{
  "best_metric": 0.17754687368869781,
  "best_model_checkpoint": "results/checkpoint-25000",
  "epoch": 9.998720081914758,
  "eval_steps": 500,
  "global_step": 26040,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.038397542557276336,
      "grad_norm": 1.0079567432403564,
      "learning_rate": 9.999643338380885e-06,
      "loss": 5.5723,
      "step": 100
    },
    {
      "epoch": 0.07679508511455267,
      "grad_norm": 0.6461474299430847,
      "learning_rate": 9.998558958654982e-06,
      "loss": 2.2782,
      "step": 200
    },
    {
      "epoch": 0.115192627671829,
      "grad_norm": 0.4909125566482544,
      "learning_rate": 9.996746982275233e-06,
      "loss": 1.8047,
      "step": 300
    },
    {
      "epoch": 0.15359017022910534,
      "grad_norm": 0.47547289729118347,
      "learning_rate": 9.994207672995245e-06,
      "loss": 1.5821,
      "step": 400
    },
    {
      "epoch": 0.19198771278638166,
      "grad_norm": 0.41358837485313416,
      "learning_rate": 9.99094140044013e-06,
      "loss": 1.4754,
      "step": 500
    },
    {
      "epoch": 0.19198771278638166,
      "eval_valid_loss": 1.4288749694824219,
      "eval_valid_runtime": 4.7117,
      "eval_valid_samples_per_second": 212.238,
      "eval_valid_steps_per_second": 6.792,
      "step": 500
    },
    {
      "epoch": 0.19198771278638166,
      "eval_valid_target_loss": 1.4590624570846558,
      "eval_valid_target_runtime": 4.684,
      "eval_valid_target_samples_per_second": 213.493,
      "eval_valid_target_steps_per_second": 6.832,
      "step": 500
    },
    {
      "epoch": 0.230385255343658,
      "grad_norm": 0.43229448795318604,
      "learning_rate": 9.986948640052719e-06,
      "loss": 1.4087,
      "step": 600
    },
    {
      "epoch": 0.26878279790093434,
      "grad_norm": 0.528977632522583,
      "learning_rate": 9.982229973024328e-06,
      "loss": 1.3245,
      "step": 700
    },
    {
      "epoch": 0.3071803404582107,
      "grad_norm": 0.5489594340324402,
      "learning_rate": 9.976786086210186e-06,
      "loss": 1.0455,
      "step": 800
    },
    {
      "epoch": 0.34557788301548703,
      "grad_norm": 0.5119125843048096,
      "learning_rate": 9.970617772029439e-06,
      "loss": 0.7605,
      "step": 900
    },
    {
      "epoch": 0.3839754255727633,
      "grad_norm": 0.5092576146125793,
      "learning_rate": 9.963725928349814e-06,
      "loss": 0.6005,
      "step": 1000
    },
    {
      "epoch": 0.3839754255727633,
      "eval_valid_loss": 0.49165624380111694,
      "eval_valid_runtime": 4.674,
      "eval_valid_samples_per_second": 213.951,
      "eval_valid_steps_per_second": 6.846,
      "step": 1000
    },
    {
      "epoch": 0.3839754255727633,
      "eval_valid_target_loss": 0.5142187476158142,
      "eval_valid_target_runtime": 4.6758,
      "eval_valid_target_samples_per_second": 213.869,
      "eval_valid_target_steps_per_second": 6.844,
      "step": 1000
    },
    {
      "epoch": 0.42237296813003966,
      "grad_norm": 0.43798330426216125,
      "learning_rate": 9.956111558356915e-06,
      "loss": 0.4887,
      "step": 1100
    },
    {
      "epoch": 0.460770510687316,
      "grad_norm": 0.3737218379974365,
      "learning_rate": 9.947775770408207e-06,
      "loss": 0.4307,
      "step": 1200
    },
    {
      "epoch": 0.49916805324459235,
      "grad_norm": 0.4303857386112213,
      "learning_rate": 9.938719777871674e-06,
      "loss": 0.4027,
      "step": 1300
    },
    {
      "epoch": 0.5375655958018687,
      "grad_norm": 0.3709202706813812,
      "learning_rate": 9.92894489894921e-06,
      "loss": 0.3799,
      "step": 1400
    },
    {
      "epoch": 0.575963138359145,
      "grad_norm": 0.3918135464191437,
      "learning_rate": 9.918452556484728e-06,
      "loss": 0.3633,
      "step": 1500
    },
    {
      "epoch": 0.575963138359145,
      "eval_valid_loss": 0.33228906989097595,
      "eval_valid_runtime": 4.7244,
      "eval_valid_samples_per_second": 211.669,
      "eval_valid_steps_per_second": 6.773,
      "step": 1500
    },
    {
      "epoch": 0.575963138359145,
      "eval_valid_target_loss": 0.3428671956062317,
      "eval_valid_target_runtime": 4.6595,
      "eval_valid_target_samples_per_second": 214.617,
      "eval_valid_target_steps_per_second": 6.868,
      "step": 1500
    },
    {
      "epoch": 0.6143606809164214,
      "grad_norm": 0.3580816686153412,
      "learning_rate": 9.907244277757053e-06,
      "loss": 0.3565,
      "step": 1600
    },
    {
      "epoch": 0.6527582234736977,
      "grad_norm": 0.34893009066581726,
      "learning_rate": 9.895321694257617e-06,
      "loss": 0.3443,
      "step": 1700
    },
    {
      "epoch": 0.6911557660309741,
      "grad_norm": 0.3050221800804138,
      "learning_rate": 9.882686541452967e-06,
      "loss": 0.3339,
      "step": 1800
    },
    {
      "epoch": 0.7295533085882504,
      "grad_norm": 0.3123306632041931,
      "learning_rate": 9.869340658532151e-06,
      "loss": 0.3278,
      "step": 1900
    },
    {
      "epoch": 0.7679508511455266,
      "grad_norm": 0.31590646505355835,
      "learning_rate": 9.85528598813901e-06,
      "loss": 0.32,
      "step": 2000
    },
    {
      "epoch": 0.7679508511455266,
      "eval_valid_loss": 0.29783594608306885,
      "eval_valid_runtime": 4.6776,
      "eval_valid_samples_per_second": 213.785,
      "eval_valid_steps_per_second": 6.841,
      "step": 2000
    },
    {
      "epoch": 0.7679508511455266,
      "eval_valid_target_loss": 0.3129218816757202,
      "eval_valid_target_runtime": 4.6598,
      "eval_valid_target_samples_per_second": 214.601,
      "eval_valid_target_steps_per_second": 6.867,
      "step": 2000
    },
    {
      "epoch": 0.806348393702803,
      "grad_norm": 0.2684693932533264,
      "learning_rate": 9.840524576089392e-06,
      "loss": 0.3194,
      "step": 2100
    },
    {
      "epoch": 0.8447459362600793,
      "grad_norm": 0.30888476967811584,
      "learning_rate": 9.82505857107337e-06,
      "loss": 0.3108,
      "step": 2200
    },
    {
      "epoch": 0.8831434788173557,
      "grad_norm": 0.2777215242385864,
      "learning_rate": 9.808890224342476e-06,
      "loss": 0.3105,
      "step": 2300
    },
    {
      "epoch": 0.921541021374632,
      "grad_norm": 0.30601397156715393,
      "learning_rate": 9.792021889381995e-06,
      "loss": 0.3055,
      "step": 2400
    },
    {
      "epoch": 0.9599385639319084,
      "grad_norm": 0.2972748875617981,
      "learning_rate": 9.774456021568404e-06,
      "loss": 0.3008,
      "step": 2500
    },
    {
      "epoch": 0.9599385639319084,
      "eval_valid_loss": 0.27842968702316284,
      "eval_valid_runtime": 4.68,
      "eval_valid_samples_per_second": 213.675,
      "eval_valid_steps_per_second": 6.838,
      "step": 2500
    },
    {
      "epoch": 0.9599385639319084,
      "eval_valid_target_loss": 0.2943359315395355,
      "eval_valid_target_runtime": 4.6764,
      "eval_valid_target_samples_per_second": 213.838,
      "eval_valid_target_steps_per_second": 6.843,
      "step": 2500
    },
    {
      "epoch": 0.9983361064891847,
      "grad_norm": 0.2891447842121124,
      "learning_rate": 9.756195177811953e-06,
      "loss": 0.2969,
      "step": 2600
    },
    {
      "epoch": 1.036733649046461,
      "grad_norm": 0.30049994587898254,
      "learning_rate": 9.737242016184486e-06,
      "loss": 0.2913,
      "step": 2700
    },
    {
      "epoch": 1.0751311916037374,
      "grad_norm": 0.25728341937065125,
      "learning_rate": 9.717599295532518e-06,
      "loss": 0.2911,
      "step": 2800
    },
    {
      "epoch": 1.1135287341610136,
      "grad_norm": 0.31619054079055786,
      "learning_rate": 9.697269875075667e-06,
      "loss": 0.2879,
      "step": 2900
    },
    {
      "epoch": 1.15192627671829,
      "grad_norm": 0.3005208671092987,
      "learning_rate": 9.676256713990448e-06,
      "loss": 0.2839,
      "step": 3000
    },
    {
      "epoch": 1.15192627671829,
      "eval_valid_loss": 0.2648593783378601,
      "eval_valid_runtime": 4.6888,
      "eval_valid_samples_per_second": 213.274,
      "eval_valid_steps_per_second": 6.825,
      "step": 3000
    },
    {
      "epoch": 1.15192627671829,
      "eval_valid_target_loss": 0.2817968726158142,
      "eval_valid_target_runtime": 4.6695,
      "eval_valid_target_samples_per_second": 214.155,
      "eval_valid_target_steps_per_second": 6.853,
      "step": 3000
    },
    {
      "epoch": 1.1903238192755663,
      "grad_norm": 0.24846772849559784,
      "learning_rate": 9.654562870979545e-06,
      "loss": 0.2803,
      "step": 3100
    },
    {
      "epoch": 1.2287213618328428,
      "grad_norm": 0.23501113057136536,
      "learning_rate": 9.632191503826574e-06,
      "loss": 0.278,
      "step": 3200
    },
    {
      "epoch": 1.267118904390119,
      "grad_norm": 0.27793240547180176,
      "learning_rate": 9.609145868936434e-06,
      "loss": 0.2776,
      "step": 3300
    },
    {
      "epoch": 1.3055164469473954,
