{
  "best_metric": 0.46183693408966064,
  "best_model_checkpoint": "Action_all_10_class/checkpoint-1600",
  "epoch": 10.0,
  "eval_steps": 100,
  "global_step": 2890,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.03,
      "grad_norm": 1.9832468032836914,
      "learning_rate": 9.965397923875432e-05,
      "loss": 2.2618,
      "step": 10
    },
    {
      "epoch": 0.07,
      "grad_norm": 1.9726794958114624,
      "learning_rate": 9.930795847750865e-05,
      "loss": 2.1513,
      "step": 20
    },
    {
      "epoch": 0.1,
      "grad_norm": 1.8990365266799927,
      "learning_rate": 9.896193771626297e-05,
      "loss": 2.0562,
      "step": 30
    },
    {
      "epoch": 0.14,
      "grad_norm": 1.9971903562545776,
      "learning_rate": 9.86159169550173e-05,
      "loss": 1.9375,
      "step": 40
    },
    {
      "epoch": 0.17,
      "grad_norm": 2.781921148300171,
      "learning_rate": 9.826989619377162e-05,
      "loss": 1.7386,
      "step": 50
    },
    {
      "epoch": 0.21,
      "grad_norm": 2.6308515071868896,
      "learning_rate": 9.792387543252595e-05,
      "loss": 1.5669,
      "step": 60
    },
    {
      "epoch": 0.24,
      "grad_norm": 3.349245309829712,
      "learning_rate": 9.757785467128027e-05,
      "loss": 1.4557,
      "step": 70
    },
    {
      "epoch": 0.28,
      "grad_norm": 3.6303634643554688,
      "learning_rate": 9.72318339100346e-05,
      "loss": 1.3707,
      "step": 80
    },
    {
      "epoch": 0.31,
      "grad_norm": 4.0813469886779785,
      "learning_rate": 9.688581314878892e-05,
      "loss": 1.3667,
      "step": 90
    },
    {
      "epoch": 0.35,
      "grad_norm": 4.506680488586426,
      "learning_rate": 9.653979238754325e-05,
      "loss": 1.1996,
      "step": 100
    },
    {
      "epoch": 0.35,
      "eval_accuracy": 0.7730061349693251,
      "eval_loss": 1.0634636878967285,
      "eval_runtime": 16.9399,
      "eval_samples_per_second": 48.111,
      "eval_steps_per_second": 6.021,
      "step": 100
    },
    {
      "epoch": 0.38,
      "grad_norm": 2.7484147548675537,
      "learning_rate": 9.619377162629759e-05,
      "loss": 1.1942,
      "step": 110
    },
    {
      "epoch": 0.42,
      "grad_norm": 3.1658833026885986,
      "learning_rate": 9.584775086505192e-05,
      "loss": 1.2623,
      "step": 120
    },
    {
      "epoch": 0.45,
      "grad_norm": 3.366626024246216,
      "learning_rate": 9.550173010380624e-05,
      "loss": 1.0979,
      "step": 130
    },
    {
      "epoch": 0.48,
      "grad_norm": 4.570044994354248,
      "learning_rate": 9.515570934256057e-05,
      "loss": 1.0803,
      "step": 140
    },
    {
      "epoch": 0.52,
      "grad_norm": 3.996645927429199,
      "learning_rate": 9.480968858131488e-05,
      "loss": 1.069,
      "step": 150
    },
    {
      "epoch": 0.55,
      "grad_norm": 4.534286975860596,
      "learning_rate": 9.446366782006922e-05,
      "loss": 1.0673,
      "step": 160
    },
    {
      "epoch": 0.59,
      "grad_norm": 4.0312581062316895,
      "learning_rate": 9.411764705882353e-05,
      "loss": 1.0314,
      "step": 170
    },
    {
      "epoch": 0.62,
      "grad_norm": 3.871572494506836,
      "learning_rate": 9.380622837370243e-05,
      "loss": 1.056,
      "step": 180
    },
    {
      "epoch": 0.66,
      "grad_norm": 4.05521821975708,
      "learning_rate": 9.346020761245676e-05,
      "loss": 0.9677,
      "step": 190
    },
    {
      "epoch": 0.69,
      "grad_norm": 5.704447269439697,
      "learning_rate": 9.311418685121109e-05,
      "loss": 1.0335,
      "step": 200
    },
    {
      "epoch": 0.69,
      "eval_accuracy": 0.7717791411042945,
      "eval_loss": 0.8391550779342651,
      "eval_runtime": 11.6297,
      "eval_samples_per_second": 70.079,
      "eval_steps_per_second": 8.771,
      "step": 200
    },
    {
      "epoch": 0.73,
      "grad_norm": 6.939559459686279,
      "learning_rate": 9.27681660899654e-05,
      "loss": 0.9873,
      "step": 210
    },
    {
      "epoch": 0.76,
      "grad_norm": 4.459825038909912,
      "learning_rate": 9.242214532871972e-05,
      "loss": 0.9642,
      "step": 220
    },
    {
      "epoch": 0.8,
      "grad_norm": 5.480250835418701,
      "learning_rate": 9.207612456747406e-05,
      "loss": 0.9012,
      "step": 230
    },
    {
      "epoch": 0.83,
      "grad_norm": 3.0692737102508545,
      "learning_rate": 9.173010380622837e-05,
      "loss": 0.9618,
      "step": 240
    },
    {
      "epoch": 0.87,
      "grad_norm": 2.740853786468506,
      "learning_rate": 9.13840830449827e-05,
      "loss": 0.8414,
      "step": 250
    },
    {
      "epoch": 0.9,
      "grad_norm": 3.637920379638672,
      "learning_rate": 9.103806228373702e-05,
      "loss": 0.908,
      "step": 260
    },
    {
      "epoch": 0.93,
      "grad_norm": 3.488283634185791,
      "learning_rate": 9.069204152249136e-05,
      "loss": 0.9001,
      "step": 270
    },
    {
      "epoch": 0.97,
      "grad_norm": 4.264199256896973,
      "learning_rate": 9.034602076124567e-05,
      "loss": 0.8283,
      "step": 280
    },
    {
      "epoch": 1.0,
      "grad_norm": 3.3874425888061523,
      "learning_rate": 9e-05,
      "loss": 0.7862,
      "step": 290
    },
    {
      "epoch": 1.04,
      "grad_norm": 3.339252233505249,
      "learning_rate": 8.965397923875432e-05,
      "loss": 0.6279,
      "step": 300
    },
    {
      "epoch": 1.04,
      "eval_accuracy": 0.8294478527607362,
      "eval_loss": 0.6463050842285156,
      "eval_runtime": 12.9239,
      "eval_samples_per_second": 63.061,
      "eval_steps_per_second": 7.892,
      "step": 300
    },
    {
      "epoch": 1.07,
      "grad_norm": 5.215779781341553,
      "learning_rate": 8.930795847750866e-05,
      "loss": 0.7314,
      "step": 310
    },
    {
      "epoch": 1.11,
      "grad_norm": 5.00214958190918,
      "learning_rate": 8.896193771626297e-05,
      "loss": 0.6857,
      "step": 320
    },
    {
      "epoch": 1.14,
      "grad_norm": 4.8398542404174805,
      "learning_rate": 8.86159169550173e-05,
      "loss": 0.767,
      "step": 330
    },
    {
      "epoch": 1.18,
      "grad_norm": 4.89760160446167,
      "learning_rate": 8.826989619377162e-05,
      "loss": 0.6085,
      "step": 340
    },
    {
      "epoch": 1.21,
      "grad_norm": 3.459028720855713,
      "learning_rate": 8.792387543252596e-05,
      "loss": 0.6735,
      "step": 350
    },
    {
      "epoch": 1.25,
      "grad_norm": 5.60736608505249,
      "learning_rate": 8.757785467128027e-05,
      "loss": 0.8843,
      "step": 360
    },
    {
      "epoch": 1.28,
      "grad_norm": 4.042627811431885,
      "learning_rate": 8.72318339100346e-05,
