,
      "grad_norm": 5.349125862121582,
      "learning_rate": 0.00015043370508054525,
      "loss": 1.1875,
      "step": 800
    },
    {
      "epoch": 0.49566294919454773,
      "eval_accuracy": 0.6468401486988847,
      "eval_loss": 1.4151387214660645,
      "eval_runtime": 105.6692,
      "eval_samples_per_second": 61.096,
      "eval_steps_per_second": 7.637,
      "step": 800
    },
    {
      "epoch": 0.5018587360594795,
      "grad_norm": 4.983338832855225,
      "learning_rate": 0.00014981412639405206,
      "loss": 1.3196,
      "step": 810
    },
    {
      "epoch": 0.5080545229244114,
      "grad_norm": 5.539896488189697,
      "learning_rate": 0.00014919454770755887,
      "loss": 1.4704,
      "step": 820
    },
    {
      "epoch": 0.5142503097893433,
      "grad_norm": 4.018556594848633,
      "learning_rate": 0.0001485749690210657,
      "loss": 1.2208,
      "step": 830
    },
    {
      "epoch": 0.5204460966542751,
      "grad_norm": 4.443668842315674,
      "learning_rate": 0.00014795539033457251,
      "loss": 1.2335,
      "step": 840
    },
    {
      "epoch": 0.5266418835192069,
      "grad_norm": 6.135488033294678,
      "learning_rate": 0.0001473358116480793,
      "loss": 1.4955,
      "step": 850
    },
    {
      "epoch": 0.5328376703841388,
      "grad_norm": 5.287814617156982,
      "learning_rate": 0.0001467162329615861,
      "loss": 1.2792,
      "step": 860
    },
    {
      "epoch": 0.5390334572490706,
      "grad_norm": 5.035948276519775,
      "learning_rate": 0.00014609665427509294,
      "loss": 1.2946,
      "step": 870
    },
    {
      "epoch": 0.5452292441140025,
      "grad_norm": 4.492195129394531,
      "learning_rate": 0.00014547707558859975,
      "loss": 1.2686,
      "step": 880
    },
    {
      "epoch": 0.5514250309789344,
      "grad_norm": 4.308188438415527,
      "learning_rate": 0.00014485749690210656,
      "loss": 1.3539,
      "step": 890
    },
    {
      "epoch": 0.5576208178438662,
      "grad_norm": 5.699028015136719,
      "learning_rate": 0.00014423791821561337,
      "loss": 1.5114,
      "step": 900
    },
    {
      "epoch": 0.5576208178438662,
      "eval_accuracy": 0.682001239157373,
      "eval_loss": 1.2709109783172607,
      "eval_runtime": 105.9622,
      "eval_samples_per_second": 60.927,
      "eval_steps_per_second": 7.616,
      "step": 900
    },
    {
      "epoch": 0.563816604708798,
      "grad_norm": 3.6500446796417236,
      "learning_rate": 0.0001436183395291202,
      "loss": 1.1699,
      "step": 910
    },
    {
      "epoch": 0.5700123915737298,
      "grad_norm": 5.26973819732666,
      "learning_rate": 0.00014299876084262702,
      "loss": 1.2285,
      "step": 920
    },
    {
      "epoch": 0.5762081784386617,
      "grad_norm": 5.995537757873535,
      "learning_rate": 0.00014237918215613383,
      "loss": 1.2008,
      "step": 930
    },
    {
      "epoch": 0.5824039653035935,
      "grad_norm": 4.208325386047363,
      "learning_rate": 0.00014175960346964067,
      "loss": 0.9493,
      "step": 940
    },
    {
      "epoch": 0.5885997521685254,
      "grad_norm": 4.716500282287598,
      "learning_rate": 0.00014114002478314748,
      "loss": 1.1492,
      "step": 950
    },
    {
      "epoch": 0.5947955390334573,
      "grad_norm": 3.3192636966705322,
      "learning_rate": 0.0001405204460966543,
      "loss": 1.3701,
      "step": 960
    },
    {
      "epoch": 0.6009913258983891,
      "grad_norm": 4.523627758026123,
      "learning_rate": 0.0001399008674101611,
      "loss": 1.1221,
      "step": 970
    },
    {
      "epoch": 0.6071871127633209,
      "grad_norm": 4.424323081970215,
      "learning_rate": 0.0001392812887236679,
      "loss": 1.2023,
      "step": 980
    },
    {
      "epoch": 0.6133828996282528,
      "grad_norm": 5.295013427734375,
      "learning_rate": 0.00013866171003717472,
      "loss": 1.0532,
      "step": 990
    },
    {
      "epoch": 0.6195786864931846,
      "grad_norm": 4.994614124298096,
      "learning_rate": 0.00013804213135068153,
      "loss": 1.3122,
      "step": 1000
    },
    {
      "epoch": 0.6195786864931846,
      "eval_accuracy": 0.6939281288723668,
      "eval_loss": 1.1940184831619263,
      "eval_runtime": 106.4084,
      "eval_samples_per_second": 60.672,
      "eval_steps_per_second": 7.584,
      "step": 1000
    },
    {
      "epoch": 0.6257744733581165,
      "grad_norm": 2.611729145050049,
      "learning_rate": 0.00013742255266418837,
      "loss": 0.9277,
      "step": 1010
    },
    {
      "epoch": 0.6319702602230484,
      "grad_norm": 2.6009323596954346,
      "learning_rate": 0.00013680297397769518,
      "loss": 1.0508,
      "step": 1020
    },
    {
      "epoch": 0.6381660470879802,
      "grad_norm": 3.2199230194091797,
      "learning_rate": 0.00013618339529120199,
      "loss": 1.11,
      "step": 1030
    },
    {
      "epoch": 0.644361833952912,
      "grad_norm": 2.762774705886841,
      "learning_rate": 0.0001355638166047088,
      "loss": 1.0821,
      "step": 1040
    },
    {
      "epoch": 0.6505576208178439,
      "grad_norm": 3.8875350952148438,
      "learning_rate": 0.00013494423791821563,
      "loss": 1.0406,
      "step": 1050
    },
    {
      "epoch": 0.6567534076827757,
      "grad_norm": 4.314332008361816,
      "learning_rate": 0.00013432465923172244,
      "loss": 1.1308,
      "step": 1060
    },
    {
      "epoch": 0.6629491945477075,
      "grad_norm": 3.6009325981140137,
      "learning_rate": 0.00013370508054522925,
      "loss": 1.1907,
      "step": 1070
    },
    {
      "epoch": 0.6691449814126395,
      "grad_norm": 5.977869033813477,
      "learning_rate": 0.00013308550185873606,
      "loss": 1.1071,
      "step": 1080
    },
    {
      "epoch": 0.6753407682775713,
      "grad_norm": 5.390667915344238,
      "learning_rate": 0.0001324659231722429,
      "loss": 1.0829,
      "step": 1090
    },
    {
      "epoch": 0.6815365551425031,
      "grad_norm": 6.445249080657959,
      "learning_rate": 0.0001318463444857497,
      "loss": 1.0721,
      "step": 1100
    },
    {
      "epoch": 0.6815365551425031,