      "grad_norm": 0.2599338889122009,
      "learning_rate": 9.5854293208613e-06,
      "loss": 0.275,
      "step": 3400
    },
    {
      "epoch": 1.3439139895046717,
      "grad_norm": 0.2431340515613556,
      "learning_rate": 9.561045311812335e-06,
      "loss": 0.2722,
      "step": 3500
    },
    {
      "epoch": 1.3439139895046717,
      "eval_valid_loss": 0.2545468807220459,
      "eval_valid_runtime": 4.7131,
      "eval_valid_samples_per_second": 212.177,
      "eval_valid_steps_per_second": 6.79,
      "step": 3500
    },
    {
      "epoch": 1.3439139895046717,
      "eval_valid_target_loss": 0.27326563000679016,
      "eval_valid_target_runtime": 4.6635,
      "eval_valid_target_samples_per_second": 214.433,
      "eval_valid_target_steps_per_second": 6.862,
      "step": 3500
    },
    {
      "epoch": 1.382311532061948,
      "grad_norm": 0.28658226132392883,
      "learning_rate": 9.535997391157174e-06,
      "loss": 0.2693,
      "step": 3600
    },
    {
      "epoch": 1.4207090746192244,
      "grad_norm": 0.26118528842926025,
      "learning_rate": 9.510289204903273e-06,
      "loss": 0.2667,
      "step": 3700
    },
    {
      "epoch": 1.4591066171765008,
      "grad_norm": 0.2761940062046051,
      "learning_rate": 9.483924495167204e-06,
      "loss": 0.2654,
      "step": 3800
    },
    {
      "epoch": 1.497504159733777,
      "grad_norm": 0.2712952792644501,
      "learning_rate": 9.456907099629933e-06,
      "loss": 0.2642,
      "step": 3900
    },
    {
      "epoch": 1.5359017022910533,
      "grad_norm": 0.23100949823856354,
      "learning_rate": 9.429240950978212e-06,
      "loss": 0.2622,
      "step": 4000
    },
    {
      "epoch": 1.5359017022910533,
      "eval_valid_loss": 0.24485936760902405,
      "eval_valid_runtime": 4.6751,
      "eval_valid_samples_per_second": 213.9,
      "eval_valid_steps_per_second": 6.845,
      "step": 4000
    },
    {
      "epoch": 1.5359017022910533,
      "eval_valid_target_loss": 0.2641640603542328,
      "eval_valid_target_runtime": 4.6656,
      "eval_valid_target_samples_per_second": 214.335,
      "eval_valid_target_steps_per_second": 6.859,
      "step": 4000
    },
    {
      "epoch": 1.5742992448483297,
      "grad_norm": 0.2676081359386444,
      "learning_rate": 9.400930076332126e-06,
      "loss": 0.2602,
      "step": 4100
    },
    {
      "epoch": 1.6126967874056062,
      "grad_norm": 0.24242335557937622,
      "learning_rate": 9.371978596658904e-06,
      "loss": 0.2573,
      "step": 4200
    },
    {
      "epoch": 1.6510943299628824,
      "grad_norm": 0.27868130803108215,
      "learning_rate": 9.342390726173065e-06,
      "loss": 0.2574,
      "step": 4300
    },
    {
      "epoch": 1.6894918725201586,
      "grad_norm": 0.2644180655479431,
      "learning_rate": 9.31217077172299e-06,
      "loss": 0.255,
      "step": 4400
    },
    {
      "epoch": 1.727889415077435,
      "grad_norm": 0.2352069914340973,
      "learning_rate": 9.281323132164013e-06,
      "loss": 0.2538,
      "step": 4500
    },
    {
      "epoch": 1.727889415077435,
      "eval_valid_loss": 0.2354765683412552,
      "eval_valid_runtime": 4.7068,
      "eval_valid_samples_per_second": 212.461,
      "eval_valid_steps_per_second": 6.799,
      "step": 4500
    },
    {
      "epoch": 1.727889415077435,
      "eval_valid_target_loss": 0.25502344965934753,
      "eval_valid_target_runtime": 4.6612,
      "eval_valid_target_samples_per_second": 214.536,
      "eval_valid_target_steps_per_second": 6.865,
      "step": 4500
    },
    {
      "epoch": 1.7662869576347113,
      "grad_norm": 0.28041261434555054,
      "learning_rate": 9.249852297718116e-06,
      "loss": 0.2507,
      "step": 4600
    },
    {
      "epoch": 1.8046845001919878,
      "grad_norm": 0.2735440135002136,
      "learning_rate": 9.217762849320334e-06,
      "loss": 0.2496,
      "step": 4700
    },
    {
      "epoch": 1.843082042749264,
      "grad_norm": 0.26316097378730774,
      "learning_rate": 9.185059457951933e-06,
      "loss": 0.2479,
      "step": 4800
    },
    {
      "epoch": 1.8814795853065402,
      "grad_norm": 0.23891638219356537,
      "learning_rate": 9.151746883960512e-06,
      "loss": 0.2457,
      "step": 4900
    },
    {
      "epoch": 1.9198771278638167,
      "grad_norm": 0.22432874143123627,
      "learning_rate": 9.117829976367072e-06,
      "loss": 0.2446,
      "step": 5000
    },
    {
      "epoch": 1.9198771278638167,
      "eval_valid_loss": 0.2283046841621399,
      "eval_valid_runtime": 4.6829,
      "eval_valid_samples_per_second": 213.544,
      "eval_valid_steps_per_second": 6.833,
      "step": 5000
    },
    {
      "epoch": 1.9198771278638167,
      "eval_valid_target_loss": 0.24809375405311584,
      "eval_valid_target_runtime": 4.6694,
      "eval_valid_target_samples_per_second": 214.162,
      "eval_valid_target_steps_per_second": 6.853,
      "step": 5000
    },
    {
      "epoch": 1.9582746704210932,
      "grad_norm": 0.27488961815834045,
      "learning_rate": 9.08331367216019e-06,
      "loss": 0.2434,
      "step": 5100
    },
    {
      "epoch": 1.9966722129783694,
      "grad_norm": 0.2284267097711563,
      "learning_rate": 9.048202995577383e-06,
      "loss": 0.24,
      "step": 5200
    },
    {
      "epoch": 2.0350697555356456,
      "grad_norm": 0.2710357904434204,
      "learning_rate": 9.012503057373769e-06,
      "loss": 0.2399,
      "step": 5300
    },
    {
      "epoch": 2.073467298092922,
      "grad_norm": 0.24398750066757202,
      "learning_rate": 8.976219054078147e-06,
      "loss": 0.2391,
      "step": 5400
    },
    {
      "epoch": 2.1118648406501985,
      "grad_norm": 0.24732039868831635,
      "learning_rate": 8.939356267236582e-06,
      "loss": 0.2374,
      "step": 5500
    },
    {
      "epoch": 2.1118648406501985,
      "eval_valid_loss": 0.22253906726837158,
      "eval_valid_runtime": 4.6969,
      "eval_valid_samples_per_second": 212.904,
      "eval_valid_steps_per_second": 6.813,
      "step": 5500
    },
    {
      "epoch": 2.1118648406501985,
      "eval_valid_target_loss": 0.24240624904632568,
      "eval_valid_target_runtime": 4.6761,
      "eval_valid_target_samples_per_second": 213.853,
      "eval_valid_target_steps_per_second": 6.843,
      "step": 5500
    },
    {
      "epoch": 2.1502623832074748,
      "grad_norm": 0.23949123919010162,
      "learning_rate": 8.901920062643607e-06,
      "loss": 0.2368,
      "step": 5600
    },
    {
      "epoch": 2.188659925764751,
      "grad_norm": 0.26010605692863464,
      "learning_rate": 8.863915889561188e-06,
      "loss": 0.2351,
      "step": 5700
    },
    {
      "epoch": 2.2270574683220272,
      "grad_norm": 0.2524034380912781,
      "learning_rate": 8.825349279925506e-06,
      "loss": 0.2333,
      "step": 5800
    },
    {
      "epoch": 2.265455010879304,
      "grad_norm": 0.24745632708072662,
      "learning_rate": 8.78622584754173e-06,
      "loss": 0.2323,
      "step": 5900
    },
    {
      "epoch": 2.30385255343658,
      "grad_norm": 0.2586907148361206,
      "learning_rate": 8.746551287266863e-06,
      "loss": 0.2312,
      "step": 6000
    },
    {
      "epoch": 2.30385255343658,
      "eval_valid_loss": 0.216859370470047,
      "eval_valid_runtime": 4.6709,
      "eval_valid_samples_per_second": 214.092,
      "eval_valid_steps_per_second": 6.851,
      "step": 6000
    },
    {
      "epoch": 2.30385255343658,
      "eval_valid_target_loss": 0.23771093785762787,
      "eval_valid_target_runtime": 4.6848,
      "eval_valid_target_samples_per_second": 213.455,
      "eval_valid_target_steps_per_second": 6.831,
      "step": 6000
    },
    {
      "epoch": 2.3422500959938564,
      "grad_norm": 0.24499697983264923,
      "learning_rate": 8.706331374180792e-06,
      "loss": 0.2301,
      "step": 6100
    },
    {
      "epoch": 2.3806476385511326,
      "grad_norm": 0.24237163364887238,
      "learning_rate": 8.665571962745655e-06,
      "loss": 0.2304,
      "step": 6200
    },
    {
      "epoch": 2.419045181108409,
      "grad_norm": 0.27395910024642944,
      "learning_rate": 8.624278985953665e-06,
      "loss": 0.2287,
      "step": 6300
    },
    {
      "epoch": 2.4574427236656855,
      "grad_norm": 0.2500033378601074,
      "learning_rate": 8.582458454463493e-06,
      "loss": 0.2279,
      "step": 6400
    },
    {
      "epoch": 2.4958402662229617,
      "grad_norm": 0.2605977952480316,
      "learning_rate": 8.540116455725346e-06,
      "loss": 0.2277,
      "step": 6500
    },
    {
      "epoch": 2.4958402662229617,
      "eval_valid_loss": 0.21196874976158142,
      "eval_valid_runtime": 4.6941,
      "eval_valid_samples_per_second": 213.035,
      "eval_valid_steps_per_second": 6.817,
      "step": 6500
    },
    {
      "epoch": 2.4958402662229617,
      "eval_valid_target_loss": 0.23328906297683716,
      "eval_valid_target_runtime": 4.6792,
      "eval_valid_target_samples_per_second": 213.712,
      "eval_valid_target_steps_per_second": 6.839,