      "loss": 0.7819,
      "step": 370
    },
    {
      "epoch": 1.31,
      "grad_norm": 3.9558098316192627,
      "learning_rate": 8.688581314878894e-05,
      "loss": 0.8151,
      "step": 380
    },
    {
      "epoch": 1.35,
      "grad_norm": 2.589348554611206,
      "learning_rate": 8.653979238754327e-05,
      "loss": 0.7008,
      "step": 390
    },
    {
      "epoch": 1.38,
      "grad_norm": 5.536679744720459,
      "learning_rate": 8.619377162629759e-05,
      "loss": 0.8633,
      "step": 400
    },
    {
      "epoch": 1.38,
      "eval_accuracy": 0.7926380368098159,
      "eval_loss": 0.7172089219093323,
      "eval_runtime": 11.6484,
      "eval_samples_per_second": 69.967,
      "eval_steps_per_second": 8.757,
      "step": 400
    },
    {
      "epoch": 1.42,
      "grad_norm": 3.1925909519195557,
      "learning_rate": 8.584775086505192e-05,
      "loss": 0.7146,
      "step": 410
    },
    {
      "epoch": 1.45,
      "grad_norm": 1.3583850860595703,
      "learning_rate": 8.550173010380624e-05,
      "loss": 0.6679,
      "step": 420
    },
    {
      "epoch": 1.49,
      "grad_norm": 2.476952075958252,
      "learning_rate": 8.515570934256055e-05,
      "loss": 0.7093,
      "step": 430
    },
    {
      "epoch": 1.52,
      "grad_norm": 4.362217426300049,
      "learning_rate": 8.480968858131489e-05,
      "loss": 0.5813,
      "step": 440
    },
    {
      "epoch": 1.56,
      "grad_norm": 6.683523178100586,
      "learning_rate": 8.44636678200692e-05,
      "loss": 0.6022,
      "step": 450
    },
    {
      "epoch": 1.59,
      "grad_norm": 5.650263786315918,
      "learning_rate": 8.411764705882354e-05,
      "loss": 0.6538,
      "step": 460
    },
    {
      "epoch": 1.63,
      "grad_norm": 2.9296507835388184,
      "learning_rate": 8.377162629757785e-05,
      "loss": 0.7312,
      "step": 470
    },
    {
      "epoch": 1.66,
      "grad_norm": 4.2093048095703125,
      "learning_rate": 8.342560553633219e-05,
      "loss": 0.5509,
      "step": 480
    },
    {
      "epoch": 1.7,
      "grad_norm": 4.481240749359131,
      "learning_rate": 8.30795847750865e-05,
      "loss": 0.7108,
      "step": 490
    },
    {
      "epoch": 1.73,
      "grad_norm": 6.15458869934082,
      "learning_rate": 8.273356401384084e-05,
      "loss": 0.5851,
      "step": 500
    },
    {
      "epoch": 1.73,
      "eval_accuracy": 0.838036809815951,
      "eval_loss": 0.5857788324356079,
      "eval_runtime": 11.1755,
      "eval_samples_per_second": 72.927,
      "eval_steps_per_second": 9.127,
      "step": 500
    },
    {
      "epoch": 1.76,
      "grad_norm": 5.6646409034729,
      "learning_rate": 8.238754325259515e-05,
      "loss": 0.7466,
      "step": 510
    },
    {
      "epoch": 1.8,
      "grad_norm": 3.708621025085449,
      "learning_rate": 8.204152249134948e-05,
      "loss": 0.6417,
      "step": 520
    },
    {
      "epoch": 1.83,
      "grad_norm": 4.252438545227051,
      "learning_rate": 8.16955017301038e-05,
      "loss": 0.5843,
      "step": 530
    },
    {
      "epoch": 1.87,
      "grad_norm": 6.532431602478027,
      "learning_rate": 8.134948096885813e-05,
      "loss": 0.7773,
      "step": 540
    },
    {
      "epoch": 1.9,
      "grad_norm": 2.0311672687530518,
      "learning_rate": 8.100346020761245e-05,
      "loss": 0.6273,
      "step": 550
    },
    {
      "epoch": 1.94,
      "grad_norm": 4.033968448638916,
      "learning_rate": 8.065743944636678e-05,
      "loss": 0.6618,
      "step": 560
    },
    {
      "epoch": 1.97,
      "grad_norm": 3.268759250640869,
      "learning_rate": 8.03114186851211e-05,
      "loss": 0.5404,
      "step": 570
    },
    {
      "epoch": 2.01,
      "grad_norm": 5.866196632385254,
      "learning_rate": 7.996539792387543e-05,
      "loss": 0.6579,
      "step": 580
    },
    {
      "epoch": 2.04,
      "grad_norm": 1.0091450214385986,
      "learning_rate": 7.961937716262977e-05,
      "loss": 0.4575,
      "step": 590
    },
    {
      "epoch": 2.08,
      "grad_norm": 4.265220642089844,
      "learning_rate": 7.92733564013841e-05,
      "loss": 0.5305,
      "step": 600
    },
    {
      "epoch": 2.08,
      "eval_accuracy": 0.8355828220858895,
      "eval_loss": 0.5779993534088135,
      "eval_runtime": 12.1337,
      "eval_samples_per_second": 67.168,
      "eval_steps_per_second": 8.406,
      "step": 600
    },
    {
      "epoch": 2.11,
      "grad_norm": 3.7668848037719727,
      "learning_rate": 7.892733564013842e-05,
      "loss": 0.5928,
      "step": 610
    },
    {
      "epoch": 2.15,
      "grad_norm": 3.6394691467285156,
      "learning_rate": 7.858131487889275e-05,
      "loss": 0.533,
      "step": 620
    },
    {
      "epoch": 2.18,
      "grad_norm": 4.226925373077393,
      "learning_rate": 7.823529411764707e-05,
      "loss": 0.4776,
      "step": 630
    },
    {
      "epoch": 2.21,
      "grad_norm": 6.159184455871582,
      "learning_rate": 7.78892733564014e-05,
      "loss": 0.5517,
      "step": 640
    },
    {
      "epoch": 2.25,
      "grad_norm": 7.021705150604248,
      "learning_rate": 7.754325259515572e-05,
      "loss": 0.4969,
      "step": 650
    },
    {
      "epoch": 2.28,
      "grad_norm": 5.880521297454834,
      "learning_rate": 7.719723183391003e-05,
      "loss": 0.5172,
      "step": 660
    },
    {
      "epoch": 2.32,
      "grad_norm": 4.049355506896973,
      "learning_rate": 7.685121107266437e-05,
      "loss": 0.5188,
      "step": 670
    },
    {
      "epoch": 2.35,
      "grad_norm": 6.603725433349609,
      "learning_rate": 7.650519031141868e-05,
      "loss": 0.5329,
      "step": 680
    },
    {
      "epoch": 2.39,
      "grad_norm": 6.908223628997803,
      "learning_rate": 7.615916955017301e-05,
      "loss": 0.5938,
      "step": 690
    },
    {
      "epoch": 2.42,
      "grad_norm": 5.728999614715576,
      "learning_rate": 7.581314878892733e-05,
      "loss": 0.5511,
      "step": 700
    },
    {
      "epoch": 2.42,
      "eval_accuracy": 0.8392638036809816,
      "eval_loss": 0.5313203930854797,
      "eval_runtime": 11.9712,
      "eval_samples_per_second": 68.08,
      "eval_steps_per_second": 8.52,
      "step": 700
    },
    {
      "epoch": 2.46,
      "grad_norm": 3.414422035217285,
      "learning_rate": 7.546712802768166e-05,
      "loss": 0.4702,
      "step": 710
    },
    {
      "epoch": 2.49,
      "grad_norm": 7.404133319854736,
      "learning_rate": 7.512110726643598e-05,
      "loss": 0.5877,
      "step": 720
    },
    {
      "epoch": 2.53,
      "grad_norm": 4.851989269256592,
      "learning_rate": 7.477508650519031e-05,