      "eval_accuracy": 0.7261462205700124,
      "eval_loss": 1.0756527185440063,
      "eval_runtime": 104.6325,
      "eval_samples_per_second": 61.702,
      "eval_steps_per_second": 7.713,
      "step": 1100
    },
    {
      "epoch": 0.6877323420074349,
      "grad_norm": 3.1488468647003174,
      "learning_rate": 0.0001312267657992565,
      "loss": 0.9928,
      "step": 1110
    },
    {
      "epoch": 0.6939281288723668,
      "grad_norm": 3.6752538681030273,
      "learning_rate": 0.00013060718711276333,
      "loss": 0.8412,
      "step": 1120
    },
    {
      "epoch": 0.7001239157372986,
      "grad_norm": 6.186413764953613,
      "learning_rate": 0.00012998760842627014,
      "loss": 0.8979,
      "step": 1130
    },
    {
      "epoch": 0.7063197026022305,
      "grad_norm": 4.457529544830322,
      "learning_rate": 0.00012936802973977695,
      "loss": 1.0586,
      "step": 1140
    },
    {
      "epoch": 0.7125154894671624,
      "grad_norm": 3.7016208171844482,
      "learning_rate": 0.00012874845105328376,
      "loss": 0.7632,
      "step": 1150
    },
    {
      "epoch": 0.7187112763320942,
      "grad_norm": 3.913440227508545,
      "learning_rate": 0.0001281288723667906,
      "loss": 0.9905,
      "step": 1160
    },
    {
      "epoch": 0.724907063197026,
      "grad_norm": 4.720458984375,
      "learning_rate": 0.0001275092936802974,
      "loss": 0.8703,
      "step": 1170
    },
    {
      "epoch": 0.7311028500619579,
      "grad_norm": 4.232792854309082,
      "learning_rate": 0.00012688971499380422,
      "loss": 1.2989,
      "step": 1180
    },
    {
      "epoch": 0.7372986369268897,
      "grad_norm": 5.886707305908203,
      "learning_rate": 0.00012627013630731103,
      "loss": 1.1108,
      "step": 1190
    },
    {
      "epoch": 0.7434944237918215,
      "grad_norm": 3.9497408866882324,
      "learning_rate": 0.00012565055762081787,
      "loss": 0.8249,
      "step": 1200
    },
    {
      "epoch": 0.7434944237918215,
      "eval_accuracy": 0.7575898389095415,
      "eval_loss": 0.9666171669960022,
      "eval_runtime": 106.5314,
      "eval_samples_per_second": 60.602,
      "eval_steps_per_second": 7.575,
      "step": 1200
    },
    {
      "epoch": 0.7496902106567535,
      "grad_norm": 5.659217357635498,
      "learning_rate": 0.00012503097893432468,
      "loss": 0.8484,
      "step": 1210
    },
    {
      "epoch": 0.7558859975216853,
      "grad_norm": 4.372608661651611,
      "learning_rate": 0.00012441140024783149,
      "loss": 0.9891,
      "step": 1220
    },
    {
      "epoch": 0.7620817843866171,
      "grad_norm": 7.296494007110596,
      "learning_rate": 0.0001237918215613383,
      "loss": 0.919,
      "step": 1230
    },
    {
      "epoch": 0.7682775712515489,
      "grad_norm": 2.1547601222991943,
      "learning_rate": 0.0001231722428748451,
      "loss": 0.8611,
      "step": 1240
    },
    {
      "epoch": 0.7744733581164808,
      "grad_norm": 4.57709264755249,
      "learning_rate": 0.00012255266418835192,
      "loss": 0.893,
      "step": 1250
    },
    {
      "epoch": 0.7806691449814126,
      "grad_norm": 4.922760009765625,
      "learning_rate": 0.00012193308550185874,
      "loss": 0.9478,
      "step": 1260
    },
    {
      "epoch": 0.7868649318463445,
      "grad_norm": 4.814465045928955,
      "learning_rate": 0.00012131350681536555,
      "loss": 0.9416,
      "step": 1270
    },
    {
      "epoch": 0.7930607187112764,
      "grad_norm": 5.223793029785156,
      "learning_rate": 0.00012069392812887237,
      "loss": 0.7468,
      "step": 1280
    },
    {
      "epoch": 0.7992565055762082,
      "grad_norm": 7.333277702331543,
      "learning_rate": 0.00012007434944237918,
      "loss": 0.8581,
      "step": 1290
    },
    {
      "epoch": 0.80545229244114,
      "grad_norm": 2.8909173011779785,
      "learning_rate": 0.000119454770755886,
      "loss": 0.7944,
      "step": 1300
    },
    {
      "epoch": 0.80545229244114,
      "eval_accuracy": 0.7707558859975217,
      "eval_loss": 0.9101163744926453,
      "eval_runtime": 106.3936,
      "eval_samples_per_second": 60.68,
      "eval_steps_per_second": 7.585,
      "step": 1300
    },
    {
      "epoch": 0.8116480793060719,
      "grad_norm": 5.546220302581787,
      "learning_rate": 0.00011883519206939282,
      "loss": 1.066,
      "step": 1310
    },
    {
      "epoch": 0.8178438661710037,
      "grad_norm": 5.582309246063232,
      "learning_rate": 0.00011821561338289964,
      "loss": 0.7992,
      "step": 1320
    },
    {
      "epoch": 0.8240396530359355,
      "grad_norm": 6.598534107208252,
      "learning_rate": 0.00011759603469640645,
      "loss": 0.8142,
      "step": 1330
    },
    {
      "epoch": 0.8302354399008675,
      "grad_norm": 3.366227865219116,
      "learning_rate": 0.00011697645600991327,
      "loss": 0.8297,
      "step": 1340
    },
    {
      "epoch": 0.8364312267657993,
      "grad_norm": 4.621030807495117,
      "learning_rate": 0.00011635687732342008,
      "loss": 0.8083,
      "step": 1350
    },
    {
      "epoch": 0.8426270136307311,
      "grad_norm": 4.285297870635986,
      "learning_rate": 0.00011573729863692691,
      "loss": 0.8572,
      "step": 1360
    },
    {
      "epoch": 0.8488228004956629,
      "grad_norm": 5.127432823181152,
      "learning_rate": 0.0001151177199504337,
      "loss": 0.7723,
      "step": 1370
    },
    {
      "epoch": 0.8550185873605948,
      "grad_norm": 8.046000480651855,
      "learning_rate": 0.00011449814126394051,
      "loss": 0.6129,
      "step": 1380
    },
    {
      "epoch": 0.8612143742255266,
      "grad_norm": 3.8149867057800293,
      "learning_rate": 0.00011387856257744734,
      "loss": 0.7718,
      "step": 1390
    },
    {
      "epoch": 0.8674101610904585,
      "grad_norm": 7.305781364440918,
      "learning_rate": 0.00011325898389095415,
      "loss": 0.8032,
      "step": 1400
    },
    {
      "epoch": 0.8674101610904585,
      "eval_accuracy": 0.7690520446096655,
      "eval_loss": 0.901136040687561,
      "eval_runtime": 106.4417,
      "eval_samples_per_second": 60.653,