      "step": 6500
    },
    {
      "epoch": 2.534237808780238,
      "grad_norm": 0.2220095992088318,
      "learning_rate": 8.497259153094875e-06,
      "loss": 0.2254,
      "step": 6600
    },
    {
      "epoch": 2.5726353513375146,
      "grad_norm": 0.24707047641277313,
      "learning_rate": 8.453892784936022e-06,
      "loss": 0.2239,
      "step": 6700
    },
    {
      "epoch": 2.611032893894791,
      "grad_norm": 0.23103290796279907,
      "learning_rate": 8.41002366371297e-06,
      "loss": 0.224,
      "step": 6800
    },
    {
      "epoch": 2.649430436452067,
      "grad_norm": 0.2249547839164734,
      "learning_rate": 8.36565817507127e-06,
      "loss": 0.2227,
      "step": 6900
    },
    {
      "epoch": 2.6878279790093433,
      "grad_norm": 0.24457262456417084,
      "learning_rate": 8.32080277690836e-06,
      "loss": 0.2209,
      "step": 7000
    },
    {
      "epoch": 2.6878279790093433,
      "eval_valid_loss": 0.20793749392032623,
      "eval_valid_runtime": 4.6727,
      "eval_valid_samples_per_second": 214.01,
      "eval_valid_steps_per_second": 6.848,
      "step": 7000
    },
    {
      "epoch": 2.6878279790093433,
      "eval_valid_target_loss": 0.22950781881809235,
      "eval_valid_target_runtime": 4.6848,
      "eval_valid_target_samples_per_second": 213.456,
      "eval_valid_target_steps_per_second": 6.831,
      "step": 7000
    },
    {
      "epoch": 2.7262255215666196,
      "grad_norm": 0.23176012933254242,
      "learning_rate": 8.275463998433537e-06,
      "loss": 0.2206,
      "step": 7100
    },
    {
      "epoch": 2.764623064123896,
      "grad_norm": 0.21723733842372894,
      "learning_rate": 8.229648439217552e-06,
      "loss": 0.2203,
      "step": 7200
    },
    {
      "epoch": 2.8030206066811725,
      "grad_norm": 0.2428179383277893,
      "learning_rate": 8.183362768231971e-06,
      "loss": 0.2192,
      "step": 7300
    },
    {
      "epoch": 2.8414181492384487,
      "grad_norm": 0.2162482738494873,
      "learning_rate": 8.136613722878437e-06,
      "loss": 0.2183,
      "step": 7400
    },
    {
      "epoch": 2.879815691795725,
      "grad_norm": 0.22231200337409973,
      "learning_rate": 8.08940810800796e-06,
      "loss": 0.2177,
      "step": 7500
    },
    {
      "epoch": 2.879815691795725,
      "eval_valid_loss": 0.20469531416893005,
      "eval_valid_runtime": 4.6819,
      "eval_valid_samples_per_second": 213.587,
      "eval_valid_steps_per_second": 6.835,
      "step": 7500
    },
    {
      "epoch": 2.879815691795725,
      "eval_valid_target_loss": 0.2264062464237213,
      "eval_valid_target_runtime": 4.6647,
      "eval_valid_target_samples_per_second": 214.376,
      "eval_valid_target_steps_per_second": 6.86,
      "step": 7500
    },
    {
      "epoch": 2.9182132343530016,
      "grad_norm": 0.2663327157497406,
      "learning_rate": 8.041752794930389e-06,
      "loss": 0.2172,
      "step": 7600
    },
    {
      "epoch": 2.956610776910278,
      "grad_norm": 0.2545444369316101,
      "learning_rate": 7.993654720414227e-06,
      "loss": 0.216,
      "step": 7700
    },
    {
      "epoch": 2.995008319467554,
      "grad_norm": 0.2252371460199356,
      "learning_rate": 7.9451208856769e-06,
      "loss": 0.2154,
      "step": 7800
    },
    {
      "epoch": 3.0334058620248303,
      "grad_norm": 0.2507840394973755,
      "learning_rate": 7.896158355365643e-06,
      "loss": 0.2151,
      "step": 7900
    },
    {
      "epoch": 3.0718034045821065,
      "grad_norm": 0.22570189833641052,
      "learning_rate": 7.846774256529178e-06,
      "loss": 0.2131,
      "step": 8000
    },
    {
      "epoch": 3.0718034045821065,
      "eval_valid_loss": 0.2014453113079071,
      "eval_valid_runtime": 4.6924,
      "eval_valid_samples_per_second": 213.111,
      "eval_valid_steps_per_second": 6.82,
      "step": 8000
    },
    {
      "epoch": 3.0718034045821065,
      "eval_valid_target_loss": 0.22346094250679016,
      "eval_valid_target_runtime": 4.6655,
      "eval_valid_target_samples_per_second": 214.339,
      "eval_valid_target_steps_per_second": 6.859,
      "step": 8000
    },
    {
      "epoch": 3.110200947139383,
      "grad_norm": 0.24750301241874695,
      "learning_rate": 7.796975777580276e-06,
      "loss": 0.2133,
      "step": 8100
    },
    {
      "epoch": 3.1485984896966595,
      "grad_norm": 0.2118765264749527,
      "learning_rate": 7.746770167249413e-06,
      "loss": 0.2124,
      "step": 8200
    },
    {
      "epoch": 3.1869960322539357,
      "grad_norm": 0.22295965254306793,
      "learning_rate": 7.696164733529628e-06,
      "loss": 0.2123,
      "step": 8300
    },
    {
      "epoch": 3.225393574811212,
      "grad_norm": 0.2226712554693222,
      "learning_rate": 7.645166842612766e-06,
      "loss": 0.2115,
      "step": 8400
    },
    {
      "epoch": 3.2637911173684886,
      "grad_norm": 0.22712872922420502,
      "learning_rate": 7.593783917817248e-06,
      "loss": 0.211,
      "step": 8500
    },
    {
      "epoch": 3.2637911173684886,
      "eval_valid_loss": 0.19893750548362732,
      "eval_valid_runtime": 4.6876,
      "eval_valid_samples_per_second": 213.327,
      "eval_valid_steps_per_second": 6.826,
      "step": 8500
    },
    {
      "epoch": 3.2637911173684886,
      "eval_valid_target_loss": 0.22138281166553497,
      "eval_valid_target_runtime": 4.6684,
      "eval_valid_target_samples_per_second": 214.206,
      "eval_valid_target_steps_per_second": 6.855,
      "step": 8500
    },
    {
      "epoch": 3.302188659925765,
      "grad_norm": 0.20663662254810333,
      "learning_rate": 7.5420234385075155e-06,
      "loss": 0.211,
      "step": 8600
    },
    {
      "epoch": 3.340586202483041,
      "grad_norm": 0.24639233946800232,
      "learning_rate": 7.489892939005333e-06,
      "loss": 0.2099,
      "step": 8700
    },
    {
      "epoch": 3.3789837450403173,
      "grad_norm": 0.21435491740703583,
      "learning_rate": 7.437400007493079e-06,
      "loss": 0.209,
      "step": 8800
    },
    {
      "epoch": 3.4173812875975935,
      "grad_norm": 0.21131959557533264,
      "learning_rate": 7.384552284909195e-06,
      "loss": 0.2081,
      "step": 8900
    },
    {
      "epoch": 3.45577883015487,
      "grad_norm": 0.2295517921447754,
      "learning_rate": 7.3313574638359734e-06,
      "loss": 0.2084,
      "step": 9000
    },
    {
      "epoch": 3.45577883015487,
      "eval_valid_loss": 0.19658593833446503,
      "eval_valid_runtime": 4.6935,
      "eval_valid_samples_per_second": 213.059,
      "eval_valid_steps_per_second": 6.818,
      "step": 9000
    },
    {
      "epoch": 3.45577883015487,
      "eval_valid_target_loss": 0.2188750058412552,
      "eval_valid_target_runtime": 4.6686,
      "eval_valid_target_samples_per_second": 214.199,
      "eval_valid_target_steps_per_second": 6.854,
      "step": 9000
    },
    {
      "epoch": 3.4941763727121464,
      "grad_norm": 0.2244088351726532,
      "learning_rate": 7.277823287379801e-06,
      "loss": 0.2084,
      "step": 9100
    },
    {
      "epoch": 3.5325739152694227,
      "grad_norm": 0.2267696112394333,
      "learning_rate": 7.2239575480440774e-06,
      "loss": 0.2085,
      "step": 9200
    },
    {
      "epoch": 3.5709714578266993,
      "grad_norm": 0.20846766233444214,
      "learning_rate": 7.169768086594913e-06,
      "loss": 0.2063,
      "step": 9300
    },
    {
      "epoch": 3.6093690003839756,
      "grad_norm": 0.23632733523845673,
      "learning_rate": 7.115262790919827e-06,
      "loss": 0.2068,
      "step": 9400
    },
    {
      "epoch": 3.647766542941252,
      "grad_norm": 0.20877471566200256,
      "learning_rate": 7.060449594879573e-06,
      "loss": 0.2059,
      "step": 9500
    },
    {
      "epoch": 3.647766542941252,
      "eval_valid_loss": 0.19441406428813934,
      "eval_valid_runtime": 4.6671,
      "eval_valid_samples_per_second": 214.264,
      "eval_valid_steps_per_second": 6.856,
      "step": 9500
    },
    {
      "epoch": 3.647766542941252,
      "eval_valid_target_loss": 0.21704687178134918,
      "eval_valid_target_runtime": 4.6648,
      "eval_valid_target_samples_per_second": 214.371,
      "eval_valid_target_steps_per_second": 6.86,
      "step": 9500
    },
    {
      "epoch": 3.686164085498528,
      "grad_norm": 0.20587915182113647,
      "learning_rate": 7.0053364771532805e-06,
      "loss": 0.2058,
      "step": 9600
    },
    {
      "epoch": 3.7245616280558043,
      "grad_norm": 0.208708256483078,
      "learning_rate": 6.949931460077058e-06,
      "loss": 0.2052,
      "step": 9700
    },
    {
      "epoch": 3.7629591706130805,
      "grad_norm": 0.21517980098724365,
      "learning_rate": 6.894242608476263e-06,
      "loss": 0.2049,
      "step": 9800
    },
    {
      "epoch": 3.801356713170357,
      "grad_norm": 0.22570070624351501,
      "learning_rate": 6.8382780284915685e-06,
      "loss": 0.2047,
      "step": 9900
    },
    {
      "epoch": 3.8397542557276334,