      "loss": 0.5855,
      "step": 730
    },
    {
      "epoch": 2.56,
      "grad_norm": 6.651949882507324,
      "learning_rate": 7.442906574394463e-05,
      "loss": 0.5618,
      "step": 740
    },
    {
      "epoch": 2.6,
      "grad_norm": 9.513969421386719,
      "learning_rate": 7.408304498269896e-05,
      "loss": 0.5871,
      "step": 750
    },
    {
      "epoch": 2.63,
      "grad_norm": 11.049631118774414,
      "learning_rate": 7.373702422145328e-05,
      "loss": 0.4098,
      "step": 760
    },
    {
      "epoch": 2.66,
      "grad_norm": 4.604189872741699,
      "learning_rate": 7.339100346020761e-05,
      "loss": 0.5752,
      "step": 770
    },
    {
      "epoch": 2.7,
      "grad_norm": 5.650027275085449,
      "learning_rate": 7.304498269896195e-05,
      "loss": 0.6265,
      "step": 780
    },
    {
      "epoch": 2.73,
      "grad_norm": 7.081886291503906,
      "learning_rate": 7.269896193771626e-05,
      "loss": 0.4814,
      "step": 790
    },
    {
      "epoch": 2.77,
      "grad_norm": 9.59929084777832,
      "learning_rate": 7.23529411764706e-05,
      "loss": 0.4657,
      "step": 800
    },
    {
      "epoch": 2.77,
      "eval_accuracy": 0.8368098159509203,
      "eval_loss": 0.5443019866943359,
      "eval_runtime": 14.7103,
      "eval_samples_per_second": 55.403,
      "eval_steps_per_second": 6.934,
      "step": 800
    },
    {
      "epoch": 2.8,
      "grad_norm": 3.963488817214966,
      "learning_rate": 7.200692041522493e-05,
      "loss": 0.4586,
      "step": 810
    },
    {
      "epoch": 2.84,
      "grad_norm": 5.124209403991699,
      "learning_rate": 7.166089965397925e-05,
      "loss": 0.4099,
      "step": 820
    },
    {
      "epoch": 2.87,
      "grad_norm": 2.2323999404907227,
      "learning_rate": 7.131487889273358e-05,
      "loss": 0.5683,
      "step": 830
    },
    {
      "epoch": 2.91,
      "grad_norm": 6.542498588562012,
      "learning_rate": 7.09688581314879e-05,
      "loss": 0.4908,
      "step": 840
    },
    {
      "epoch": 2.94,
      "grad_norm": 5.042673110961914,
      "learning_rate": 7.062283737024223e-05,
      "loss": 0.5441,
      "step": 850
    },
    {
      "epoch": 2.98,
      "grad_norm": 7.417181491851807,
      "learning_rate": 7.027681660899654e-05,
      "loss": 0.4688,
      "step": 860
    },
    {
      "epoch": 3.01,
      "grad_norm": 3.395101547241211,
      "learning_rate": 6.993079584775086e-05,
      "loss": 0.6159,
      "step": 870
    },
    {
      "epoch": 3.04,
      "grad_norm": 4.752573013305664,
      "learning_rate": 6.95847750865052e-05,
      "loss": 0.4304,
      "step": 880
    },
    {
      "epoch": 3.08,
      "grad_norm": 6.356733322143555,
      "learning_rate": 6.923875432525951e-05,
      "loss": 0.4116,
      "step": 890
    },
    {
      "epoch": 3.11,
      "grad_norm": 3.9821412563323975,
      "learning_rate": 6.889273356401384e-05,
      "loss": 0.3615,
      "step": 900
    },
    {
      "epoch": 3.11,
      "eval_accuracy": 0.8429447852760736,
      "eval_loss": 0.5038174986839294,
      "eval_runtime": 11.7579,
      "eval_samples_per_second": 69.315,
      "eval_steps_per_second": 8.675,
      "step": 900
    },
    {
      "epoch": 3.15,
      "grad_norm": 4.009409427642822,
      "learning_rate": 6.854671280276816e-05,
      "loss": 0.3844,
      "step": 910
    },
    {
      "epoch": 3.18,
      "grad_norm": 6.364121437072754,
      "learning_rate": 6.82006920415225e-05,
      "loss": 0.3863,
      "step": 920
    },
    {
      "epoch": 3.22,
      "grad_norm": 5.382229804992676,
      "learning_rate": 6.785467128027681e-05,
      "loss": 0.4159,
      "step": 930
    },
    {
      "epoch": 3.25,
      "grad_norm": 3.8384909629821777,
      "learning_rate": 6.750865051903114e-05,
      "loss": 0.4783,
      "step": 940
    },
    {
      "epoch": 3.29,
      "grad_norm": 0.2637701630592346,
      "learning_rate": 6.716262975778546e-05,
      "loss": 0.4072,
      "step": 950
    },
    {
      "epoch": 3.32,
      "grad_norm": 3.8876795768737793,
      "learning_rate": 6.68166089965398e-05,
      "loss": 0.3698,
      "step": 960
    },
    {
      "epoch": 3.36,
      "grad_norm": 2.9388973712921143,
      "learning_rate": 6.647058823529411e-05,
      "loss": 0.4197,
      "step": 970
    },
    {
      "epoch": 3.39,
      "grad_norm": 1.423226237297058,
      "learning_rate": 6.612456747404844e-05,
      "loss": 0.344,
      "step": 980
    },
    {
      "epoch": 3.43,
      "grad_norm": 5.080854892730713,
      "learning_rate": 6.577854671280278e-05,
      "loss": 0.3596,
      "step": 990
    },
    {
      "epoch": 3.46,
      "grad_norm": 4.027952671051025,
      "learning_rate": 6.54325259515571e-05,
      "loss": 0.5301,
      "step": 1000
    },
    {
      "epoch": 3.46,
      "eval_accuracy": 0.8503067484662576,
      "eval_loss": 0.5100666284561157,
      "eval_runtime": 11.1502,
      "eval_samples_per_second": 73.093,
      "eval_steps_per_second": 9.148,
      "step": 1000
    },
    {
      "epoch": 3.49,
      "grad_norm": 7.976325988769531,
      "learning_rate": 6.508650519031142e-05,
      "loss": 0.3365,
      "step": 1010
    },
    {
      "epoch": 3.53,
      "grad_norm": 5.674125671386719,
      "learning_rate": 6.474048442906576e-05,
      "loss": 0.3745,
      "step": 1020
    },
    {
      "epoch": 3.56,
      "grad_norm": 2.311941623687744,
      "learning_rate": 6.439446366782007e-05,
      "loss": 0.4887,
      "step": 1030
    },
    {
      "epoch": 3.6,
      "grad_norm": 2.7413113117218018,
      "learning_rate": 6.40484429065744e-05,
      "loss": 0.4548,
      "step": 1040
    },
    {
      "epoch": 3.63,
      "grad_norm": 4.1170830726623535,
      "learning_rate": 6.370242214532872e-05,
      "loss": 0.5191,
      "step": 1050
    },
    {
      "epoch": 3.67,
      "grad_norm": 3.560255765914917,
      "learning_rate": 6.335640138408306e-05,
      "loss": 0.496,
      "step": 1060
    },
    {
      "epoch": 3.7,
      "grad_norm": 7.074652671813965,
      "learning_rate": 6.301038062283737e-05,
      "loss": 0.5216,
      "step": 1070
    },
    {
      "epoch": 3.74,
      "grad_norm": 2.202094793319702,
      "learning_rate": 6.26643598615917e-05,
      "loss": 0.3693,
      "step": 1080
    },
    {
      "epoch": 3.77,
      "grad_norm": 5.584536075592041,
      "learning_rate": 6.231833910034602e-05,
      "loss": 0.3557,
      "step": 1090
    },
    {
      "epoch": 3.81,
      "grad_norm": 4.547154903411865,
      "learning_rate": 6.197231833910034e-05,
      "loss": 0.4108,
      "step": 1100
    },
    {
      "epoch": 3.81,