      "eval_steps_per_second": 7.582,
      "step": 1400
    },
    {
      "epoch": 0.8736059479553904,
      "grad_norm": 5.850032329559326,
      "learning_rate": 0.00011263940520446097,
      "loss": 0.9454,
      "step": 1410
    },
    {
      "epoch": 0.8798017348203222,
      "grad_norm": 5.336400032043457,
      "learning_rate": 0.00011201982651796778,
      "loss": 0.7002,
      "step": 1420
    },
    {
      "epoch": 0.885997521685254,
      "grad_norm": 3.1872918605804443,
      "learning_rate": 0.0001114002478314746,
      "loss": 0.8848,
      "step": 1430
    },
    {
      "epoch": 0.8921933085501859,
      "grad_norm": 5.940222263336182,
      "learning_rate": 0.00011078066914498142,
      "loss": 0.8746,
      "step": 1440
    },
    {
      "epoch": 0.8983890954151177,
      "grad_norm": 4.074731349945068,
      "learning_rate": 0.00011016109045848824,
      "loss": 0.9426,
      "step": 1450
    },
    {
      "epoch": 0.9045848822800495,
      "grad_norm": 4.467647552490234,
      "learning_rate": 0.00010954151177199505,
      "loss": 0.7319,
      "step": 1460
    },
    {
      "epoch": 0.9107806691449815,
      "grad_norm": 4.298549175262451,
      "learning_rate": 0.00010892193308550187,
      "loss": 0.7648,
      "step": 1470
    },
    {
      "epoch": 0.9169764560099133,
      "grad_norm": 5.923393249511719,
      "learning_rate": 0.00010830235439900868,
      "loss": 0.66,
      "step": 1480
    },
    {
      "epoch": 0.9231722428748451,
      "grad_norm": 3.252465009689331,
      "learning_rate": 0.0001076827757125155,
      "loss": 0.6732,
      "step": 1490
    },
    {
      "epoch": 0.929368029739777,
      "grad_norm": 5.450772285461426,
      "learning_rate": 0.0001070631970260223,
      "loss": 0.7479,
      "step": 1500
    },
    {
      "epoch": 0.929368029739777,
      "eval_accuracy": 0.8066914498141264,
      "eval_loss": 0.7409122586250305,
      "eval_runtime": 106.9474,
      "eval_samples_per_second": 60.366,
      "eval_steps_per_second": 7.546,
      "step": 1500
    },
    {
      "epoch": 0.9355638166047088,
      "grad_norm": 3.8321099281311035,
      "learning_rate": 0.00010644361833952911,
      "loss": 0.6568,
      "step": 1510
    },
    {
      "epoch": 0.9417596034696406,
      "grad_norm": 2.7507283687591553,
      "learning_rate": 0.00010582403965303594,
      "loss": 0.7641,
      "step": 1520
    },
    {
      "epoch": 0.9479553903345725,
      "grad_norm": 4.787430286407471,
      "learning_rate": 0.00010520446096654275,
      "loss": 0.8192,
      "step": 1530
    },
    {
      "epoch": 0.9541511771995044,
      "grad_norm": 5.063214302062988,
      "learning_rate": 0.00010458488228004957,
      "loss": 0.7953,
      "step": 1540
    },
    {
      "epoch": 0.9603469640644362,
      "grad_norm": 3.0914242267608643,
      "learning_rate": 0.00010396530359355638,
      "loss": 0.6188,
      "step": 1550
    },
    {
      "epoch": 0.966542750929368,
      "grad_norm": 2.4344420433044434,
      "learning_rate": 0.0001033457249070632,
      "loss": 0.7707,
      "step": 1560
    },
    {
      "epoch": 0.9727385377942999,
      "grad_norm": 5.563531398773193,
      "learning_rate": 0.00010272614622057001,
      "loss": 0.657,
      "step": 1570
    },
    {
      "epoch": 0.9789343246592317,
      "grad_norm": 2.2125167846679688,
      "learning_rate": 0.00010210656753407684,
      "loss": 0.7362,
      "step": 1580
    },
    {
      "epoch": 0.9851301115241635,
      "grad_norm": 7.253428936004639,
      "learning_rate": 0.00010148698884758365,
      "loss": 0.974,
      "step": 1590
    },
    {
      "epoch": 0.9913258983890955,
      "grad_norm": 5.762598037719727,
      "learning_rate": 0.00010086741016109047,
      "loss": 0.5997,
      "step": 1600
    },
    {
      "epoch": 0.9913258983890955,
      "eval_accuracy": 0.8110285006195787,
      "eval_loss": 0.7325805425643921,
      "eval_runtime": 105.9884,
      "eval_samples_per_second": 60.912,
      "eval_steps_per_second": 7.614,
      "step": 1600
    },
    {
      "epoch": 0.9975216852540273,
      "grad_norm": 7.013967037200928,
      "learning_rate": 0.00010024783147459728,
      "loss": 0.7562,
      "step": 1610
    },
    {
      "epoch": 1.003717472118959,
      "grad_norm": 4.252784252166748,
      "learning_rate": 9.962825278810409e-05,
      "loss": 0.5769,
      "step": 1620
    },
    {
      "epoch": 1.009913258983891,
      "grad_norm": 2.1795663833618164,
      "learning_rate": 9.900867410161091e-05,
      "loss": 0.4631,
      "step": 1630
    },
    {
      "epoch": 1.016109045848823,
      "grad_norm": 2.469095468521118,
      "learning_rate": 9.838909541511772e-05,
      "loss": 0.3613,
      "step": 1640
    },
    {
      "epoch": 1.0223048327137547,
      "grad_norm": 4.4682297706604,
      "learning_rate": 9.776951672862455e-05,
      "loss": 0.4184,
      "step": 1650
    },
    {
      "epoch": 1.0285006195786865,
      "grad_norm": 1.6269049644470215,
      "learning_rate": 9.714993804213134e-05,
      "loss": 0.3376,
      "step": 1660
    },
    {
      "epoch": 1.0346964064436184,
      "grad_norm": 4.2464423179626465,
      "learning_rate": 9.653035935563817e-05,
      "loss": 0.4305,
      "step": 1670
    },
    {
      "epoch": 1.0408921933085502,
      "grad_norm": 3.627943992614746,
      "learning_rate": 9.591078066914498e-05,
      "loss": 0.3696,
      "step": 1680
    },
    {
      "epoch": 1.047087980173482,
      "grad_norm": 2.6817383766174316,
      "learning_rate": 9.52912019826518e-05,
      "loss": 0.3897,
      "step": 1690
    },
    {
      "epoch": 1.0532837670384139,
      "grad_norm": 7.518842697143555,
      "learning_rate": 9.467162329615861e-05,
      "loss": 0.5005,
      "step": 1700
    },
    {
      "epoch": 1.0532837670384139,
      "eval_accuracy": 0.8210966542750929,
      "eval_loss": 0.6769081950187683,
      "eval_runtime": 105.6921,
      "eval_samples_per_second": 61.083,
      "eval_steps_per_second": 7.635,
      "step": 1700
    },
    {
      "epoch": 1.0594795539033457,
      "grad_norm": 3.0254740715026855,