      "grad_norm": 0.22346258163452148,
      "learning_rate": 6.782045866399023e-06,
      "loss": 0.2037,
      "step": 10000
    },
    {
      "epoch": 3.8397542557276334,
      "eval_valid_loss": 0.1928359419107437,
      "eval_valid_runtime": 4.6748,
      "eval_valid_samples_per_second": 213.912,
      "eval_valid_steps_per_second": 6.845,
      "step": 10000
    },
    {
      "epoch": 3.8397542557276334,
      "eval_valid_target_loss": 0.21531249582767487,
      "eval_valid_target_runtime": 4.6773,
      "eval_valid_target_samples_per_second": 213.8,
      "eval_valid_target_steps_per_second": 6.842,
      "step": 10000
    },
    {
      "epoch": 3.8781517982849096,
      "grad_norm": 0.2544507086277008,
      "learning_rate": 6.725554307424274e-06,
      "loss": 0.2036,
      "step": 10100
    },
    {
      "epoch": 3.9165493408421863,
      "grad_norm": 0.27723318338394165,
      "learning_rate": 6.668811574551106e-06,
      "loss": 0.2039,
      "step": 10200
    },
    {
      "epoch": 3.9549468833994625,
      "grad_norm": 0.22496485710144043,
      "learning_rate": 6.6118259273245065e-06,
      "loss": 0.2032,
      "step": 10300
    },
    {
      "epoch": 3.9933444259567388,
      "grad_norm": 0.22093619406223297,
      "learning_rate": 6.55460566064838e-06,
      "loss": 0.2027,
      "step": 10400
    },
    {
      "epoch": 4.031741968514015,
      "grad_norm": 0.2137976437807083,
      "learning_rate": 6.497159103578143e-06,
      "loss": 0.2016,
      "step": 10500
    },
    {
      "epoch": 4.031741968514015,
      "eval_valid_loss": 0.19111718237400055,
      "eval_valid_runtime": 4.6833,
      "eval_valid_samples_per_second": 213.523,
      "eval_valid_steps_per_second": 6.833,
      "step": 10500
    },
    {
      "epoch": 4.031741968514015,
      "eval_valid_target_loss": 0.2142656296491623,
      "eval_valid_target_runtime": 4.6587,
      "eval_valid_target_samples_per_second": 214.65,
      "eval_valid_target_steps_per_second": 6.869,
      "step": 10500
    },
    {
      "epoch": 4.070139511071291,
      "grad_norm": 0.20360158383846283,
      "learning_rate": 6.439494618108332e-06,
      "loss": 0.2013,
      "step": 10600
    },
    {
      "epoch": 4.1085370536285675,
      "grad_norm": 0.21878282725811005,
      "learning_rate": 6.38162059795542e-06,
      "loss": 0.2006,
      "step": 10700
    },
    {
      "epoch": 4.146934596185844,
      "grad_norm": 0.2319776862859726,
      "learning_rate": 6.323545467336017e-06,
      "loss": 0.2012,
      "step": 10800
    },
    {
      "epoch": 4.185332138743121,
      "grad_norm": 0.20898312330245972,
      "learning_rate": 6.26527767974063e-06,
      "loss": 0.2005,
      "step": 10900
    },
    {
      "epoch": 4.223729681300397,
      "grad_norm": 0.21366915106773376,
      "learning_rate": 6.206825716703166e-06,
      "loss": 0.2,
      "step": 11000
    },
    {
      "epoch": 4.223729681300397,
      "eval_valid_loss": 0.18977344036102295,
      "eval_valid_runtime": 4.7328,
      "eval_valid_samples_per_second": 211.293,
      "eval_valid_steps_per_second": 6.761,
      "step": 11000
    },
    {
      "epoch": 4.223729681300397,
      "eval_valid_target_loss": 0.21274219453334808,
      "eval_valid_target_runtime": 4.6506,
      "eval_valid_target_samples_per_second": 215.026,
      "eval_valid_target_steps_per_second": 6.881,
      "step": 11000
    },
    {
      "epoch": 4.262127223857673,
      "grad_norm": 0.20968745648860931,
      "learning_rate": 6.1481980865663405e-06,
      "loss": 0.1993,
      "step": 11100
    },
    {
      "epoch": 4.3005247664149495,
      "grad_norm": 0.20683012902736664,
      "learning_rate": 6.089403323243203e-06,
      "loss": 0.1992,
      "step": 11200
    },
    {
      "epoch": 4.338922308972226,
      "grad_norm": 0.20785097777843475,
      "learning_rate": 6.030449984974916e-06,
      "loss": 0.199,
      "step": 11300
    },
    {
      "epoch": 4.377319851529502,
      "grad_norm": 0.20532238483428955,
      "learning_rate": 5.971346653085025e-06,
      "loss": 0.199,
      "step": 11400
    },
    {
      "epoch": 4.415717394086778,
      "grad_norm": 0.21589842438697815,
      "learning_rate": 5.912101930730329e-06,
      "loss": 0.1992,
      "step": 11500
    },
    {
      "epoch": 4.415717394086778,
      "eval_valid_loss": 0.18833594024181366,
      "eval_valid_runtime": 4.6904,
      "eval_valid_samples_per_second": 213.203,
      "eval_valid_steps_per_second": 6.823,
      "step": 11500
    },
    {
      "epoch": 4.415717394086778,
      "eval_valid_target_loss": 0.211976557970047,
      "eval_valid_target_runtime": 4.658,
      "eval_valid_target_samples_per_second": 214.686,
      "eval_valid_target_steps_per_second": 6.87,
      "step": 11500
    },
    {
      "epoch": 4.4541149366440544,
      "grad_norm": 0.2021540254354477,
      "learning_rate": 5.852724441648614e-06,
      "loss": 0.1987,
      "step": 11600
    },
    {
      "epoch": 4.492512479201331,
      "grad_norm": 0.24406403303146362,
      "learning_rate": 5.7932228289033506e-06,
      "loss": 0.1984,
      "step": 11700
    },
    {
      "epoch": 4.530910021758608,
      "grad_norm": 0.20519228279590607,
      "learning_rate": 5.7336057536256216e-06,
      "loss": 0.1984,
      "step": 11800
    },
    {
      "epoch": 4.569307564315884,
      "grad_norm": 0.21227143704891205,
      "learning_rate": 5.67388189375337e-06,
      "loss": 0.1976,
      "step": 11900
    },
    {
      "epoch": 4.60770510687316,
      "grad_norm": 0.2325662076473236,
      "learning_rate": 5.614059942768254e-06,
      "loss": 0.1977,
      "step": 12000
    },
    {
      "epoch": 4.60770510687316,
      "eval_valid_loss": 0.18742187321186066,
      "eval_valid_runtime": 4.6831,
      "eval_valid_samples_per_second": 213.535,
      "eval_valid_steps_per_second": 6.833,
      "step": 12000
    },
    {
      "epoch": 4.60770510687316,
      "eval_valid_target_loss": 0.21108593046665192,
      "eval_valid_target_runtime": 4.6502,
      "eval_valid_target_samples_per_second": 215.046,
      "eval_valid_target_steps_per_second": 6.881,
      "step": 12000
    },
    {
      "epoch": 4.6461026494304365,
      "grad_norm": 0.2245544046163559,
      "learning_rate": 5.554148608430192e-06,
      "loss": 0.1965,
      "step": 12100
    },
    {
      "epoch": 4.684500191987713,
      "grad_norm": 0.22662824392318726,
      "learning_rate": 5.4941566115098614e-06,
      "loss": 0.1971,
      "step": 12200
    },
    {
      "epoch": 4.722897734544989,
      "grad_norm": 0.19245535135269165,
      "learning_rate": 5.4340926845192874e-06,
      "loss": 0.1974,
      "step": 12300
    },
    {
      "epoch": 4.761295277102265,
      "grad_norm": 0.18942756950855255,
      "learning_rate": 5.373965570440729e-06,
      "loss": 0.1966,
      "step": 12400
    },
    {
      "epoch": 4.799692819659541,
      "grad_norm": 0.1962059736251831,
      "learning_rate": 5.3137840214540395e-06,
      "loss": 0.1958,
      "step": 12500
    },
    {
      "epoch": 4.799692819659541,
      "eval_valid_loss": 0.18663281202316284,
      "eval_valid_runtime": 4.6972,
      "eval_valid_samples_per_second": 212.895,
      "eval_valid_steps_per_second": 6.813,
      "step": 12500
    },
    {
      "epoch": 4.799692819659541,
      "eval_valid_target_loss": 0.21009375154972076,
      "eval_valid_target_runtime": 4.6708,
      "eval_valid_target_samples_per_second": 214.096,
      "eval_valid_target_steps_per_second": 6.851,
      "step": 12500
    },
    {
      "epoch": 4.838090362216818,
      "grad_norm": 0.2151457667350769,
      "learning_rate": 5.2535567976626846e-06,
      "loss": 0.1963,
      "step": 12600
    },
    {
      "epoch": 4.876487904774095,
      "grad_norm": 0.18380814790725708,
      "learning_rate": 5.1932926658186166e-06,
      "loss": 0.1959,
      "step": 12700
    },
    {
      "epoch": 4.914885447331371,
      "grad_norm": 0.19516663253307343,
      "learning_rate": 5.133000398046168e-06,
      "loss": 0.1953,
      "step": 12800
    },
    {
      "epoch": 4.953282989888647,
      "grad_norm": 0.24182352423667908,
      "learning_rate": 5.072688770565177e-06,
      "loss": 0.1953,
      "step": 12900
    },
    {
      "epoch": 4.9916805324459235,
      "grad_norm": 0.23720215260982513,
      "learning_rate": 5.012366562413501e-06,
      "loss": 0.1955,
      "step": 13000
    },
    {
      "epoch": 4.9916805324459235,
      "eval_valid_loss": 0.18524999916553497,
      "eval_valid_runtime": 4.6908,
      "eval_valid_samples_per_second": 213.184,
      "eval_valid_steps_per_second": 6.822,
      "step": 13000
    },
    {
      "epoch": 4.9916805324459235,
      "eval_valid_target_loss": 0.20893749594688416,
      "eval_valid_target_runtime": 4.6667,
      "eval_valid_target_samples_per_second": 214.285,
      "eval_valid_target_steps_per_second": 6.857,
      "step": 13000
    },
    {
      "epoch": 5.0300780750032,
      "grad_norm": 0.20271484553813934,