      "eval_accuracy": 0.8478527607361963,
      "eval_loss": 0.5211547613143921,
      "eval_runtime": 11.576,
      "eval_samples_per_second": 70.404,
      "eval_steps_per_second": 8.811,
      "step": 1100
    },
    {
      "epoch": 3.84,
      "grad_norm": 2.4568958282470703,
      "learning_rate": 6.162629757785467e-05,
      "loss": 0.4367,
      "step": 1110
    },
    {
      "epoch": 3.88,
      "grad_norm": 3.632456064224243,
      "learning_rate": 6.128027681660899e-05,
      "loss": 0.4811,
      "step": 1120
    },
    {
      "epoch": 3.91,
      "grad_norm": 5.1356329917907715,
      "learning_rate": 6.0934256055363324e-05,
      "loss": 0.3121,
      "step": 1130
    },
    {
      "epoch": 3.94,
      "grad_norm": 2.870601177215576,
      "learning_rate": 6.058823529411765e-05,
      "loss": 0.42,
      "step": 1140
    },
    {
      "epoch": 3.98,
      "grad_norm": 6.305639743804932,
      "learning_rate": 6.0242214532871973e-05,
      "loss": 0.4733,
      "step": 1150
    },
    {
      "epoch": 4.01,
      "grad_norm": 5.95077657699585,
      "learning_rate": 5.98961937716263e-05,
      "loss": 0.3604,
      "step": 1160
    },
    {
      "epoch": 4.05,
      "grad_norm": 4.88993501663208,
      "learning_rate": 5.955017301038063e-05,
      "loss": 0.3475,
      "step": 1170
    },
    {
      "epoch": 4.08,
      "grad_norm": 2.9184603691101074,
      "learning_rate": 5.920415224913495e-05,
      "loss": 0.395,
      "step": 1180
    },
    {
      "epoch": 4.12,
      "grad_norm": 3.779115915298462,
      "learning_rate": 5.885813148788928e-05,
      "loss": 0.4199,
      "step": 1190
    },
    {
      "epoch": 4.15,
      "grad_norm": 4.5922322273254395,
      "learning_rate": 5.854671280276816e-05,
      "loss": 0.4223,
      "step": 1200
    },
    {
      "epoch": 4.15,
      "eval_accuracy": 0.8429447852760736,
      "eval_loss": 0.5328340530395508,
      "eval_runtime": 11.5883,
      "eval_samples_per_second": 70.33,
      "eval_steps_per_second": 8.802,
      "step": 1200
    },
    {
      "epoch": 4.19,
      "grad_norm": 5.457287311553955,
      "learning_rate": 5.8200692041522495e-05,
      "loss": 0.2775,
      "step": 1210
    },
    {
      "epoch": 4.22,
      "grad_norm": 5.398082733154297,
      "learning_rate": 5.785467128027682e-05,
      "loss": 0.3019,
      "step": 1220
    },
    {
      "epoch": 4.26,
      "grad_norm": 5.261845111846924,
      "learning_rate": 5.750865051903115e-05,
      "loss": 0.4365,
      "step": 1230
    },
    {
      "epoch": 4.29,
      "grad_norm": 5.616456508636475,
      "learning_rate": 5.716262975778547e-05,
      "loss": 0.412,
      "step": 1240
    },
    {
      "epoch": 4.33,
      "grad_norm": 3.715630292892456,
      "learning_rate": 5.68166089965398e-05,
      "loss": 0.3644,
      "step": 1250
    },
    {
      "epoch": 4.36,
      "grad_norm": 5.539583683013916,
      "learning_rate": 5.647058823529412e-05,
      "loss": 0.2519,
      "step": 1260
    },
    {
      "epoch": 4.39,
      "grad_norm": 7.531044960021973,
      "learning_rate": 5.612456747404845e-05,
      "loss": 0.4028,
      "step": 1270
    },
    {
      "epoch": 4.43,
      "grad_norm": 4.245101451873779,
      "learning_rate": 5.577854671280277e-05,
      "loss": 0.483,
      "step": 1280
    },
    {
      "epoch": 4.46,
      "grad_norm": 4.674008846282959,
      "learning_rate": 5.54325259515571e-05,
      "loss": 0.4101,
      "step": 1290
    },
    {
      "epoch": 4.5,
      "grad_norm": 5.6747331619262695,
      "learning_rate": 5.508650519031142e-05,
      "loss": 0.3877,
      "step": 1300
    },
    {
      "epoch": 4.5,
      "eval_accuracy": 0.8294478527607362,
      "eval_loss": 0.5815134048461914,
      "eval_runtime": 11.4501,
      "eval_samples_per_second": 71.178,
      "eval_steps_per_second": 8.908,
      "step": 1300
    },
    {
      "epoch": 4.53,
      "grad_norm": 4.670008659362793,
      "learning_rate": 5.4740484429065744e-05,
      "loss": 0.4169,
      "step": 1310
    },
    {
      "epoch": 4.57,
      "grad_norm": 6.947169780731201,
      "learning_rate": 5.4394463667820076e-05,
      "loss": 0.3955,
      "step": 1320
    },
    {
      "epoch": 4.6,
      "grad_norm": 0.6032439470291138,
      "learning_rate": 5.4048442906574394e-05,
      "loss": 0.3438,
      "step": 1330
    },
    {
      "epoch": 4.64,
      "grad_norm": 2.0391006469726562,
      "learning_rate": 5.3702422145328725e-05,
      "loss": 0.3032,
      "step": 1340
    },
    {
      "epoch": 4.67,
      "grad_norm": 1.2689828872680664,
      "learning_rate": 5.3356401384083043e-05,
      "loss": 0.3091,
      "step": 1350
    },
    {
      "epoch": 4.71,
      "grad_norm": 9.588872909545898,
      "learning_rate": 5.3010380622837375e-05,
      "loss": 0.3301,
      "step": 1360
    },
    {
      "epoch": 4.74,
      "grad_norm": 9.207768440246582,
      "learning_rate": 5.266435986159169e-05,
      "loss": 0.4184,
      "step": 1370
    },
    {
      "epoch": 4.78,
      "grad_norm": 3.842557191848755,
      "learning_rate": 5.2318339100346025e-05,
      "loss": 0.3254,
      "step": 1380
    },
    {
      "epoch": 4.81,
      "grad_norm": 5.358198642730713,
      "learning_rate": 5.197231833910034e-05,
      "loss": 0.2876,
      "step": 1390
    },
    {
      "epoch": 4.84,
      "grad_norm": 4.02155065536499,
      "learning_rate": 5.1626297577854675e-05,
      "loss": 0.3879,
      "step": 1400
    },
    {
      "epoch": 4.84,
      "eval_accuracy": 0.8503067484662576,
      "eval_loss": 0.5150614380836487,
      "eval_runtime": 11.4927,
      "eval_samples_per_second": 70.914,
      "eval_steps_per_second": 8.875,
      "step": 1400
    },
    {
      "epoch": 4.88,
      "grad_norm": 4.721061706542969,
      "learning_rate": 5.1280276816609e-05,
      "loss": 0.4051,
      "step": 1410
    },
    {
      "epoch": 4.91,
      "grad_norm": 6.063509941101074,
      "learning_rate": 5.0934256055363325e-05,
      "loss": 0.436,
      "step": 1420
    },
    {
      "epoch": 4.95,
      "grad_norm": 4.278838634490967,
      "learning_rate": 5.058823529411765e-05,
      "loss": 0.3533,
      "step": 1430
    },
    {
      "epoch": 4.98,
      "grad_norm": 5.166705131530762,
      "learning_rate": 5.024221453287198e-05,
      "loss": 0.3281,
      "step": 1440
    },
    {
      "epoch": 5.02,
      "grad_norm": 4.790309906005859,
      "learning_rate": 4.98961937716263e-05,
      "loss": 0.3324,
      "step": 1450
    },
    {
      "epoch": 5.05,
      "grad_norm": 6.665502548217773,
      "learning_rate": 4.9550173010380624e-05,
      "loss": 0.4309,
      "step": 1460
    },
    {
      "epoch": 5.09,