      "learning_rate": 9.405204460966544e-05,
      "loss": 0.3827,
      "step": 1710
    },
    {
      "epoch": 1.0656753407682775,
      "grad_norm": 4.391673564910889,
      "learning_rate": 9.343246592317225e-05,
      "loss": 0.3535,
      "step": 1720
    },
    {
      "epoch": 1.0718711276332094,
      "grad_norm": 3.73157000541687,
      "learning_rate": 9.281288723667906e-05,
      "loss": 0.3238,
      "step": 1730
    },
    {
      "epoch": 1.0780669144981412,
      "grad_norm": 2.160573720932007,
      "learning_rate": 9.219330855018588e-05,
      "loss": 0.383,
      "step": 1740
    },
    {
      "epoch": 1.084262701363073,
      "grad_norm": 4.27864408493042,
      "learning_rate": 9.157372986369269e-05,
      "loss": 0.3361,
      "step": 1750
    },
    {
      "epoch": 1.090458488228005,
      "grad_norm": 3.1258535385131836,
      "learning_rate": 9.095415117719951e-05,
      "loss": 0.2797,
      "step": 1760
    },
    {
      "epoch": 1.096654275092937,
      "grad_norm": 3.8895909786224365,
      "learning_rate": 9.033457249070632e-05,
      "loss": 0.4367,
      "step": 1770
    },
    {
      "epoch": 1.1028500619578687,
      "grad_norm": 2.674630880355835,
      "learning_rate": 8.971499380421315e-05,
      "loss": 0.3925,
      "step": 1780
    },
    {
      "epoch": 1.1090458488228006,
      "grad_norm": 7.267265319824219,
      "learning_rate": 8.909541511771994e-05,
      "loss": 0.4627,
      "step": 1790
    },
    {
      "epoch": 1.1152416356877324,
      "grad_norm": 4.650302886962891,
      "learning_rate": 8.847583643122677e-05,
      "loss": 0.4107,
      "step": 1800
    },
    {
      "epoch": 1.1152416356877324,
      "eval_accuracy": 0.837360594795539,
      "eval_loss": 0.6374781131744385,
      "eval_runtime": 106.1067,
      "eval_samples_per_second": 60.844,
      "eval_steps_per_second": 7.606,
      "step": 1800
    },
    {
      "epoch": 1.1214374225526642,
      "grad_norm": 2.1244664192199707,
      "learning_rate": 8.785625774473358e-05,
      "loss": 0.3463,
      "step": 1810
    },
    {
      "epoch": 1.127633209417596,
      "grad_norm": 1.8372740745544434,
      "learning_rate": 8.72366790582404e-05,
      "loss": 0.3116,
      "step": 1820
    },
    {
      "epoch": 1.1338289962825279,
      "grad_norm": 3.6969428062438965,
      "learning_rate": 8.661710037174722e-05,
      "loss": 0.4069,
      "step": 1830
    },
    {
      "epoch": 1.1400247831474597,
      "grad_norm": 3.857111930847168,
      "learning_rate": 8.599752168525403e-05,
      "loss": 0.4478,
      "step": 1840
    },
    {
      "epoch": 1.1462205700123915,
      "grad_norm": 2.005557060241699,
      "learning_rate": 8.537794299876086e-05,
      "loss": 0.2794,
      "step": 1850
    },
    {
      "epoch": 1.1524163568773234,
      "grad_norm": 5.883118629455566,
      "learning_rate": 8.475836431226765e-05,
      "loss": 0.3732,
      "step": 1860
    },
    {
      "epoch": 1.1586121437422552,
      "grad_norm": 5.240428924560547,
      "learning_rate": 8.413878562577448e-05,
      "loss": 0.3042,
      "step": 1870
    },
    {
      "epoch": 1.164807930607187,
      "grad_norm": 2.857640027999878,
      "learning_rate": 8.351920693928129e-05,
      "loss": 0.3013,
      "step": 1880
    },
    {
      "epoch": 1.1710037174721188,
      "grad_norm": 5.086670398712158,
      "learning_rate": 8.289962825278811e-05,
      "loss": 0.2786,
      "step": 1890
    },
    {
      "epoch": 1.177199504337051,
      "grad_norm": 4.9353413581848145,
      "learning_rate": 8.228004956629492e-05,
      "loss": 0.4596,
      "step": 1900
    },
    {
      "epoch": 1.177199504337051,
      "eval_accuracy": 0.8303903345724907,
      "eval_loss": 0.6301799416542053,
      "eval_runtime": 106.1215,
      "eval_samples_per_second": 60.836,
      "eval_steps_per_second": 7.604,
      "step": 1900
    },
    {
      "epoch": 1.1833952912019827,
      "grad_norm": 3.3738842010498047,
      "learning_rate": 8.166047087980174e-05,
      "loss": 0.2521,
      "step": 1910
    },
    {
      "epoch": 1.1895910780669146,
      "grad_norm": 3.060638427734375,
      "learning_rate": 8.104089219330855e-05,
      "loss": 0.3277,
      "step": 1920
    },
    {
      "epoch": 1.1957868649318464,
      "grad_norm": 3.8044564723968506,
      "learning_rate": 8.042131350681536e-05,
      "loss": 0.3403,
      "step": 1930
    },
    {
      "epoch": 1.2019826517967782,
      "grad_norm": 4.681379795074463,
      "learning_rate": 7.980173482032219e-05,
      "loss": 0.3816,
      "step": 1940
    },
    {
      "epoch": 1.20817843866171,
      "grad_norm": 5.634707450866699,
      "learning_rate": 7.9182156133829e-05,
      "loss": 0.3161,
      "step": 1950
    },
    {
      "epoch": 1.2143742255266419,
      "grad_norm": 2.39424204826355,
      "learning_rate": 7.856257744733582e-05,
      "loss": 0.3638,
      "step": 1960
    },
    {
      "epoch": 1.2205700123915737,
      "grad_norm": 2.7920138835906982,
      "learning_rate": 7.794299876084263e-05,
      "loss": 0.2685,
      "step": 1970
    },
    {
      "epoch": 1.2267657992565055,
      "grad_norm": 6.918692111968994,
      "learning_rate": 7.732342007434946e-05,
      "loss": 0.2768,
      "step": 1980
    },
    {
      "epoch": 1.2329615861214374,
      "grad_norm": 4.783863544464111,
      "learning_rate": 7.670384138785625e-05,
      "loss": 0.3513,
      "step": 1990
    },
    {
      "epoch": 1.2391573729863692,
      "grad_norm": 1.5945993661880493,
      "learning_rate": 7.608426270136308e-05,
      "loss": 0.2544,
      "step": 2000
    },
    {
      "epoch": 1.2391573729863692,
      "eval_accuracy": 0.8399938042131351,
      "eval_loss": 0.5804997086524963,
      "eval_runtime": 106.7403,
      "eval_samples_per_second": 60.483,
      "eval_steps_per_second": 7.56,
      "step": 2000
    },
    {
      "epoch": 1.2453531598513012,
      "grad_norm": 0.8452507853507996,
      "learning_rate": 7.546468401486989e-05,
      "loss": 0.2581,
      "step": 2010
    },
    {
      "epoch": 1.251548946716233,
      "grad_norm": 3.4717066287994385,