      "learning_rate": 4.952042554169138e-06,
      "loss": 0.1948,
      "step": 13100
    },
    {
      "epoch": 5.068475617560476,
      "grad_norm": 0.2053770273923874,
      "learning_rate": 4.891725526672107e-06,
      "loss": 0.1947,
      "step": 13200
    },
    {
      "epoch": 5.106873160117752,
      "grad_norm": 0.20811979472637177,
      "learning_rate": 4.8314242597463e-06,
      "loss": 0.1939,
      "step": 13300
    },
    {
      "epoch": 5.145270702675028,
      "grad_norm": 0.19889037311077118,
      "learning_rate": 4.771147530921483e-06,
      "loss": 0.1943,
      "step": 13400
    },
    {
      "epoch": 5.1836682452323055,
      "grad_norm": 0.2038932591676712,
      "learning_rate": 4.710904114155621e-06,
      "loss": 0.1938,
      "step": 13500
    },
    {
      "epoch": 5.1836682452323055,
      "eval_valid_loss": 0.1847265660762787,
      "eval_valid_runtime": 4.698,
      "eval_valid_samples_per_second": 212.854,
      "eval_valid_steps_per_second": 6.811,
      "step": 13500
    },
    {
      "epoch": 5.1836682452323055,
      "eval_valid_target_loss": 0.20839843153953552,
      "eval_valid_target_runtime": 4.6593,
      "eval_valid_target_samples_per_second": 214.626,
      "eval_valid_target_steps_per_second": 6.868,
      "step": 13500
    },
    {
      "epoch": 5.222065787789582,
      "grad_norm": 0.19585560262203217,
      "learning_rate": 4.650702778557736e-06,
      "loss": 0.1932,
      "step": 13600
    },
    {
      "epoch": 5.260463330346858,
      "grad_norm": 0.23953603208065033,
      "learning_rate": 4.59055228711146e-06,
      "loss": 0.1933,
      "step": 13700
    },
    {
      "epoch": 5.298860872904134,
      "grad_norm": 0.21477288007736206,
      "learning_rate": 4.530461395399485e-06,
      "loss": 0.1929,
      "step": 13800
    },
    {
      "epoch": 5.33725841546141,
      "grad_norm": 0.22662727534770966,
      "learning_rate": 4.470438850329089e-06,
      "loss": 0.1935,
      "step": 13900
    },
    {
      "epoch": 5.375655958018687,
      "grad_norm": 0.18912354111671448,
      "learning_rate": 4.410493388858925e-06,
      "loss": 0.1931,
      "step": 14000
    },
    {
      "epoch": 5.375655958018687,
      "eval_valid_loss": 0.18379686772823334,
      "eval_valid_runtime": 4.6729,
      "eval_valid_samples_per_second": 214.001,
      "eval_valid_steps_per_second": 6.848,
      "step": 14000
    },
    {
      "epoch": 5.375655958018687,
      "eval_valid_target_loss": 0.20746874809265137,
      "eval_valid_target_runtime": 4.6581,
      "eval_valid_target_samples_per_second": 214.682,
      "eval_valid_target_steps_per_second": 6.87,
      "step": 14000
    },
    {
      "epoch": 5.414053500575963,
      "grad_norm": 0.21155835688114166,
      "learning_rate": 4.350633736727259e-06,
      "loss": 0.193,
      "step": 14100
    },
    {
      "epoch": 5.452451043133239,
      "grad_norm": 0.2160138338804245,
      "learning_rate": 4.29086860718184e-06,
      "loss": 0.1931,
      "step": 14200
    },
    {
      "epoch": 5.490848585690516,
      "grad_norm": 0.19270409643650055,
      "learning_rate": 4.231206699711587e-06,
      "loss": 0.1925,
      "step": 14300
    },
    {
      "epoch": 5.5292461282477925,
      "grad_norm": 0.18501386046409607,
      "learning_rate": 4.171656698780281e-06,
      "loss": 0.1925,
      "step": 14400
    },
    {
      "epoch": 5.567643670805069,
      "grad_norm": 0.20564299821853638,
      "learning_rate": 4.112227272562447e-06,
      "loss": 0.1918,
      "step": 14500
    },
    {
      "epoch": 5.567643670805069,
      "eval_valid_loss": 0.18317969143390656,
      "eval_valid_runtime": 4.679,
      "eval_valid_samples_per_second": 213.72,
      "eval_valid_steps_per_second": 6.839,
      "step": 14500
    },
    {
      "epoch": 5.567643670805069,
      "eval_valid_target_loss": 0.20700781047344208,
      "eval_valid_target_runtime": 4.674,
      "eval_valid_target_samples_per_second": 213.95,
      "eval_valid_target_steps_per_second": 6.846,
      "step": 14500
    },
    {
      "epoch": 5.606041213362345,
      "grad_norm": 0.21509169042110443,
      "learning_rate": 4.052927071681593e-06,
      "loss": 0.1919,
      "step": 14600
    },
    {
      "epoch": 5.644438755919621,
      "grad_norm": 0.18730491399765015,
      "learning_rate": 3.99376472795103e-06,
      "loss": 0.1921,
      "step": 14700
    },
    {
      "epoch": 5.682836298476897,
      "grad_norm": 0.21269969642162323,
      "learning_rate": 3.934748853117398e-06,
      "loss": 0.1918,
      "step": 14800
    },
    {
      "epoch": 5.721233841034174,
      "grad_norm": 0.18910899758338928,
      "learning_rate": 3.8758880376071415e-06,
      "loss": 0.1914,
      "step": 14900
    },
    {
      "epoch": 5.75963138359145,
      "grad_norm": 0.22251802682876587,
      "learning_rate": 3.8171908492760665e-06,
      "loss": 0.1916,
      "step": 15000
    },
    {
      "epoch": 5.75963138359145,
      "eval_valid_loss": 0.18259374797344208,
      "eval_valid_runtime": 4.67,
      "eval_valid_samples_per_second": 214.134,
      "eval_valid_steps_per_second": 6.852,
      "step": 15000
    },
    {
      "epoch": 5.75963138359145,
      "eval_valid_target_loss": 0.20646093785762787,
      "eval_valid_target_runtime": 4.67,
      "eval_valid_target_samples_per_second": 214.131,
      "eval_valid_target_steps_per_second": 6.852,
      "step": 15000
    },
    {
      "epoch": 5.798028926148726,
      "grad_norm": 0.17328619956970215,
      "learning_rate": 3.758665832162203e-06,
      "loss": 0.1911,
      "step": 15100
    },
    {
      "epoch": 5.836426468706003,
      "grad_norm": 0.20850612223148346,
      "learning_rate": 3.7003215052421116e-06,
      "loss": 0.1915,
      "step": 15200
    },
    {
      "epoch": 5.8748240112632795,
      "grad_norm": 0.1912785917520523,
      "learning_rate": 3.642166361190859e-06,
      "loss": 0.1908,
      "step": 15300
    },
    {
      "epoch": 5.913221553820556,
      "grad_norm": 0.2138790339231491,
      "learning_rate": 3.584208865145812e-06,
      "loss": 0.1907,
      "step": 15400
    },
    {
      "epoch": 5.951619096377832,
      "grad_norm": 0.19723013043403625,
      "learning_rate": 3.5264574534744373e-06,
      "loss": 0.1913,
      "step": 15500
    },
    {
      "epoch": 5.951619096377832,
      "eval_valid_loss": 0.1817968785762787,
      "eval_valid_runtime": 4.6726,
      "eval_valid_samples_per_second": 214.016,
      "eval_valid_steps_per_second": 6.849,
      "step": 15500
    },
    {
      "epoch": 5.951619096377832,
      "eval_valid_target_loss": 0.20574218034744263,
      "eval_valid_target_runtime": 4.6817,
      "eval_valid_target_samples_per_second": 213.599,
      "eval_valid_target_steps_per_second": 6.835,
      "step": 15500
    },
    {
      "epoch": 5.990016638935108,
      "grad_norm": 0.19212548434734344,
      "learning_rate": 3.4689205325462997e-06,
      "loss": 0.1907,
      "step": 15600
    },
    {
      "epoch": 6.028414181492384,
      "grad_norm": 0.19529464840888977,
      "learning_rate": 3.4116064775094126e-06,
      "loss": 0.1901,
      "step": 15700
    },
    {
      "epoch": 6.066811724049661,
      "grad_norm": 0.2088070809841156,
      "learning_rate": 3.354523631071147e-06,
      "loss": 0.1902,
      "step": 15800
    },
    {
      "epoch": 6.105209266606937,
      "grad_norm": 0.19294045865535736,
      "learning_rate": 3.2976803022838514e-06,
      "loss": 0.1903,
      "step": 15900
    },
    {
      "epoch": 6.143606809164213,
      "grad_norm": 0.20844899117946625,
      "learning_rate": 3.2410847653353805e-06,
      "loss": 0.1897,
      "step": 16000
    },
    {
      "epoch": 6.143606809164213,
      "eval_valid_loss": 0.1809999942779541,
      "eval_valid_runtime": 4.6789,
      "eval_valid_samples_per_second": 213.724,
      "eval_valid_steps_per_second": 6.839,
      "step": 16000
    },
    {
      "epoch": 6.143606809164213,
      "eval_valid_target_loss": 0.20546874403953552,
      "eval_valid_target_runtime": 4.6614,
      "eval_valid_target_samples_per_second": 214.53,
      "eval_valid_target_steps_per_second": 6.865,
      "step": 16000
    },
    {
      "epoch": 6.18200435172149,
      "grad_norm": 0.19932307302951813,
      "learning_rate": 3.184745258344688e-06,
      "loss": 0.1894,
      "step": 16100
    },
    {
      "epoch": 6.220401894278766,
      "grad_norm": 0.19776058197021484,
      "learning_rate": 3.128669982162681e-06,
      "loss": 0.1899,
      "step": 16200
    },
    {
      "epoch": 6.258799436836043,
      "grad_norm": 0.20467509329319,
      "learning_rate": 3.07286709917849e-06,
      "loss": 0.1898,
      "step": 16300
    },
    {
      "epoch": 6.297196979393319,
      "grad_norm": 0.19593088328838348,
      "learning_rate": 3.017344732131342e-06,
      "loss": 0.1895,
      "step": 16400
    },
    {
      "epoch": 6.335594521950595,
      "grad_norm": 0.20078891515731812,
      "learning_rate": 2.9621109629282064e-06,