      "grad_norm": 2.508721113204956,
      "learning_rate": 4.920415224913495e-05,
      "loss": 0.4029,
      "step": 1470
    },
    {
      "epoch": 5.12,
      "grad_norm": 3.7062671184539795,
      "learning_rate": 4.8858131487889274e-05,
      "loss": 0.3016,
      "step": 1480
    },
    {
      "epoch": 5.16,
      "grad_norm": 1.6657217741012573,
      "learning_rate": 4.85121107266436e-05,
      "loss": 0.2906,
      "step": 1490
    },
    {
      "epoch": 5.19,
      "grad_norm": 5.122851848602295,
      "learning_rate": 4.8166089965397924e-05,
      "loss": 0.2797,
      "step": 1500
    },
    {
      "epoch": 5.19,
      "eval_accuracy": 0.8564417177914111,
      "eval_loss": 0.5160326361656189,
      "eval_runtime": 11.658,
      "eval_samples_per_second": 69.909,
      "eval_steps_per_second": 8.749,
      "step": 1500
    },
    {
      "epoch": 5.22,
      "grad_norm": 1.0947901010513306,
      "learning_rate": 4.782006920415225e-05,
      "loss": 0.2272,
      "step": 1510
    },
    {
      "epoch": 5.26,
      "grad_norm": 6.223177909851074,
      "learning_rate": 4.747404844290658e-05,
      "loss": 0.4336,
      "step": 1520
    },
    {
      "epoch": 5.29,
      "grad_norm": 4.786419868469238,
      "learning_rate": 4.7128027681660905e-05,
      "loss": 0.3644,
      "step": 1530
    },
    {
      "epoch": 5.33,
      "grad_norm": 2.7506916522979736,
      "learning_rate": 4.678200692041523e-05,
      "loss": 0.3686,
      "step": 1540
    },
    {
      "epoch": 5.36,
      "grad_norm": 5.9919352531433105,
      "learning_rate": 4.6435986159169555e-05,
      "loss": 0.2422,
      "step": 1550
    },
    {
      "epoch": 5.4,
      "grad_norm": 2.702258825302124,
      "learning_rate": 4.608996539792388e-05,
      "loss": 0.4158,
      "step": 1560
    },
    {
      "epoch": 5.43,
      "grad_norm": 0.17089949548244476,
      "learning_rate": 4.5743944636678205e-05,
      "loss": 0.2326,
      "step": 1570
    },
    {
      "epoch": 5.47,
      "grad_norm": 1.4725955724716187,
      "learning_rate": 4.539792387543253e-05,
      "loss": 0.2891,
      "step": 1580
    },
    {
      "epoch": 5.5,
      "grad_norm": 3.754875659942627,
      "learning_rate": 4.5051903114186854e-05,
      "loss": 0.3169,
      "step": 1590
    },
    {
      "epoch": 5.54,
      "grad_norm": 1.3562734127044678,
      "learning_rate": 4.470588235294118e-05,
      "loss": 0.2628,
      "step": 1600
    },
    {
      "epoch": 5.54,
      "eval_accuracy": 0.8699386503067484,
      "eval_loss": 0.46183693408966064,
      "eval_runtime": 12.3863,
      "eval_samples_per_second": 65.799,
      "eval_steps_per_second": 8.235,
      "step": 1600
    },
    {
      "epoch": 5.57,
      "grad_norm": 1.3551547527313232,
      "learning_rate": 4.4359861591695504e-05,
      "loss": 0.2924,
      "step": 1610
    },
    {
      "epoch": 5.61,
      "grad_norm": 2.6148874759674072,
      "learning_rate": 4.401384083044983e-05,
      "loss": 0.2792,
      "step": 1620
    },
    {
      "epoch": 5.64,
      "grad_norm": 3.0203983783721924,
      "learning_rate": 4.3667820069204154e-05,
      "loss": 0.2955,
      "step": 1630
    },
    {
      "epoch": 5.67,
      "grad_norm": 10.279276847839355,
      "learning_rate": 4.332179930795848e-05,
      "loss": 0.4448,
      "step": 1640
    },
    {
      "epoch": 5.71,
      "grad_norm": 0.37972888350486755,
      "learning_rate": 4.2975778546712804e-05,
      "loss": 0.2687,
      "step": 1650
    },
    {
      "epoch": 5.74,
      "grad_norm": 5.29146671295166,
      "learning_rate": 4.262975778546713e-05,
      "loss": 0.2522,
      "step": 1660
    },
    {
      "epoch": 5.78,
      "grad_norm": 1.9747039079666138,
      "learning_rate": 4.2283737024221454e-05,
      "loss": 0.2579,
      "step": 1670
    },
    {
      "epoch": 5.81,
      "grad_norm": 1.1432113647460938,
      "learning_rate": 4.193771626297578e-05,
      "loss": 0.2522,
      "step": 1680
    },
    {
      "epoch": 5.85,
      "grad_norm": 0.5862625241279602,
      "learning_rate": 4.15916955017301e-05,
      "loss": 0.265,
      "step": 1690
    },
    {
      "epoch": 5.88,
      "grad_norm": 5.743435859680176,
      "learning_rate": 4.124567474048443e-05,
      "loss": 0.3404,
      "step": 1700
    },
    {
      "epoch": 5.88,
      "eval_accuracy": 0.8674846625766871,
      "eval_loss": 0.4903254508972168,
      "eval_runtime": 11.4951,
      "eval_samples_per_second": 70.9,
      "eval_steps_per_second": 8.873,
      "step": 1700
    },
    {
      "epoch": 5.92,
      "grad_norm": 2.558229446411133,
      "learning_rate": 4.089965397923875e-05,
      "loss": 0.3082,
      "step": 1710
    },
    {
      "epoch": 5.95,
      "grad_norm": 1.225205659866333,
      "learning_rate": 4.055363321799308e-05,
      "loss": 0.2875,
      "step": 1720
    },
    {
      "epoch": 5.99,
      "grad_norm": 0.44406113028526306,
      "learning_rate": 4.020761245674741e-05,
      "loss": 0.2987,
      "step": 1730
    },
    {
      "epoch": 6.02,
      "grad_norm": 6.770223617553711,
      "learning_rate": 3.9861591695501735e-05,
      "loss": 0.381,
      "step": 1740
    },
    {
      "epoch": 6.06,
      "grad_norm": 5.378862380981445,
      "learning_rate": 3.951557093425606e-05,
      "loss": 0.3109,
      "step": 1750
    },
    {
      "epoch": 6.09,
      "grad_norm": 5.086582183837891,
      "learning_rate": 3.9169550173010384e-05,
      "loss": 0.3031,
      "step": 1760
    },
    {
      "epoch": 6.12,
      "grad_norm": 3.3848397731781006,
      "learning_rate": 3.882352941176471e-05,
      "loss": 0.3092,
      "step": 1770
    },
    {
      "epoch": 6.16,
      "grad_norm": 1.4123233556747437,
      "learning_rate": 3.8477508650519034e-05,
      "loss": 0.352,
      "step": 1780
    },
    {
      "epoch": 6.19,
      "grad_norm": 1.3674020767211914,
      "learning_rate": 3.813148788927336e-05,
      "loss": 0.1128,
      "step": 1790
    },
    {
      "epoch": 6.23,
      "grad_norm": 3.944838047027588,
      "learning_rate": 3.7785467128027684e-05,
      "loss": 0.3033,
      "step": 1800
    },
    {
      "epoch": 6.23,
      "eval_accuracy": 0.8662576687116564,
      "eval_loss": 0.4860667288303375,
      "eval_runtime": 11.4194,
      "eval_samples_per_second": 71.37,
      "eval_steps_per_second": 8.932,
      "step": 1800
    },
    {
      "epoch": 6.26,
      "grad_norm": 4.455410957336426,
      "learning_rate": 3.743944636678201e-05,
      "loss": 0.2315,
      "step": 1810
    },
    {
      "epoch": 6.3,
      "grad_norm": 3.5710418224334717,
      "learning_rate": 3.7093425605536334e-05,
      "loss": 0.2175,