      "learning_rate": 7.484510532837671e-05,
      "loss": 0.2793,
      "step": 2020
    },
    {
      "epoch": 1.257744733581165,
      "grad_norm": 3.5942156314849854,
      "learning_rate": 7.422552664188352e-05,
      "loss": 0.2798,
      "step": 2030
    },
    {
      "epoch": 1.2639405204460967,
      "grad_norm": 5.311221599578857,
      "learning_rate": 7.360594795539034e-05,
      "loss": 0.3501,
      "step": 2040
    },
    {
      "epoch": 1.2701363073110286,
      "grad_norm": 3.8793325424194336,
      "learning_rate": 7.298636926889715e-05,
      "loss": 0.3645,
      "step": 2050
    },
    {
      "epoch": 1.2763320941759604,
      "grad_norm": 2.781317949295044,
      "learning_rate": 7.236679058240396e-05,
      "loss": 0.3294,
      "step": 2060
    },
    {
      "epoch": 1.2825278810408922,
      "grad_norm": 1.0684071779251099,
      "learning_rate": 7.174721189591079e-05,
      "loss": 0.2507,
      "step": 2070
    },
    {
      "epoch": 1.288723667905824,
      "grad_norm": 1.6585029363632202,
      "learning_rate": 7.11276332094176e-05,
      "loss": 0.3451,
      "step": 2080
    },
    {
      "epoch": 1.2949194547707559,
      "grad_norm": 1.9376587867736816,
      "learning_rate": 7.050805452292442e-05,
      "loss": 0.3423,
      "step": 2090
    },
    {
      "epoch": 1.3011152416356877,
      "grad_norm": 2.892873525619507,
      "learning_rate": 6.988847583643123e-05,
      "loss": 0.2983,
      "step": 2100
    },
    {
      "epoch": 1.3011152416356877,
      "eval_accuracy": 0.8500619578686494,
      "eval_loss": 0.5480403304100037,
      "eval_runtime": 107.2113,
      "eval_samples_per_second": 60.218,
      "eval_steps_per_second": 7.527,
      "step": 2100
    },
    {
      "epoch": 1.3073110285006195,
      "grad_norm": 0.8831340670585632,
      "learning_rate": 6.926889714993805e-05,
      "loss": 0.3744,
      "step": 2110
    },
    {
      "epoch": 1.3135068153655514,
      "grad_norm": 5.404819011688232,
      "learning_rate": 6.864931846344485e-05,
      "loss": 0.2856,
      "step": 2120
    },
    {
      "epoch": 1.3197026022304832,
      "grad_norm": 5.324650287628174,
      "learning_rate": 6.802973977695167e-05,
      "loss": 0.2055,
      "step": 2130
    },
    {
      "epoch": 1.325898389095415,
      "grad_norm": 1.5021191835403442,
      "learning_rate": 6.741016109045848e-05,
      "loss": 0.3426,
      "step": 2140
    },
    {
      "epoch": 1.3320941759603468,
      "grad_norm": 3.320554256439209,
      "learning_rate": 6.679058240396531e-05,
      "loss": 0.1865,
      "step": 2150
    },
    {
      "epoch": 1.3382899628252787,
      "grad_norm": 6.635782241821289,
      "learning_rate": 6.617100371747212e-05,
      "loss": 0.2956,
      "step": 2160
    },
    {
      "epoch": 1.3444857496902107,
      "grad_norm": 7.695481300354004,
      "learning_rate": 6.555142503097894e-05,
      "loss": 0.2729,
      "step": 2170
    },
    {
      "epoch": 1.3506815365551426,
      "grad_norm": 2.3640730381011963,
      "learning_rate": 6.493184634448575e-05,
      "loss": 0.2946,
      "step": 2180
    },
    {
      "epoch": 1.3568773234200744,
      "grad_norm": 6.867854595184326,
      "learning_rate": 6.431226765799256e-05,
      "loss": 0.269,
      "step": 2190
    },
    {
      "epoch": 1.3630731102850062,
      "grad_norm": 5.83229923248291,
      "learning_rate": 6.369268897149939e-05,
      "loss": 0.3214,
      "step": 2200
    },
    {
      "epoch": 1.3630731102850062,
      "eval_accuracy": 0.8683395291201983,
      "eval_loss": 0.5052544474601746,
      "eval_runtime": 106.9004,
      "eval_samples_per_second": 60.393,
      "eval_steps_per_second": 7.549,
      "step": 2200
    },
    {
      "epoch": 1.369268897149938,
      "grad_norm": 4.513510227203369,
      "learning_rate": 6.30731102850062e-05,
      "loss": 0.2435,
      "step": 2210
    },
    {
      "epoch": 1.3754646840148699,
      "grad_norm": 0.9935147762298584,
      "learning_rate": 6.245353159851302e-05,
      "loss": 0.237,
      "step": 2220
    },
    {
      "epoch": 1.3816604708798017,
      "grad_norm": 5.048427104949951,
      "learning_rate": 6.183395291201983e-05,
      "loss": 0.1811,
      "step": 2230
    },
    {
      "epoch": 1.3878562577447335,
      "grad_norm": 4.882187366485596,
      "learning_rate": 6.121437422552665e-05,
      "loss": 0.2591,
      "step": 2240
    },
    {
      "epoch": 1.3940520446096654,
      "grad_norm": 3.1441776752471924,
      "learning_rate": 6.0594795539033456e-05,
      "loss": 0.2503,
      "step": 2250
    },
    {
      "epoch": 1.4002478314745972,
      "grad_norm": 5.158385753631592,
      "learning_rate": 5.997521685254027e-05,
      "loss": 0.3771,
      "step": 2260
    },
    {
      "epoch": 1.4064436183395292,
      "grad_norm": 4.472280979156494,
      "learning_rate": 5.935563816604709e-05,
      "loss": 0.2673,
      "step": 2270
    },
    {
      "epoch": 1.412639405204461,
      "grad_norm": 3.1100497245788574,
      "learning_rate": 5.8736059479553906e-05,
      "loss": 0.3471,
      "step": 2280
    },
    {
      "epoch": 1.418835192069393,
      "grad_norm": 1.5550055503845215,
      "learning_rate": 5.811648079306072e-05,
      "loss": 0.2452,
      "step": 2290
    },
    {
      "epoch": 1.4250309789343247,
      "grad_norm": 1.051927924156189,
      "learning_rate": 5.749690210656754e-05,
      "loss": 0.2384,
      "step": 2300
    },
    {
      "epoch": 1.4250309789343247,
      "eval_accuracy": 0.8712825278810409,
      "eval_loss": 0.4928523004055023,
      "eval_runtime": 106.1123,
      "eval_samples_per_second": 60.841,
      "eval_steps_per_second": 7.605,
      "step": 2300
    },
    {
      "epoch": 1.4312267657992566,
      "grad_norm": 3.616842031478882,
      "learning_rate": 5.687732342007436e-05,
      "loss": 0.251,
      "step": 2310
    },
    {
      "epoch": 1.4374225526641884,
      "grad_norm": 3.94722580909729,
      "learning_rate": 5.625774473358117e-05,
      "loss": 0.1996,
      "step": 2320
    },
    {
      "epoch": 1.4436183395291202,