      "loss": 0.1897,
      "step": 16500
    },
    {
      "epoch": 6.335594521950595,
      "eval_valid_loss": 0.1807578057050705,
      "eval_valid_runtime": 4.7017,
      "eval_valid_samples_per_second": 212.687,
      "eval_valid_steps_per_second": 6.806,
      "step": 16500
    },
    {
      "epoch": 6.335594521950595,
      "eval_valid_target_loss": 0.2052578181028366,
      "eval_valid_target_runtime": 4.6726,
      "eval_valid_target_samples_per_second": 214.013,
      "eval_valid_target_steps_per_second": 6.848,
      "step": 16500
    },
    {
      "epoch": 6.373992064507871,
      "grad_norm": 0.17822235822677612,
      "learning_rate": 2.9071738314673758e-06,
      "loss": 0.1889,
      "step": 16600
    },
    {
      "epoch": 6.412389607065148,
      "grad_norm": 0.21160703897476196,
      "learning_rate": 2.8525413344681797e-06,
      "loss": 0.1889,
      "step": 16700
    },
    {
      "epoch": 6.450787149622424,
      "grad_norm": 0.19472962617874146,
      "learning_rate": 2.798221424306953e-06,
      "loss": 0.1894,
      "step": 16800
    },
    {
      "epoch": 6.4891846921797,
      "grad_norm": 0.17923222482204437,
      "learning_rate": 2.744222007859506e-06,
      "loss": 0.1891,
      "step": 16900
    },
    {
      "epoch": 6.527582234736977,
      "grad_norm": 0.18077126145362854,
      "learning_rate": 2.690550945350157e-06,
      "loss": 0.1886,
      "step": 17000
    },
    {
      "epoch": 6.527582234736977,
      "eval_valid_loss": 0.18031249940395355,
      "eval_valid_runtime": 4.6828,
      "eval_valid_samples_per_second": 213.548,
      "eval_valid_steps_per_second": 6.834,
      "step": 17000
    },
    {
      "epoch": 6.527582234736977,
      "eval_valid_target_loss": 0.20450781285762787,
      "eval_valid_target_runtime": 4.6685,
      "eval_valid_target_samples_per_second": 214.203,
      "eval_valid_target_steps_per_second": 6.854,
      "step": 17000
    },
    {
      "epoch": 6.565979777294253,
      "grad_norm": 0.19065329432487488,
      "learning_rate": 2.637216049207615e-06,
      "loss": 0.188,
      "step": 17100
    },
    {
      "epoch": 6.60437731985153,
      "grad_norm": 0.20368430018424988,
      "learning_rate": 2.5842250829277724e-06,
      "loss": 0.189,
      "step": 17200
    },
    {
      "epoch": 6.642774862408806,
      "grad_norm": 0.21131780743598938,
      "learning_rate": 2.5315857599436575e-06,
      "loss": 0.1887,
      "step": 17300
    },
    {
      "epoch": 6.681172404966082,
      "grad_norm": 0.2033446729183197,
      "learning_rate": 2.4793057425026467e-06,
      "loss": 0.1887,
      "step": 17400
    },
    {
      "epoch": 6.719569947523358,
      "grad_norm": 0.19689294695854187,
      "learning_rate": 2.427392640551137e-06,
      "loss": 0.1887,
      "step": 17500
    },
    {
      "epoch": 6.719569947523358,
      "eval_valid_loss": 0.17996874451637268,
      "eval_valid_runtime": 4.7043,
      "eval_valid_samples_per_second": 212.57,
      "eval_valid_steps_per_second": 6.802,
      "step": 17500
    },
    {
      "epoch": 6.719569947523358,
      "eval_valid_target_loss": 0.20432811975479126,
      "eval_valid_target_runtime": 4.6638,
      "eval_valid_target_samples_per_second": 214.416,
      "eval_valid_target_steps_per_second": 6.861,
      "step": 17500
    },
    {
      "epoch": 6.757967490080635,
      "grad_norm": 0.1994999349117279,
      "learning_rate": 2.3758540106268406e-06,
      "loss": 0.1881,
      "step": 17600
    },
    {
      "epoch": 6.796365032637911,
      "grad_norm": 0.19650602340698242,
      "learning_rate": 2.32469735475884e-06,
      "loss": 0.1881,
      "step": 17700
    },
    {
      "epoch": 6.834762575195187,
      "grad_norm": 0.21248474717140198,
      "learning_rate": 2.273930119375586e-06,
      "loss": 0.1882,
      "step": 17800
    },
    {
      "epoch": 6.873160117752464,
      "grad_norm": 0.19042810797691345,
      "learning_rate": 2.2235596942209776e-06,
      "loss": 0.188,
      "step": 17900
    },
    {
      "epoch": 6.91155766030974,
      "grad_norm": 0.23096908628940582,
      "learning_rate": 2.173593411278714e-06,
      "loss": 0.1886,
      "step": 18000
    },
    {
      "epoch": 6.91155766030974,
      "eval_valid_loss": 0.17952343821525574,
      "eval_valid_runtime": 4.6878,
      "eval_valid_samples_per_second": 213.321,
      "eval_valid_steps_per_second": 6.826,
      "step": 18000
    },
    {
      "epoch": 6.91155766030974,
      "eval_valid_target_loss": 0.20391406118869781,
      "eval_valid_target_runtime": 4.6595,
      "eval_valid_target_samples_per_second": 214.617,
      "eval_valid_target_steps_per_second": 6.868,
      "step": 18000
    },
    {
      "epoch": 6.949955202867017,
      "grad_norm": 0.21275204420089722,
      "learning_rate": 2.124038543705034e-06,
      "loss": 0.1878,
      "step": 18100
    },
    {
      "epoch": 6.988352745424293,
      "grad_norm": 0.20453621447086334,
      "learning_rate": 2.0749023047700285e-06,
      "loss": 0.188,
      "step": 18200
    },
    {
      "epoch": 7.026750287981569,
      "grad_norm": 0.20724526047706604,
      "learning_rate": 2.026191846807663e-06,
      "loss": 0.1883,
      "step": 18300
    },
    {
      "epoch": 7.065147830538845,
      "grad_norm": 0.1886543333530426,
      "learning_rate": 1.9779142601746825e-06,
      "loss": 0.1874,
      "step": 18400
    },
    {
      "epoch": 7.1035453730961216,
      "grad_norm": 0.20411571860313416,
      "learning_rate": 1.9300765722185265e-06,
      "loss": 0.187,
      "step": 18500
    },
    {
      "epoch": 7.1035453730961216,
      "eval_valid_loss": 0.17924219369888306,
      "eval_valid_runtime": 4.6825,
      "eval_valid_samples_per_second": 213.561,
      "eval_valid_steps_per_second": 6.834,
      "step": 18500
    },
    {
      "epoch": 7.1035453730961216,
      "eval_valid_target_loss": 0.20393750071525574,
      "eval_valid_target_runtime": 4.6736,
      "eval_valid_target_samples_per_second": 213.97,
      "eval_valid_target_steps_per_second": 6.847,
      "step": 18500
    },
    {
      "epoch": 7.141942915653398,
      "grad_norm": 0.18996645510196686,
      "learning_rate": 1.8826857462544129e-06,
      "loss": 0.1871,
      "step": 18600
    },
    {
      "epoch": 7.180340458210675,
      "grad_norm": 0.21018381416797638,
      "learning_rate": 1.8357486805517615e-06,
      "loss": 0.1874,
      "step": 18700
    },
    {
      "epoch": 7.218738000767951,
      "grad_norm": 0.19617675244808197,
      "learning_rate": 1.7892722073300627e-06,
      "loss": 0.1869,
      "step": 18800
    },
    {
      "epoch": 7.257135543325227,
      "grad_norm": 0.2340448796749115,
      "learning_rate": 1.743263091764379e-06,
      "loss": 0.187,
      "step": 18900
    },
    {
      "epoch": 7.295533085882504,
      "grad_norm": 0.22970305383205414,
      "learning_rate": 1.6977280310005845e-06,
      "loss": 0.1873,
      "step": 19000
    },
    {
      "epoch": 7.295533085882504,
      "eval_valid_loss": 0.1788671910762787,
      "eval_valid_runtime": 4.6706,
      "eval_valid_samples_per_second": 214.105,
      "eval_valid_steps_per_second": 6.851,
      "step": 19000
    },
    {
      "epoch": 7.295533085882504,
      "eval_valid_target_loss": 0.20334374904632568,
      "eval_valid_target_runtime": 4.6842,
      "eval_valid_target_samples_per_second": 213.484,
      "eval_valid_target_steps_per_second": 6.831,
      "step": 19000
    },
    {
      "epoch": 7.33393062843978,
      "grad_norm": 0.20527499914169312,
      "learning_rate": 1.6526736531805354e-06,
      "loss": 0.1873,
      "step": 19100
    },
    {
      "epoch": 7.372328170997056,
      "grad_norm": 0.1835908442735672,
      "learning_rate": 1.6081065164772624e-06,
      "loss": 0.187,
      "step": 19200
    },
    {
      "epoch": 7.410725713554332,
      "grad_norm": 0.18936371803283691,
      "learning_rate": 1.564033108140348e-06,
      "loss": 0.1865,
      "step": 19300
    },
    {
      "epoch": 7.4491232561116085,
      "grad_norm": 0.19136998057365417,
      "learning_rate": 1.520459843551646e-06,
      "loss": 0.1872,
      "step": 19400
    },
    {
      "epoch": 7.487520798668886,
      "grad_norm": 0.19691316783428192,
      "learning_rate": 1.4773930652914426e-06,
      "loss": 0.187,
      "step": 19500
    },
    {
      "epoch": 7.487520798668886,
      "eval_valid_loss": 0.17878125607967377,
      "eval_valid_runtime": 4.6602,
      "eval_valid_samples_per_second": 214.581,
      "eval_valid_steps_per_second": 6.867,
      "step": 19500
    },
    {
      "epoch": 7.487520798668886,
      "eval_valid_target_loss": 0.20325781404972076,
      "eval_valid_target_runtime": 4.6796,
      "eval_valid_target_samples_per_second": 213.695,
      "eval_valid_target_steps_per_second": 6.838,
      "step": 19500
    },
    {
      "epoch": 7.525918341226162,
      "grad_norm": 0.18792080879211426,
      "learning_rate": 1.434839042215227e-06,
      "loss": 0.1868,
      "step": 19600
    },
    {