      "step": 1820
    },
    {
      "epoch": 6.33,
      "grad_norm": 6.50364875793457,
      "learning_rate": 3.674740484429066e-05,
      "loss": 0.2477,
      "step": 1830
    },
    {
      "epoch": 6.37,
      "grad_norm": 2.710531711578369,
      "learning_rate": 3.640138408304499e-05,
      "loss": 0.221,
      "step": 1840
    },
    {
      "epoch": 6.4,
      "grad_norm": 2.1464736461639404,
      "learning_rate": 3.605536332179931e-05,
      "loss": 0.2387,
      "step": 1850
    },
    {
      "epoch": 6.44,
      "grad_norm": 7.897326469421387,
      "learning_rate": 3.570934256055363e-05,
      "loss": 0.2723,
      "step": 1860
    },
    {
      "epoch": 6.47,
      "grad_norm": 2.9161367416381836,
      "learning_rate": 3.536332179930796e-05,
      "loss": 0.2621,
      "step": 1870
    },
    {
      "epoch": 6.51,
      "grad_norm": 7.048449516296387,
      "learning_rate": 3.501730103806228e-05,
      "loss": 0.2955,
      "step": 1880
    },
    {
      "epoch": 6.54,
      "grad_norm": 5.2737250328063965,
      "learning_rate": 3.467128027681661e-05,
      "loss": 0.3076,
      "step": 1890
    },
    {
      "epoch": 6.57,
      "grad_norm": 4.985352039337158,
      "learning_rate": 3.432525951557093e-05,
      "loss": 0.214,
      "step": 1900
    },
    {
      "epoch": 6.57,
      "eval_accuracy": 0.8687116564417178,
      "eval_loss": 0.4853067100048065,
      "eval_runtime": 11.47,
      "eval_samples_per_second": 71.055,
      "eval_steps_per_second": 8.893,
      "step": 1900
    },
    {
      "epoch": 6.61,
      "grad_norm": 6.846111297607422,
      "learning_rate": 3.397923875432526e-05,
      "loss": 0.2693,
      "step": 1910
    },
    {
      "epoch": 6.64,
      "grad_norm": 2.8362717628479004,
      "learning_rate": 3.363321799307958e-05,
      "loss": 0.2582,
      "step": 1920
    },
    {
      "epoch": 6.68,
      "grad_norm": 1.0130655765533447,
      "learning_rate": 3.3287197231833914e-05,
      "loss": 0.278,
      "step": 1930
    },
    {
      "epoch": 6.71,
      "grad_norm": 1.3184999227523804,
      "learning_rate": 3.294117647058824e-05,
      "loss": 0.3776,
      "step": 1940
    },
    {
      "epoch": 6.75,
      "grad_norm": 3.637080430984497,
      "learning_rate": 3.2595155709342564e-05,
      "loss": 0.1598,
      "step": 1950
    },
    {
      "epoch": 6.78,
      "grad_norm": 4.093498706817627,
      "learning_rate": 3.224913494809689e-05,
      "loss": 0.2604,
      "step": 1960
    },
    {
      "epoch": 6.82,
      "grad_norm": 6.110915184020996,
      "learning_rate": 3.1903114186851214e-05,
      "loss": 0.3471,
      "step": 1970
    },
    {
      "epoch": 6.85,
      "grad_norm": 6.045133590698242,
      "learning_rate": 3.155709342560554e-05,
      "loss": 0.1886,
      "step": 1980
    },
    {
      "epoch": 6.89,
      "grad_norm": 0.4137258231639862,
      "learning_rate": 3.1211072664359864e-05,
      "loss": 0.183,
      "step": 1990
    },
    {
      "epoch": 6.92,
      "grad_norm": 3.405696392059326,
      "learning_rate": 3.086505190311419e-05,
      "loss": 0.2763,
      "step": 2000
    },
    {
      "epoch": 6.92,
      "eval_accuracy": 0.8736196319018404,
      "eval_loss": 0.4705219566822052,
      "eval_runtime": 11.4342,
      "eval_samples_per_second": 71.277,
      "eval_steps_per_second": 8.921,
      "step": 2000
    },
    {
      "epoch": 6.96,
      "grad_norm": 6.817298889160156,
      "learning_rate": 3.0519031141868513e-05,
      "loss": 0.2825,
      "step": 2010
    },
    {
      "epoch": 6.99,
      "grad_norm": 5.460371494293213,
      "learning_rate": 3.0173010380622842e-05,
      "loss": 0.2853,
      "step": 2020
    },
    {
      "epoch": 7.02,
      "grad_norm": 6.517573833465576,
      "learning_rate": 2.9826989619377167e-05,
      "loss": 0.2913,
      "step": 2030
    },
    {
      "epoch": 7.06,
      "grad_norm": 2.988943099975586,
      "learning_rate": 2.948096885813149e-05,
      "loss": 0.3647,
      "step": 2040
    },
    {
      "epoch": 7.09,
      "grad_norm": 3.221747636795044,
      "learning_rate": 2.9134948096885816e-05,
      "loss": 0.2344,
      "step": 2050
    },
    {
      "epoch": 7.13,
      "grad_norm": 3.342573404312134,
      "learning_rate": 2.878892733564014e-05,
      "loss": 0.2741,
      "step": 2060
    },
    {
      "epoch": 7.16,
      "grad_norm": 4.326902866363525,
      "learning_rate": 2.8442906574394463e-05,
      "loss": 0.2503,
      "step": 2070
    },
    {
      "epoch": 7.2,
      "grad_norm": 7.965173244476318,
      "learning_rate": 2.8096885813148788e-05,
      "loss": 0.2061,
      "step": 2080
    },
    {
      "epoch": 7.23,
      "grad_norm": 0.314100056886673,
      "learning_rate": 2.7750865051903113e-05,
      "loss": 0.2205,
      "step": 2090
    },
    {
      "epoch": 7.27,
      "grad_norm": 4.415227890014648,
      "learning_rate": 2.7404844290657437e-05,
      "loss": 0.3009,
      "step": 2100
    },
    {
      "epoch": 7.27,
      "eval_accuracy": 0.8625766871165644,
      "eval_loss": 0.47227197885513306,
      "eval_runtime": 11.4055,
      "eval_samples_per_second": 71.457,
      "eval_steps_per_second": 8.943,
      "step": 2100
    },
    {
      "epoch": 7.3,
      "grad_norm": 4.705826759338379,
      "learning_rate": 2.7058823529411766e-05,
      "loss": 0.2255,
      "step": 2110
    },
    {
      "epoch": 7.34,
      "grad_norm": 1.2122201919555664,
      "learning_rate": 2.671280276816609e-05,
      "loss": 0.2103,
      "step": 2120
    },
    {
      "epoch": 7.37,
      "grad_norm": 11.807892799377441,
      "learning_rate": 2.6366782006920416e-05,
      "loss": 0.2001,
      "step": 2130
    },
    {
      "epoch": 7.4,
      "grad_norm": 5.092467784881592,
      "learning_rate": 2.602076124567474e-05,
      "loss": 0.2208,
      "step": 2140
    },
    {
      "epoch": 7.44,
      "grad_norm": 2.7508227825164795,
      "learning_rate": 2.5674740484429065e-05,
      "loss": 0.2397,
      "step": 2150
    },
    {
      "epoch": 7.47,
      "grad_norm": 5.095257759094238,
      "learning_rate": 2.532871972318339e-05,
      "loss": 0.1861,
      "step": 2160
    },
    {
      "epoch": 7.51,
      "grad_norm": 2.8208508491516113,
      "learning_rate": 2.498269896193772e-05,
      "loss": 0.1776,
      "step": 2170
    },
    {
      "epoch": 7.54,
      "grad_norm": 16.025461196899414,
      "learning_rate": 2.4636678200692043e-05,
      "loss": 0.2899,
      "step": 2180
    },
    {
      "epoch": 7.58,
      "grad_norm": 8.661275863647461,
      "learning_rate": 2.4290657439446368e-05,
      "loss": 0.3018,
      "step": 2190
    },
    {