      "grad_norm": 6.4395222663879395,
      "learning_rate": 5.5638166047087984e-05,
      "loss": 0.3224,
      "step": 2330
    },
    {
      "epoch": 1.449814126394052,
      "grad_norm": 2.801499366760254,
      "learning_rate": 5.50185873605948e-05,
      "loss": 0.2018,
      "step": 2340
    },
    {
      "epoch": 1.4560099132589839,
      "grad_norm": 4.119659900665283,
      "learning_rate": 5.439900867410162e-05,
      "loss": 0.2661,
      "step": 2350
    },
    {
      "epoch": 1.4622057001239157,
      "grad_norm": 8.405607223510742,
      "learning_rate": 5.3779429987608434e-05,
      "loss": 0.3412,
      "step": 2360
    },
    {
      "epoch": 1.4684014869888475,
      "grad_norm": 3.3943393230438232,
      "learning_rate": 5.315985130111525e-05,
      "loss": 0.303,
      "step": 2370
    },
    {
      "epoch": 1.4745972738537794,
      "grad_norm": 3.990785837173462,
      "learning_rate": 5.2540272614622054e-05,
      "loss": 0.1961,
      "step": 2380
    },
    {
      "epoch": 1.4807930607187112,
      "grad_norm": 8.141942977905273,
      "learning_rate": 5.192069392812887e-05,
      "loss": 0.2683,
      "step": 2390
    },
    {
      "epoch": 1.486988847583643,
      "grad_norm": 2.5247440338134766,
      "learning_rate": 5.130111524163569e-05,
      "loss": 0.2397,
      "step": 2400
    },
    {
      "epoch": 1.486988847583643,
      "eval_accuracy": 0.8742255266418835,
      "eval_loss": 0.4664279520511627,
      "eval_runtime": 106.5781,
      "eval_samples_per_second": 60.575,
      "eval_steps_per_second": 7.572,
      "step": 2400
    },
    {
      "epoch": 1.4931846344485749,
      "grad_norm": 2.132350206375122,
      "learning_rate": 5.0681536555142505e-05,
      "loss": 0.3258,
      "step": 2410
    },
    {
      "epoch": 1.4993804213135067,
      "grad_norm": 4.359376430511475,
      "learning_rate": 5.006195786864932e-05,
      "loss": 0.2781,
      "step": 2420
    },
    {
      "epoch": 1.5055762081784385,
      "grad_norm": 7.171940326690674,
      "learning_rate": 4.944237918215613e-05,
      "loss": 0.2828,
      "step": 2430
    },
    {
      "epoch": 1.5117719950433703,
      "grad_norm": 1.664962649345398,
      "learning_rate": 4.882280049566295e-05,
      "loss": 0.206,
      "step": 2440
    },
    {
      "epoch": 1.5179677819083024,
      "grad_norm": 7.0570268630981445,
      "learning_rate": 4.820322180916977e-05,
      "loss": 0.2365,
      "step": 2450
    },
    {
      "epoch": 1.5241635687732342,
      "grad_norm": 4.6403279304504395,
      "learning_rate": 4.758364312267658e-05,
      "loss": 0.3308,
      "step": 2460
    },
    {
      "epoch": 1.530359355638166,
      "grad_norm": 0.445726215839386,
      "learning_rate": 4.69640644361834e-05,
      "loss": 0.171,
      "step": 2470
    },
    {
      "epoch": 1.5365551425030979,
      "grad_norm": 6.475937366485596,
      "learning_rate": 4.6344485749690216e-05,
      "loss": 0.2941,
      "step": 2480
    },
    {
      "epoch": 1.5427509293680297,
      "grad_norm": 1.953753113746643,
      "learning_rate": 4.5724907063197026e-05,
      "loss": 0.1691,
      "step": 2490
    },
    {
      "epoch": 1.5489467162329615,
      "grad_norm": 4.187342166900635,
      "learning_rate": 4.510532837670384e-05,
      "loss": 0.3448,
      "step": 2500
    },
    {
      "epoch": 1.5489467162329615,
      "eval_accuracy": 0.8754646840148699,
      "eval_loss": 0.46897682547569275,
      "eval_runtime": 107.2096,
      "eval_samples_per_second": 60.218,
      "eval_steps_per_second": 7.527,
      "step": 2500
    },
    {
      "epoch": 1.5551425030978936,
      "grad_norm": 4.653651714324951,
      "learning_rate": 4.448574969021066e-05,
      "loss": 0.2035,
      "step": 2510
    },
    {
      "epoch": 1.5613382899628254,
      "grad_norm": 1.159033179283142,
      "learning_rate": 4.3866171003717476e-05,
      "loss": 0.2354,
      "step": 2520
    },
    {
      "epoch": 1.5675340768277573,
      "grad_norm": 0.841773271560669,
      "learning_rate": 4.3246592317224286e-05,
      "loss": 0.1738,
      "step": 2530
    },
    {
      "epoch": 1.573729863692689,
      "grad_norm": 6.38914155960083,
      "learning_rate": 4.26270136307311e-05,
      "loss": 0.2743,
      "step": 2540
    },
    {
      "epoch": 1.579925650557621,
      "grad_norm": 0.2980528473854065,
      "learning_rate": 4.200743494423792e-05,
      "loss": 0.2383,
      "step": 2550
    },
    {
      "epoch": 1.5861214374225527,
      "grad_norm": 3.9257161617279053,
      "learning_rate": 4.1387856257744737e-05,
      "loss": 0.1728,
      "step": 2560
    },
    {
      "epoch": 1.5923172242874846,
      "grad_norm": 4.1586785316467285,
      "learning_rate": 4.0768277571251553e-05,
      "loss": 0.2226,
      "step": 2570
    },
    {
      "epoch": 1.5985130111524164,
      "grad_norm": 1.591169834136963,
      "learning_rate": 4.014869888475837e-05,
      "loss": 0.1998,
      "step": 2580
    },
    {
      "epoch": 1.6047087980173482,
      "grad_norm": 1.5540215969085693,
      "learning_rate": 3.952912019826518e-05,
      "loss": 0.101,
      "step": 2590
    },
    {
      "epoch": 1.61090458488228,
      "grad_norm": 1.4260759353637695,
      "learning_rate": 3.8909541511772e-05,
      "loss": 0.3129,
      "step": 2600
    },
    {
      "epoch": 1.61090458488228,
      "eval_accuracy": 0.8842936802973977,
      "eval_loss": 0.4350809156894684,
      "eval_runtime": 106.4924,
      "eval_samples_per_second": 60.624,
      "eval_steps_per_second": 7.578,
      "step": 2600
    },
    {
      "epoch": 1.6171003717472119,
      "grad_norm": 4.335544586181641,
      "learning_rate": 3.8289962825278814e-05,
      "loss": 0.2159,
      "step": 2610
    },
    {
      "epoch": 1.6232961586121437,
      "grad_norm": 1.1200919151306152,
      "learning_rate": 3.7670384138785624e-05,
      "loss": 0.2679,
      "step": 2620
    },
    {
      "epoch": 1.6294919454770755,
      "grad_norm": 2.773334264755249,
      "learning_rate": 3.705080545229244e-05,
      "loss": 0.2927,
      "step": 2630
    },
    {
      "epoch": 1.6356877323420074,