      "epoch": 7.564315883783438,
      "grad_norm": 0.1945939064025879,
      "learning_rate": 1.3928039685411793e-06,
      "loss": 0.1869,
      "step": 19700
    },
    {
      "epoch": 7.602713426340714,
      "grad_norm": 0.17974095046520233,
      "learning_rate": 1.3512939629485456e-06,
      "loss": 0.187,
      "step": 19800
    },
    {
      "epoch": 7.641110968897991,
      "grad_norm": 0.22416825592517853,
      "learning_rate": 1.3103150676869864e-06,
      "loss": 0.1871,
      "step": 19900
    },
    {
      "epoch": 7.679508511455267,
      "grad_norm": 0.19613422453403473,
      "learning_rate": 1.2698732476970627e-06,
      "loss": 0.1869,
      "step": 20000
    },
    {
      "epoch": 7.679508511455267,
      "eval_valid_loss": 0.1783437430858612,
      "eval_valid_runtime": 4.6716,
      "eval_valid_samples_per_second": 214.058,
      "eval_valid_steps_per_second": 6.85,
      "step": 20000
    },
    {
      "epoch": 7.679508511455267,
      "eval_valid_target_loss": 0.2031562477350235,
      "eval_valid_target_runtime": 4.6803,
      "eval_valid_target_samples_per_second": 213.661,
      "eval_valid_target_steps_per_second": 6.837,
      "step": 20000
    },
    {
      "epoch": 7.717906054012543,
      "grad_norm": 0.20145875215530396,
      "learning_rate": 1.229974389741964e-06,
      "loss": 0.187,
      "step": 20100
    },
    {
      "epoch": 7.756303596569819,
      "grad_norm": 0.18396620452404022,
      "learning_rate": 1.1906243015506375e-06,
      "loss": 0.1867,
      "step": 20200
    },
    {
      "epoch": 7.7947011391270955,
      "grad_norm": 0.18105918169021606,
      "learning_rate": 1.1518287109723958e-06,
      "loss": 0.1862,
      "step": 20300
    },
    {
      "epoch": 7.833098681684373,
      "grad_norm": 0.20986780524253845,
      "learning_rate": 1.1135932651431651e-06,
      "loss": 0.1863,
      "step": 20400
    },
    {
      "epoch": 7.871496224241649,
      "grad_norm": 0.21804456412792206,
      "learning_rate": 1.075923529663489e-06,
      "loss": 0.1869,
      "step": 20500
    },
    {
      "epoch": 7.871496224241649,
      "eval_valid_loss": 0.17836718261241913,
      "eval_valid_runtime": 4.6832,
      "eval_valid_samples_per_second": 213.531,
      "eval_valid_steps_per_second": 6.833,
      "step": 20500
    },
    {
      "epoch": 7.871496224241649,
      "eval_valid_target_loss": 0.20322656631469727,
      "eval_valid_target_runtime": 4.6763,
      "eval_valid_target_samples_per_second": 213.843,
      "eval_valid_target_steps_per_second": 6.843,
      "step": 20500
    },
    {
      "epoch": 7.909893766798925,
      "grad_norm": 0.22019818425178528,
      "learning_rate": 1.0388249877883827e-06,
      "loss": 0.1858,
      "step": 20600
    },
    {
      "epoch": 7.948291309356201,
      "grad_norm": 0.1965310275554657,
      "learning_rate": 1.0023030396291916e-06,
      "loss": 0.1866,
      "step": 20700
    },
    {
      "epoch": 7.9866888519134775,
      "grad_norm": 0.18218408524990082,
      "learning_rate": 9.66363001367534e-07,
      "loss": 0.1869,
      "step": 20800
    },
    {
      "epoch": 8.025086394470755,
      "grad_norm": 0.1850380003452301,
      "learning_rate": 9.310101044814835e-07,
      "loss": 0.1861,
      "step": 20900
    },
    {
      "epoch": 8.06348393702803,
      "grad_norm": 0.18823818862438202,
      "learning_rate": 8.962494949840577e-07,
      "loss": 0.186,
      "step": 21000
    },
    {
      "epoch": 8.06348393702803,
      "eval_valid_loss": 0.17808593809604645,
      "eval_valid_runtime": 4.6916,
      "eval_valid_samples_per_second": 213.147,
      "eval_valid_steps_per_second": 6.821,
      "step": 21000
    },
    {
      "epoch": 8.06348393702803,
      "eval_valid_target_loss": 0.20311719179153442,
      "eval_valid_target_runtime": 4.6653,
      "eval_valid_target_samples_per_second": 214.347,
      "eval_valid_target_steps_per_second": 6.859,
      "step": 21000
    },
    {
      "epoch": 8.101881479585307,
      "grad_norm": 0.20501789450645447,
      "learning_rate": 8.620862326741658e-07,
      "loss": 0.1862,
      "step": 21100
    },
    {
      "epoch": 8.140279022142582,
      "grad_norm": 0.19500133395195007,
      "learning_rate": 8.285252904000906e-07,
      "loss": 0.1862,
      "step": 21200
    },
    {
      "epoch": 8.17867656469986,
      "grad_norm": 0.18742544949054718,
      "learning_rate": 7.955715533356367e-07,
      "loss": 0.1863,
      "step": 21300
    },
    {
      "epoch": 8.217074107257135,
      "grad_norm": 0.20386624336242676,
      "learning_rate": 7.632298182690473e-07,
      "loss": 0.186,
      "step": 21400
    },
    {
      "epoch": 8.255471649814412,
      "grad_norm": 0.17727358639240265,
      "learning_rate": 7.315047929047608e-07,
      "loss": 0.1861,
      "step": 21500
    },
    {
      "epoch": 8.255471649814412,
      "eval_valid_loss": 0.17788280546665192,
      "eval_valid_runtime": 4.679,
      "eval_valid_samples_per_second": 213.72,
      "eval_valid_steps_per_second": 6.839,
      "step": 21500
    },
    {
      "epoch": 8.255471649814412,
      "eval_valid_target_loss": 0.2026640623807907,
      "eval_valid_target_runtime": 4.6709,
      "eval_valid_target_samples_per_second": 214.093,
      "eval_valid_target_steps_per_second": 6.851,
      "step": 21500
    },
    {
      "epoch": 8.293869192371687,
      "grad_norm": 0.19971401989459991,
      "learning_rate": 7.004010951781648e-07,
      "loss": 0.1858,
      "step": 21600
    },
    {
      "epoch": 8.332266734928965,
      "grad_norm": 0.17827193439006805,
      "learning_rate": 6.699232525833987e-07,
      "loss": 0.1868,
      "step": 21700
    },
    {
      "epoch": 8.370664277486242,
      "grad_norm": 0.18275295197963715,
      "learning_rate": 6.400757015143266e-07,
      "loss": 0.1858,
      "step": 21800
    },
    {
      "epoch": 8.409061820043517,
      "grad_norm": 0.19496768712997437,
      "learning_rate": 6.108627866187661e-07,
      "loss": 0.1854,
      "step": 21900
    },
    {
      "epoch": 8.447459362600794,
      "grad_norm": 0.19046269357204437,
      "learning_rate": 5.822887601660832e-07,
      "loss": 0.1862,
      "step": 22000
    },
    {
      "epoch": 8.447459362600794,
      "eval_valid_loss": 0.17781250178813934,
      "eval_valid_runtime": 4.6746,
      "eval_valid_samples_per_second": 213.921,
      "eval_valid_steps_per_second": 6.845,
      "step": 22000
    },
    {
      "epoch": 8.447459362600794,
      "eval_valid_target_loss": 0.2026640623807907,
      "eval_valid_target_runtime": 4.6755,
      "eval_valid_target_samples_per_second": 213.88,
      "eval_valid_target_steps_per_second": 6.844,
      "step": 22000
    },
    {
      "epoch": 8.48585690515807,
      "grad_norm": 0.20896296203136444,
      "learning_rate": 5.543577814282219e-07,
      "loss": 0.1856,
      "step": 22100
    },
    {
      "epoch": 8.524254447715347,
      "grad_norm": 0.19562530517578125,
      "learning_rate": 5.270739160742738e-07,
      "loss": 0.1857,
      "step": 22200
    },
    {
      "epoch": 8.562651990272622,
      "grad_norm": 0.1972120851278305,
      "learning_rate": 5.004411355786792e-07,
      "loss": 0.1863,
      "step": 22300
    },
    {
      "epoch": 8.601049532829899,
      "grad_norm": 0.19712330400943756,
      "learning_rate": 4.7446331664312786e-07,
      "loss": 0.1855,
      "step": 22400
    },
    {
      "epoch": 8.639447075387174,
      "grad_norm": 0.20409992337226868,
      "learning_rate": 4.4914424063226937e-07,
      "loss": 0.1857,
      "step": 22500
    },
    {
      "epoch": 8.639447075387174,
      "eval_valid_loss": 0.17765624821186066,
      "eval_valid_runtime": 4.6769,
      "eval_valid_samples_per_second": 213.818,
      "eval_valid_steps_per_second": 6.842,
      "step": 22500
    },
    {
      "epoch": 8.639447075387174,
      "eval_valid_target_loss": 0.2025781273841858,
      "eval_valid_target_runtime": 4.6696,
      "eval_valid_target_samples_per_second": 214.151,
      "eval_valid_target_steps_per_second": 6.853,
      "step": 22500
    },
    {
      "epoch": 8.677844617944451,
      "grad_norm": 0.21083636581897736,
      "learning_rate": 4.2448759302328336e-07,
      "loss": 0.1861,
      "step": 22600
    },
    {
      "epoch": 8.716242160501729,
      "grad_norm": 0.18778979778289795,
      "learning_rate": 4.0049696286942496e-07,
      "loss": 0.1862,
      "step": 22700
    },
    {
      "epoch": 8.754639703059004,
      "grad_norm": 0.18586015701293945,
      "learning_rate": 3.7717584227759117e-07,
      "loss": 0.1857,
      "step": 22800
    },
    {
      "epoch": 8.793037245616281,
      "grad_norm": 0.1977422684431076,
      "learning_rate": 3.54527625900013e-07,
      "loss": 0.1856,
      "step": 22900
    },
    {
      "epoch": 8.831434788173556,
      "grad_norm": 0.18881608545780182,
      "learning_rate": 3.3255561044011564e-07,
      "loss": 0.1857,
      "step": 23000
    },
    {
      "epoch": 8.831434788173556,