      "epoch": 7.61,
      "grad_norm": 4.188193321228027,
      "learning_rate": 2.3944636678200693e-05,
      "loss": 0.1543,
      "step": 2200
    },
    {
      "epoch": 7.61,
      "eval_accuracy": 0.8638036809815951,
      "eval_loss": 0.4983120858669281,
      "eval_runtime": 11.4805,
      "eval_samples_per_second": 70.99,
      "eval_steps_per_second": 8.885,
      "step": 2200
    },
    {
      "epoch": 7.65,
      "grad_norm": 2.8907430171966553,
      "learning_rate": 2.3598615916955018e-05,
      "loss": 0.2289,
      "step": 2210
    },
    {
      "epoch": 7.68,
      "grad_norm": 6.132763385772705,
      "learning_rate": 2.3252595155709346e-05,
      "loss": 0.1687,
      "step": 2220
    },
    {
      "epoch": 7.72,
      "grad_norm": 8.132895469665527,
      "learning_rate": 2.290657439446367e-05,
      "loss": 0.1936,
      "step": 2230
    },
    {
      "epoch": 7.75,
      "grad_norm": 2.7748658657073975,
      "learning_rate": 2.2560553633217993e-05,
      "loss": 0.2018,
      "step": 2240
    },
    {
      "epoch": 7.79,
      "grad_norm": 3.723444938659668,
      "learning_rate": 2.2214532871972318e-05,
      "loss": 0.331,
      "step": 2250
    },
    {
      "epoch": 7.82,
      "grad_norm": 0.9698171615600586,
      "learning_rate": 2.1868512110726642e-05,
      "loss": 0.2071,
      "step": 2260
    },
    {
      "epoch": 7.85,
      "grad_norm": 5.589043140411377,
      "learning_rate": 2.152249134948097e-05,
      "loss": 0.1308,
      "step": 2270
    },
    {
      "epoch": 7.89,
      "grad_norm": 1.0597747564315796,
      "learning_rate": 2.1176470588235296e-05,
      "loss": 0.2911,
      "step": 2280
    },
    {
      "epoch": 7.92,
      "grad_norm": 0.17304980754852295,
      "learning_rate": 2.083044982698962e-05,
      "loss": 0.3127,
      "step": 2290
    },
    {
      "epoch": 7.96,
      "grad_norm": 2.3576455116271973,
      "learning_rate": 2.0484429065743945e-05,
      "loss": 0.2407,
      "step": 2300
    },
    {
      "epoch": 7.96,
      "eval_accuracy": 0.8650306748466258,
      "eval_loss": 0.47417759895324707,
      "eval_runtime": 11.7355,
      "eval_samples_per_second": 69.447,
      "eval_steps_per_second": 8.692,
      "step": 2300
    },
    {
      "epoch": 7.99,
      "grad_norm": 6.587134838104248,
      "learning_rate": 2.013840830449827e-05,
      "loss": 0.1954,
      "step": 2310
    },
    {
      "epoch": 8.03,
      "grad_norm": 5.971607685089111,
      "learning_rate": 1.9792387543252595e-05,
      "loss": 0.139,
      "step": 2320
    },
    {
      "epoch": 8.06,
      "grad_norm": 2.352544069290161,
      "learning_rate": 1.9446366782006923e-05,
      "loss": 0.1648,
      "step": 2330
    },
    {
      "epoch": 8.1,
      "grad_norm": 1.0423003435134888,
      "learning_rate": 1.910034602076125e-05,
      "loss": 0.3657,
      "step": 2340
    },
    {
      "epoch": 8.13,
      "grad_norm": 3.612190008163452,
      "learning_rate": 1.875432525951557e-05,
      "loss": 0.1846,
      "step": 2350
    },
    {
      "epoch": 8.17,
      "grad_norm": 3.1284472942352295,
      "learning_rate": 1.8408304498269895e-05,
      "loss": 0.2876,
      "step": 2360
    },
    {
      "epoch": 8.2,
      "grad_norm": 6.02339506149292,
      "learning_rate": 1.8062283737024223e-05,
      "loss": 0.3388,
      "step": 2370
    },
    {
      "epoch": 8.24,
      "grad_norm": 7.388739109039307,
      "learning_rate": 1.7716262975778548e-05,
      "loss": 0.3095,
      "step": 2380
    },
    {
      "epoch": 8.27,
      "grad_norm": 6.4854912757873535,
      "learning_rate": 1.7370242214532873e-05,
      "loss": 0.1786,
      "step": 2390
    },
    {
      "epoch": 8.3,
      "grad_norm": 3.4759864807128906,
      "learning_rate": 1.7024221453287198e-05,
      "loss": 0.2679,
      "step": 2400
    },
    {
      "epoch": 8.3,
      "eval_accuracy": 0.8723926380368098,
      "eval_loss": 0.4934893846511841,
      "eval_runtime": 11.4149,
      "eval_samples_per_second": 71.398,
      "eval_steps_per_second": 8.936,
      "step": 2400
    },
    {
      "epoch": 8.34,
      "grad_norm": 0.9941917657852173,
      "learning_rate": 1.6678200692041523e-05,
      "loss": 0.1666,
      "step": 2410
    },
    {
      "epoch": 8.37,
      "grad_norm": 2.615389108657837,
      "learning_rate": 1.6332179930795848e-05,
      "loss": 0.2039,
      "step": 2420
    },
    {
      "epoch": 8.41,
      "grad_norm": 0.12697245180606842,
      "learning_rate": 1.5986159169550176e-05,
      "loss": 0.2026,
      "step": 2430
    },
    {
      "epoch": 8.44,
      "grad_norm": 2.1617543697357178,
      "learning_rate": 1.56401384083045e-05,
      "loss": 0.2141,
      "step": 2440
    },
    {
      "epoch": 8.48,
      "grad_norm": 6.248876094818115,
      "learning_rate": 1.5294117647058826e-05,
      "loss": 0.1039,
      "step": 2450
    },
    {
      "epoch": 8.51,
      "grad_norm": 3.2735066413879395,
      "learning_rate": 1.4948096885813149e-05,
      "loss": 0.269,
      "step": 2460
    },
    {
      "epoch": 8.55,
      "grad_norm": 0.2865382730960846,
      "learning_rate": 1.4602076124567474e-05,
      "loss": 0.1724,
      "step": 2470
    },
    {
      "epoch": 8.58,
      "grad_norm": 0.17031227052211761,
      "learning_rate": 1.4256055363321799e-05,
      "loss": 0.1477,
      "step": 2480
    },
    {
      "epoch": 8.62,
      "grad_norm": 5.686491966247559,
      "learning_rate": 1.3910034602076125e-05,
      "loss": 0.2774,
      "step": 2490
    },
    {
      "epoch": 8.65,
      "grad_norm": 0.2951521873474121,
      "learning_rate": 1.356401384083045e-05,
      "loss": 0.1508,
      "step": 2500
    },
    {
      "epoch": 8.65,
      "eval_accuracy": 0.8674846625766871,
      "eval_loss": 0.4825909435749054,
      "eval_runtime": 11.5579,
      "eval_samples_per_second": 70.515,
      "eval_steps_per_second": 8.825,
      "step": 2500
    },
    {
      "epoch": 8.69,
      "grad_norm": 4.706648826599121,
      "learning_rate": 1.3217993079584775e-05,
      "loss": 0.1022,
      "step": 2510
    },
    {
      "epoch": 8.72,
      "grad_norm": 7.546770095825195,
      "learning_rate": 1.2871972318339102e-05,
      "loss": 0.1778,
      "step": 2520
    },
    {
      "epoch": 8.75,
      "grad_norm": 8.35574722290039,
      "learning_rate": 1.2525951557093426e-05,
      "loss": 0.242,
      "step": 2530
    },
    {
      "epoch": 8.79,
      "grad_norm": 2.115328788757324,
      "learning_rate": 1.2179930795847751e-05,
      "loss": 0.1634,
      "step": 2540
    },
    {
      "epoch": 8.82,
      "grad_norm": 7.35970401763916,
      "learning_rate": 1.1833910034602076e-05,