      "grad_norm": 2.635826349258423,
      "learning_rate": 3.643122676579926e-05,
      "loss": 0.2937,
      "step": 2640
    },
    {
      "epoch": 1.6418835192069392,
      "grad_norm": 2.025951385498047,
      "learning_rate": 3.5811648079306074e-05,
      "loss": 0.2392,
      "step": 2650
    },
    {
      "epoch": 1.648079306071871,
      "grad_norm": 9.39108943939209,
      "learning_rate": 3.5192069392812884e-05,
      "loss": 0.3255,
      "step": 2660
    },
    {
      "epoch": 1.6542750929368029,
      "grad_norm": 1.8526005744934082,
      "learning_rate": 3.45724907063197e-05,
      "loss": 0.1944,
      "step": 2670
    },
    {
      "epoch": 1.6604708798017347,
      "grad_norm": 0.7823792695999146,
      "learning_rate": 3.3952912019826525e-05,
      "loss": 0.149,
      "step": 2680
    },
    {
      "epoch": 1.6666666666666665,
      "grad_norm": 7.365529537200928,
      "learning_rate": 3.3333333333333335e-05,
      "loss": 0.312,
      "step": 2690
    },
    {
      "epoch": 1.6728624535315983,
      "grad_norm": 3.5820822715759277,
      "learning_rate": 3.271375464684015e-05,
      "loss": 0.1027,
      "step": 2700
    },
    {
      "epoch": 1.6728624535315983,
      "eval_accuracy": 0.8846034696406444,
      "eval_loss": 0.4310809373855591,
      "eval_runtime": 106.7243,
      "eval_samples_per_second": 60.492,
      "eval_steps_per_second": 7.562,
      "step": 2700
    },
    {
      "epoch": 1.6790582403965304,
      "grad_norm": 0.8703699111938477,
      "learning_rate": 3.209417596034697e-05,
      "loss": 0.1744,
      "step": 2710
    },
    {
      "epoch": 1.6852540272614622,
      "grad_norm": 3.368072986602783,
      "learning_rate": 3.147459727385378e-05,
      "loss": 0.2112,
      "step": 2720
    },
    {
      "epoch": 1.691449814126394,
      "grad_norm": 2.1907970905303955,
      "learning_rate": 3.0855018587360595e-05,
      "loss": 0.227,
      "step": 2730
    },
    {
      "epoch": 1.6976456009913259,
      "grad_norm": 3.725156784057617,
      "learning_rate": 3.0235439900867412e-05,
      "loss": 0.184,
      "step": 2740
    },
    {
      "epoch": 1.7038413878562577,
      "grad_norm": 0.17684808373451233,
      "learning_rate": 2.9615861214374226e-05,
      "loss": 0.1559,
      "step": 2750
    },
    {
      "epoch": 1.7100371747211895,
      "grad_norm": 5.654155731201172,
      "learning_rate": 2.8996282527881043e-05,
      "loss": 0.2551,
      "step": 2760
    },
    {
      "epoch": 1.7162329615861216,
      "grad_norm": 1.2272543907165527,
      "learning_rate": 2.837670384138786e-05,
      "loss": 0.2352,
      "step": 2770
    },
    {
      "epoch": 1.7224287484510534,
      "grad_norm": 0.7282238006591797,
      "learning_rate": 2.7757125154894676e-05,
      "loss": 0.2647,
      "step": 2780
    },
    {
      "epoch": 1.7286245353159853,
      "grad_norm": 2.949235677719116,
      "learning_rate": 2.7137546468401486e-05,
      "loss": 0.1656,
      "step": 2790
    },
    {
      "epoch": 1.734820322180917,
      "grad_norm": 5.691123962402344,
      "learning_rate": 2.6517967781908303e-05,
      "loss": 0.2086,
      "step": 2800
    },
    {
      "epoch": 1.734820322180917,
      "eval_accuracy": 0.8897149938042132,
      "eval_loss": 0.4087870121002197,
      "eval_runtime": 107.499,
      "eval_samples_per_second": 60.056,
      "eval_steps_per_second": 7.507,
      "step": 2800
    },
    {
      "epoch": 1.741016109045849,
      "grad_norm": 0.18745893239974976,
      "learning_rate": 2.589838909541512e-05,
      "loss": 0.1878,
      "step": 2810
    },
    {
      "epoch": 1.7472118959107807,
      "grad_norm": 6.888800621032715,
      "learning_rate": 2.5278810408921933e-05,
      "loss": 0.1727,
      "step": 2820
    },
    {
      "epoch": 1.7534076827757126,
      "grad_norm": 1.913245439529419,
      "learning_rate": 2.465923172242875e-05,
      "loss": 0.1732,
      "step": 2830
    },
    {
      "epoch": 1.7596034696406444,
      "grad_norm": 2.651405096054077,
      "learning_rate": 2.4039653035935564e-05,
      "loss": 0.1928,
      "step": 2840
    },
    {
      "epoch": 1.7657992565055762,
      "grad_norm": 1.3924200534820557,
      "learning_rate": 2.342007434944238e-05,
      "loss": 0.1397,
      "step": 2850
    },
    {
      "epoch": 1.771995043370508,
      "grad_norm": 7.028463840484619,
      "learning_rate": 2.2800495662949194e-05,
      "loss": 0.2461,
      "step": 2860
    },
    {
      "epoch": 1.77819083023544,
      "grad_norm": 1.74459707736969,
      "learning_rate": 2.218091697645601e-05,
      "loss": 0.137,
      "step": 2870
    },
    {
      "epoch": 1.7843866171003717,
      "grad_norm": 0.35962000489234924,
      "learning_rate": 2.1561338289962827e-05,
      "loss": 0.1767,
      "step": 2880
    },
    {
      "epoch": 1.7905824039653035,
      "grad_norm": 1.6832956075668335,
      "learning_rate": 2.094175960346964e-05,
      "loss": 0.2574,
      "step": 2890
    },
    {
      "epoch": 1.7967781908302354,
      "grad_norm": 5.7223029136657715,
      "learning_rate": 2.0322180916976458e-05,
      "loss": 0.1683,
      "step": 2900
    },
    {
      "epoch": 1.7967781908302354,
      "eval_accuracy": 0.8918835192069393,
      "eval_loss": 0.41334882378578186,
      "eval_runtime": 106.6472,
      "eval_samples_per_second": 60.536,
      "eval_steps_per_second": 7.567,
      "step": 2900
    },
    {
      "epoch": 1.8029739776951672,
      "grad_norm": 6.862188339233398,
      "learning_rate": 1.970260223048327e-05,
      "loss": 0.1923,
      "step": 2910
    },
    {
      "epoch": 1.809169764560099,
      "grad_norm": 1.456938624382019,
      "learning_rate": 1.9083023543990088e-05,
      "loss": 0.167,
      "step": 2920
    },
    {
      "epoch": 1.8153655514250309,
      "grad_norm": 1.2363471984863281,
      "learning_rate": 1.8463444857496905e-05,
      "loss": 0.1129,
      "step": 2930
    },
    {
      "epoch": 1.8215613382899627,
      "grad_norm": 2.139641761779785,
      "learning_rate": 1.7843866171003718e-05,
      "loss": 0.2046,
      "step": 2940
    },
    {