      "eval_valid_loss": 0.17771874368190765,
      "eval_valid_runtime": 4.6727,
      "eval_valid_samples_per_second": 214.01,
      "eval_valid_steps_per_second": 6.848,
      "step": 23000
    },
    {
      "epoch": 8.831434788173556,
      "eval_valid_target_loss": 0.20250000059604645,
      "eval_valid_target_runtime": 4.6666,
      "eval_valid_target_samples_per_second": 214.287,
      "eval_valid_target_steps_per_second": 6.857,
      "step": 23000
    },
    {
      "epoch": 8.869832330730834,
      "grad_norm": 0.2037239372730255,
      "learning_rate": 3.112629941726547e-07,
      "loss": 0.1856,
      "step": 23100
    },
    {
      "epoch": 8.908229873288109,
      "grad_norm": 0.18967826664447784,
      "learning_rate": 2.9065287647816744e-07,
      "loss": 0.1855,
      "step": 23200
    },
    {
      "epoch": 8.946627415845386,
      "grad_norm": 0.17752571403980255,
      "learning_rate": 2.707282573918213e-07,
      "loss": 0.1858,
      "step": 23300
    },
    {
      "epoch": 8.985024958402661,
      "grad_norm": 0.18709731101989746,
      "learning_rate": 2.514920371667301e-07,
      "loss": 0.1854,
      "step": 23400
    },
    {
      "epoch": 9.023422500959938,
      "grad_norm": 0.21643956005573273,
      "learning_rate": 2.3294701585178213e-07,
      "loss": 0.1858,
      "step": 23500
    },
    {
      "epoch": 9.023422500959938,
      "eval_valid_loss": 0.17762500047683716,
      "eval_valid_runtime": 4.6791,
      "eval_valid_samples_per_second": 213.717,
      "eval_valid_steps_per_second": 6.839,
      "step": 23500
    },
    {
      "epoch": 9.023422500959938,
      "eval_valid_target_loss": 0.20255468785762787,
      "eval_valid_target_runtime": 4.7103,
      "eval_valid_target_samples_per_second": 212.301,
      "eval_valid_target_steps_per_second": 6.794,
      "step": 23500
    },
    {
      "epoch": 9.061820043517216,
      "grad_norm": 0.18775244057178497,
      "learning_rate": 2.1509589288407183e-07,
      "loss": 0.1855,
      "step": 23600
    },
    {
      "epoch": 9.100217586074491,
      "grad_norm": 0.17277489602565765,
      "learning_rate": 1.9794126669595403e-07,
      "loss": 0.1859,
      "step": 23700
    },
    {
      "epoch": 9.138615128631768,
      "grad_norm": 0.18996348977088928,
      "learning_rate": 1.8148563433682264e-07,
      "loss": 0.1852,
      "step": 23800
    },
    {
      "epoch": 9.177012671189043,
      "grad_norm": 0.1894453912973404,
      "learning_rate": 1.6573139110963087e-07,
      "loss": 0.1854,
      "step": 23900
    },
    {
      "epoch": 9.21541021374632,
      "grad_norm": 0.2011975795030594,
      "learning_rate": 1.5068083022223346e-07,
      "loss": 0.1855,
      "step": 24000
    },
    {
      "epoch": 9.21541021374632,
      "eval_valid_loss": 0.17754687368869781,
      "eval_valid_runtime": 4.6668,
      "eval_valid_samples_per_second": 214.279,
      "eval_valid_steps_per_second": 6.857,
      "step": 24000
    },
    {
      "epoch": 9.21541021374632,
      "eval_valid_target_loss": 0.20250000059604645,
      "eval_valid_target_runtime": 4.6766,
      "eval_valid_target_samples_per_second": 213.829,
      "eval_valid_target_steps_per_second": 6.843,
      "step": 24000
    },
    {
      "epoch": 9.253807756303596,
      "grad_norm": 0.2087700515985489,
      "learning_rate": 1.3633614245357807e-07,
      "loss": 0.1858,
      "step": 24100
    },
    {
      "epoch": 9.292205298860873,
      "grad_norm": 0.18402153253555298,
      "learning_rate": 1.2269941583481548e-07,
      "loss": 0.1859,
      "step": 24200
    },
    {
      "epoch": 9.330602841418148,
      "grad_norm": 0.17724697291851044,
      "learning_rate": 1.0977263534536597e-07,
      "loss": 0.1856,
      "step": 24300
    },
    {
      "epoch": 9.369000383975425,
      "grad_norm": 0.1847800761461258,
      "learning_rate": 9.755768262397936e-08,
      "loss": 0.1858,
      "step": 24400
    },
    {
      "epoch": 9.407397926532703,
      "grad_norm": 0.1905263364315033,
      "learning_rate": 8.605633569484184e-08,
      "loss": 0.1856,
      "step": 24500
    },
    {
      "epoch": 9.407397926532703,
      "eval_valid_loss": 0.1775546818971634,
      "eval_valid_runtime": 4.6591,
      "eval_valid_samples_per_second": 214.636,
      "eval_valid_steps_per_second": 6.868,
      "step": 24500
    },
    {
      "epoch": 9.407397926532703,
      "eval_valid_target_loss": 0.2024531215429306,
      "eval_valid_target_runtime": 4.6763,
      "eval_valid_target_samples_per_second": 213.844,
      "eval_valid_target_steps_per_second": 6.843,
      "step": 24500
    },
    {
      "epoch": 9.445795469089978,
      "grad_norm": 0.17600856721401215,
      "learning_rate": 7.52702687087653e-08,
      "loss": 0.1855,
      "step": 24600
    },
    {
      "epoch": 9.484193011647255,
      "grad_norm": 0.19071801006793976,
      "learning_rate": 6.520105169949609e-08,
      "loss": 0.1856,
      "step": 24700
    },
    {
      "epoch": 9.52259055420453,
      "grad_norm": 0.20268982648849487,
      "learning_rate": 5.5850150355178936e-08,
      "loss": 0.1855,
      "step": 24800
    },
    {
      "epoch": 9.560988096761807,
      "grad_norm": 0.18069659173488617,
      "learning_rate": 4.721892580500709e-08,
      "loss": 0.1852,
      "step": 24900
    },
    {
      "epoch": 9.599385639319083,
      "grad_norm": 0.19809788465499878,
      "learning_rate": 3.9308634421098e-08,
      "loss": 0.1853,
      "step": 25000
    },
    {
      "epoch": 9.599385639319083,
      "eval_valid_loss": 0.17754687368869781,
      "eval_valid_runtime": 4.6689,
      "eval_valid_samples_per_second": 214.182,
      "eval_valid_steps_per_second": 6.854,
      "step": 25000
    },
    {
      "epoch": 9.599385639319083,
      "eval_valid_target_loss": 0.20237499475479126,
      "eval_valid_target_runtime": 4.688,
      "eval_valid_target_samples_per_second": 213.313,
      "eval_valid_target_steps_per_second": 6.826,
      "step": 25000
    },
    {
      "epoch": 9.63778318187636,
      "grad_norm": 0.1990041732788086,
      "learning_rate": 3.2120427635613517e-08,
      "loss": 0.1852,
      "step": 25100
    },
    {
      "epoch": 9.676180724433635,
      "grad_norm": 0.20578785240650177,
      "learning_rate": 2.565535177315226e-08,
      "loss": 0.185,
      "step": 25200
    },
    {
      "epoch": 9.714578266990912,
      "grad_norm": 0.19831426441669464,
      "learning_rate": 1.991434789845037e-08,
      "loss": 0.1858,
      "step": 25300
    },
    {
      "epoch": 9.75297580954819,
      "grad_norm": 0.18692290782928467,
      "learning_rate": 1.489825167939607e-08,
      "loss": 0.1848,
      "step": 25400
    },
    {
      "epoch": 9.791373352105465,
      "grad_norm": 0.20175856351852417,
      "learning_rate": 1.0607793265389742e-08,
      "loss": 0.1854,
      "step": 25500
    },
    {
      "epoch": 9.791373352105465,
      "eval_valid_loss": 0.17751562595367432,
      "eval_valid_runtime": 4.667,
      "eval_valid_samples_per_second": 214.272,
      "eval_valid_steps_per_second": 6.857,
      "step": 25500
    },
    {
      "epoch": 9.791373352105465,
      "eval_valid_target_loss": 0.20240625739097595,
      "eval_valid_target_runtime": 4.6781,
      "eval_valid_target_samples_per_second": 213.763,
      "eval_valid_target_steps_per_second": 6.84,
      "step": 25500
    },
    {
      "epoch": 9.829770894662742,
      "grad_norm": 0.20650416612625122,
      "learning_rate": 7.0435971810606244e-09,
      "loss": 0.1859,
      "step": 25600
    },
    {
      "epoch": 9.868168437220017,
      "grad_norm": 0.1880464404821396,
      "learning_rate": 4.206182235363399e-09,
      "loss": 0.1857,
      "step": 25700
    },
    {
      "epoch": 9.906565979777294,
      "grad_norm": 0.19517436623573303,
      "learning_rate": 2.095961446056949e-09,
      "loss": 0.1851,
      "step": 25800
    },
    {
      "epoch": 9.94496352233457,
      "grad_norm": 0.21848323941230774,
      "learning_rate": 7.132419795868872e-10,
      "loss": 0.1858,
      "step": 25900
    },
    {
      "epoch": 9.983361064891847,
      "grad_norm": 0.20499403774738312,
      "learning_rate": 5.82251063713235e-11,
      "loss": 0.1851,
      "step": 26000
    },
    {
      "epoch": 9.983361064891847,
      "eval_valid_loss": 0.1775234341621399,
      "eval_valid_runtime": 4.6706,
      "eval_valid_samples_per_second": 214.106,
      "eval_valid_steps_per_second": 6.851,
      "step": 26000
    },
    {
      "epoch": 9.983361064891847,
      "eval_valid_target_loss": 0.20242968201637268,
      "eval_valid_target_runtime": 4.679,
      "eval_valid_target_samples_per_second": 213.721,
      "eval_valid_target_steps_per_second": 6.839,
      "step": 26000
    }
  ],
  "logging_steps": 100,
  "max_steps": 26040,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 10,
  "save_steps": 5000,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 2.475781022436819e+19,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": null
}