      "loss": 0.1618,
      "step": 2550
    },
    {
      "epoch": 8.86,
      "grad_norm": 0.3500753939151764,
      "learning_rate": 1.1487889273356401e-05,
      "loss": 0.2685,
      "step": 2560
    },
    {
      "epoch": 8.89,
      "grad_norm": 4.974752426147461,
      "learning_rate": 1.1141868512110728e-05,
      "loss": 0.1975,
      "step": 2570
    },
    {
      "epoch": 8.93,
      "grad_norm": 2.2218804359436035,
      "learning_rate": 1.0795847750865053e-05,
      "loss": 0.1887,
      "step": 2580
    },
    {
      "epoch": 8.96,
      "grad_norm": 2.4438514709472656,
      "learning_rate": 1.0449826989619377e-05,
      "loss": 0.2348,
      "step": 2590
    },
    {
      "epoch": 9.0,
      "grad_norm": 7.3225998878479,
      "learning_rate": 1.0103806228373702e-05,
      "loss": 0.2129,
      "step": 2600
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.8711656441717791,
      "eval_loss": 0.49807000160217285,
      "eval_runtime": 11.7207,
      "eval_samples_per_second": 69.535,
      "eval_steps_per_second": 8.703,
      "step": 2600
    },
    {
      "epoch": 9.03,
      "grad_norm": 3.7790167331695557,
      "learning_rate": 9.757785467128027e-06,
      "loss": 0.1668,
      "step": 2610
    },
    {
      "epoch": 9.07,
      "grad_norm": 0.14809288084506989,
      "learning_rate": 9.411764705882354e-06,
      "loss": 0.1686,
      "step": 2620
    },
    {
      "epoch": 9.1,
      "grad_norm": 0.21380960941314697,
      "learning_rate": 9.065743944636679e-06,
      "loss": 0.2266,
      "step": 2630
    },
    {
      "epoch": 9.13,
      "grad_norm": 1.157197117805481,
      "learning_rate": 8.719723183391004e-06,
      "loss": 0.1608,
      "step": 2640
    },
    {
      "epoch": 9.17,
      "grad_norm": 1.3275336027145386,
      "learning_rate": 8.37370242214533e-06,
      "loss": 0.1909,
      "step": 2650
    },
    {
      "epoch": 9.2,
      "grad_norm": 0.963979959487915,
      "learning_rate": 8.027681660899653e-06,
      "loss": 0.1803,
      "step": 2660
    },
    {
      "epoch": 9.24,
      "grad_norm": NaN,
      "learning_rate": 7.716262975778547e-06,
      "loss": 0.1407,
      "step": 2670
    },
    {
      "epoch": 9.27,
      "grad_norm": 2.7852437496185303,
      "learning_rate": 7.370242214532873e-06,
      "loss": 0.2366,
      "step": 2680
    },
    {
      "epoch": 9.31,
      "grad_norm": 0.2830531895160675,
      "learning_rate": 7.024221453287197e-06,
      "loss": 0.2925,
      "step": 2690
    },
    {
      "epoch": 9.34,
      "grad_norm": 2.379260301589966,
      "learning_rate": 6.678200692041523e-06,
      "loss": 0.1131,
      "step": 2700
    },
    {
      "epoch": 9.34,
      "eval_accuracy": 0.8711656441717791,
      "eval_loss": 0.47180497646331787,
      "eval_runtime": 11.6646,
      "eval_samples_per_second": 69.87,
      "eval_steps_per_second": 8.744,
      "step": 2700
    },
    {
      "epoch": 9.38,
      "grad_norm": 7.354953289031982,
      "learning_rate": 6.3321799307958475e-06,
      "loss": 0.1997,
      "step": 2710
    },
    {
      "epoch": 9.41,
      "grad_norm": 0.7862864136695862,
      "learning_rate": 5.986159169550173e-06,
      "loss": 0.1147,
      "step": 2720
    },
    {
      "epoch": 9.45,
      "grad_norm": 2.8488690853118896,
      "learning_rate": 5.640138408304498e-06,
      "loss": 0.1345,
      "step": 2730
    },
    {
      "epoch": 9.48,
      "grad_norm": 4.022710800170898,
      "learning_rate": 5.294117647058824e-06,
      "loss": 0.1466,
      "step": 2740
    },
    {
      "epoch": 9.52,
      "grad_norm": 1.0064830780029297,
      "learning_rate": 4.948096885813149e-06,
      "loss": 0.2983,
      "step": 2750
    },
    {
      "epoch": 9.55,
      "grad_norm": 2.4876391887664795,
      "learning_rate": 4.602076124567474e-06,
      "loss": 0.2516,
      "step": 2760
    },
    {
      "epoch": 9.58,
      "grad_norm": 8.971895217895508,
      "learning_rate": 4.2560553633217994e-06,
      "loss": 0.2621,
      "step": 2770
    },
    {
      "epoch": 9.62,
      "grad_norm": 6.835991859436035,
      "learning_rate": 3.910034602076125e-06,
      "loss": 0.1548,
      "step": 2780
    },
    {
      "epoch": 9.65,
      "grad_norm": 3.8968513011932373,
      "learning_rate": 3.5640138408304496e-06,
      "loss": 0.2505,
      "step": 2790
    },
    {
      "epoch": 9.69,
      "grad_norm": 2.595637083053589,
      "learning_rate": 3.2179930795847754e-06,
      "loss": 0.2144,
      "step": 2800
    },
    {
      "epoch": 9.69,
      "eval_accuracy": 0.8711656441717791,
      "eval_loss": 0.4744942784309387,
      "eval_runtime": 11.7899,
      "eval_samples_per_second": 69.127,
      "eval_steps_per_second": 8.651,
      "step": 2800
    },
    {
      "epoch": 9.72,
      "grad_norm": 8.595541954040527,
      "learning_rate": 2.8719723183391003e-06,
      "loss": 0.1468,
      "step": 2810
    },
    {
      "epoch": 9.76,
      "grad_norm": 3.5496714115142822,
      "learning_rate": 2.5259515570934256e-06,
      "loss": 0.2275,
      "step": 2820
    },
    {
      "epoch": 9.79,
      "grad_norm": 6.138185977935791,
      "learning_rate": 2.179930795847751e-06,
      "loss": 0.3254,
      "step": 2830
    },
    {
      "epoch": 9.83,
      "grad_norm": 1.1723095178604126,
      "learning_rate": 1.8339100346020762e-06,
      "loss": 0.109,
      "step": 2840
    },
    {
      "epoch": 9.86,
      "grad_norm": 5.896970272064209,
      "learning_rate": 1.4878892733564015e-06,
      "loss": 0.2334,
      "step": 2850
    },
    {
      "epoch": 9.9,
      "grad_norm": 2.945650577545166,
      "learning_rate": 1.1764705882352942e-06,
      "loss": 0.3131,
      "step": 2860
    },
    {
      "epoch": 9.93,
      "grad_norm": 5.7923583984375,
      "learning_rate": 8.304498269896194e-07,
      "loss": 0.3139,
      "step": 2870
    },
    {
      "epoch": 9.97,
      "grad_norm": 2.053051471710205,
      "learning_rate": 4.844290657439446e-07,
      "loss": 0.2643,
      "step": 2880
    },
    {
      "epoch": 10.0,
      "grad_norm": 0.1167406514286995,
      "learning_rate": 1.384083044982699e-07,
      "loss": 0.1631,
      "step": 2890
    },
    {
      "epoch": 10.0,
      "step": 2890,
      "total_flos": 3.57806169439359e+18,
      "train_loss": 0.4476649462557994,
      "train_runtime": 1596.8664,
      "train_samples_per_second": 28.913,
      "train_steps_per_second": 1.81
    }
  ],
  "logging_steps": 10,
  "max_steps": 2890,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 10,
  "save_steps": 100,
  "total_flos": 3.57806169439359e+18,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": null
}