      "epoch": 1.8277571251548945,
      "grad_norm": 7.9193620681762695,
      "learning_rate": 1.7224287484510535e-05,
      "loss": 0.1843,
      "step": 2950
    },
    {
      "epoch": 1.8339529120198264,
      "grad_norm": 0.4252839684486389,
      "learning_rate": 1.660470879801735e-05,
      "loss": 0.1719,
      "step": 2960
    },
    {
      "epoch": 1.8401486988847584,
      "grad_norm": 2.5650665760040283,
      "learning_rate": 1.5985130111524162e-05,
      "loss": 0.2294,
      "step": 2970
    },
    {
      "epoch": 1.8463444857496902,
      "grad_norm": 4.639560222625732,
      "learning_rate": 1.536555142503098e-05,
      "loss": 0.163,
      "step": 2980
    },
    {
      "epoch": 1.852540272614622,
      "grad_norm": 1.6396369934082031,
      "learning_rate": 1.4745972738537794e-05,
      "loss": 0.0855,
      "step": 2990
    },
    {
      "epoch": 1.858736059479554,
      "grad_norm": 3.6492340564727783,
      "learning_rate": 1.412639405204461e-05,
      "loss": 0.2767,
      "step": 3000
    },
    {
      "epoch": 1.858736059479554,
      "eval_accuracy": 0.8963754646840149,
      "eval_loss": 0.38507798314094543,
      "eval_runtime": 106.3075,
      "eval_samples_per_second": 60.73,
      "eval_steps_per_second": 7.591,
      "step": 3000
    },
    {
      "epoch": 1.8649318463444857,
      "grad_norm": 2.8184783458709717,
      "learning_rate": 1.3506815365551426e-05,
      "loss": 0.1552,
      "step": 3010
    },
    {
      "epoch": 1.8711276332094176,
      "grad_norm": 0.4257136881351471,
      "learning_rate": 1.288723667905824e-05,
      "loss": 0.169,
      "step": 3020
    },
    {
      "epoch": 1.8773234200743496,
      "grad_norm": 6.326307773590088,
      "learning_rate": 1.2267657992565058e-05,
      "loss": 0.1842,
      "step": 3030
    },
    {
      "epoch": 1.8835192069392814,
      "grad_norm": 0.6836357712745667,
      "learning_rate": 1.1648079306071871e-05,
      "loss": 0.1587,
      "step": 3040
    },
    {
      "epoch": 1.8897149938042133,
      "grad_norm": 6.755892753601074,
      "learning_rate": 1.1028500619578686e-05,
      "loss": 0.1954,
      "step": 3050
    },
    {
      "epoch": 1.895910780669145,
      "grad_norm": 2.74873948097229,
      "learning_rate": 1.0408921933085503e-05,
      "loss": 0.1719,
      "step": 3060
    },
    {
      "epoch": 1.902106567534077,
      "grad_norm": 4.230051517486572,
      "learning_rate": 9.789343246592318e-06,
      "loss": 0.1562,
      "step": 3070
    },
    {
      "epoch": 1.9083023543990087,
      "grad_norm": 5.3604512214660645,
      "learning_rate": 9.169764560099132e-06,
      "loss": 0.2365,
      "step": 3080
    },
    {
      "epoch": 1.9144981412639406,
      "grad_norm": 0.3976893723011017,
      "learning_rate": 8.550185873605949e-06,
      "loss": 0.1322,
      "step": 3090
    },
    {
      "epoch": 1.9206939281288724,
      "grad_norm": 3.6880292892456055,
      "learning_rate": 7.930607187112764e-06,
      "loss": 0.1582,
      "step": 3100
    },
    {
      "epoch": 1.9206939281288724,
      "eval_accuracy": 0.9017967781908303,
      "eval_loss": 0.3703024089336395,
      "eval_runtime": 107.4091,
      "eval_samples_per_second": 60.107,
      "eval_steps_per_second": 7.513,
      "step": 3100
    },
    {
      "epoch": 1.9268897149938042,
      "grad_norm": 9.823491096496582,
      "learning_rate": 7.31102850061958e-06,
      "loss": 0.2671,
      "step": 3110
    },
    {
      "epoch": 1.933085501858736,
      "grad_norm": 5.724573135375977,
      "learning_rate": 6.691449814126394e-06,
      "loss": 0.2088,
      "step": 3120
    },
    {
      "epoch": 1.939281288723668,
      "grad_norm": 6.375148773193359,
      "learning_rate": 6.071871127633209e-06,
      "loss": 0.1613,
      "step": 3130
    },
    {
      "epoch": 1.9454770755885997,
      "grad_norm": 3.657437324523926,
      "learning_rate": 5.452292441140025e-06,
      "loss": 0.1824,
      "step": 3140
    },
    {
      "epoch": 1.9516728624535316,
      "grad_norm": 0.8294070959091187,
      "learning_rate": 4.832713754646841e-06,
      "loss": 0.1733,
      "step": 3150
    },
    {
      "epoch": 1.9578686493184634,
      "grad_norm": 2.991377592086792,
      "learning_rate": 4.213135068153655e-06,
      "loss": 0.2467,
      "step": 3160
    },
    {
      "epoch": 1.9640644361833952,
      "grad_norm": 3.435967445373535,
      "learning_rate": 3.5935563816604712e-06,
      "loss": 0.1882,
      "step": 3170
    },
    {
      "epoch": 1.970260223048327,
      "grad_norm": 4.247952938079834,
      "learning_rate": 2.9739776951672864e-06,
      "loss": 0.1783,
      "step": 3180
    },
    {
      "epoch": 1.9764560099132589,
      "grad_norm": 0.5351110100746155,
      "learning_rate": 2.3543990086741015e-06,
      "loss": 0.1373,
      "step": 3190
    },
    {
      "epoch": 1.9826517967781907,
      "grad_norm": 2.167306661605835,
      "learning_rate": 1.7348203221809173e-06,
      "loss": 0.1421,
      "step": 3200
    },
    {
      "epoch": 1.9826517967781907,
      "eval_accuracy": 0.90272614622057,
      "eval_loss": 0.36434125900268555,
      "eval_runtime": 107.4144,
      "eval_samples_per_second": 60.104,
      "eval_steps_per_second": 7.513,
      "step": 3200
    },
    {
      "epoch": 1.9888475836431225,
      "grad_norm": 1.093064785003662,
      "learning_rate": 1.1152416356877324e-06,
      "loss": 0.1911,
      "step": 3210
    },
    {
      "epoch": 1.9950433705080544,
      "grad_norm": 3.7282161712646484,
      "learning_rate": 4.956629491945477e-07,
      "loss": 0.2041,
      "step": 3220
    },
    {
      "epoch": 2.0,
      "step": 3228,
      "total_flos": 4.004423768814723e+18,
      "train_loss": 0.9726454161726114,
      "train_runtime": 5145.5507,
      "train_samples_per_second": 10.036,
      "train_steps_per_second": 0.627
    }
  ],
  "logging_steps": 10,
  "max_steps": 3228,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 2,
  "save_steps": 100,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 4.004423768814723e+18,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": null
}