{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 1.790861159929701,
  "eval_steps": 500,
  "global_step": 510,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.0351493848857645,
      "grad_norm": 6.278096259282107,
      "learning_rate": 5.625e-06,
      "loss": 0.9084915161132813,
      "step": 10
    },
    {
      "epoch": 0.070298769771529,
      "grad_norm": 3.6331372712446486,
      "learning_rate": 9.999090054896397e-06,
      "loss": 0.8169801712036133,
      "step": 20
    },
    {
      "epoch": 0.1054481546572935,
      "grad_norm": 2.9908653115492245,
      "learning_rate": 9.98292246503335e-06,
      "loss": 0.7732008934020996,
      "step": 30
    },
    {
      "epoch": 0.140597539543058,
      "grad_norm": 3.0042485647099246,
      "learning_rate": 9.946609117610508e-06,
      "loss": 0.7605951309204102,
      "step": 40
    },
    {
      "epoch": 0.1757469244288225,
      "grad_norm": 3.223296409918759,
      "learning_rate": 9.89029682604881e-06,
      "loss": 0.7571172714233398,
      "step": 50
    },
    {
      "epoch": 0.210896309314587,
      "grad_norm": 3.006856371599571,
      "learning_rate": 9.814213258705813e-06,
      "loss": 0.7587286949157714,
      "step": 60
    },
    {
      "epoch": 0.2460456942003515,
      "grad_norm": 2.831241538535402,
      "learning_rate": 9.71866601842156e-06,
      "loss": 0.7482702255249023,
      "step": 70
    },
    {
      "epoch": 0.281195079086116,
      "grad_norm": 2.952447234736118,
      "learning_rate": 9.604041398892528e-06,
      "loss": 0.7249799728393554,
      "step": 80
    },
    {
      "epoch": 0.3163444639718805,
      "grad_norm": 2.653656602471469,
      "learning_rate": 9.470802822901558e-06,
      "loss": 0.7315261840820313,
      "step": 90
    },
    {
      "epoch": 0.351493848857645,
      "grad_norm": 2.864053404566767,
      "learning_rate": 9.31948896871795e-06,
      "loss": 0.7015914440155029,
      "step": 100
    },
    {
      "epoch": 0.3866432337434095,
      "grad_norm": 2.3045514301049708,
      "learning_rate": 9.150711592242627e-06,
      "loss": 0.7025287628173829,
      "step": 110
    },
    {
      "epoch": 0.421792618629174,
      "grad_norm": 2.5545038781024387,
      "learning_rate": 8.965153053703325e-06,
      "loss": 0.7210351943969726,
      "step": 120
    },
    {
      "epoch": 0.45694200351493847,
      "grad_norm": 2.4553034267827023,
      "learning_rate": 8.763563558899317e-06,
      "loss": 0.6796262741088868,
      "step": 130
    },
    {
      "epoch": 0.492091388400703,
      "grad_norm": 2.221687697158951,
      "learning_rate": 8.546758126149148e-06,
      "loss": 0.6920151710510254,
      "step": 140
    },
    {
      "epoch": 0.5272407732864675,
      "grad_norm": 2.2066268668928903,
      "learning_rate": 8.315613291203977e-06,
      "loss": 0.7003077030181885,
      "step": 150
    },
    {
      "epoch": 0.562390158172232,
      "grad_norm": 2.1448081876793483,
      "learning_rate": 8.071063563448341e-06,
      "loss": 0.7108040809631347,
      "step": 160
    },
    {
      "epoch": 0.5975395430579965,
      "grad_norm": 2.1756459600523694,
      "learning_rate": 7.814097647715848e-06,
      "loss": 0.7066202163696289,
      "step": 170
    },
    {
      "epoch": 0.632688927943761,
      "grad_norm": 2.629165746175461,
      "learning_rate": 7.5457544469948164e-06,
      "loss": 0.6910131454467774,
      "step": 180
    },
    {
      "epoch": 0.6678383128295254,
      "grad_norm": 2.3229845016911375,
      "learning_rate": 7.267118862184767e-06,
      "loss": 0.6542898654937744,
      "step": 190
    },
    {
      "epoch": 0.70298769771529,
      "grad_norm": 2.1923869810324756,
      "learning_rate": 6.9793174058851805e-06,
      "loss": 0.693804931640625,
      "step": 200
    },
    {
      "epoch": 0.7381370826010545,
      "grad_norm": 2.1367751639035713,
      "learning_rate": 6.683513647949826e-06,
      "loss": 0.7076559066772461,
      "step": 210
    },
    {
      "epoch": 0.773286467486819,
      "grad_norm": 2.0456065016164997,
      "learning_rate": 6.38090351122006e-06,
      "loss": 0.6664181709289551,
      "step": 220
    },
    {
      "epoch": 0.8084358523725835,
      "grad_norm": 2.123407040948006,
      "learning_rate": 6.072710436456293e-06,
      "loss": 0.7038721084594727,
      "step": 230
    },
    {
      "epoch": 0.843585237258348,
      "grad_norm": 2.1497633512403014,
      "learning_rate": 5.760180436015604e-06,
      "loss": 0.6690019607543946,
      "step": 240
    },
    {
      "epoch": 0.8787346221441125,
      "grad_norm": 2.2002946157420222,
      "learning_rate": 5.444577056273284e-06,
      "loss": 0.7084089279174804,
      "step": 250
    },
    {
      "epoch": 0.9138840070298769,
      "grad_norm": 2.424827713866129,
      "learning_rate": 5.1271762691550375e-06,
      "loss": 0.6761907577514649,
      "step": 260
    },
    {
      "epoch": 0.9490333919156415,
      "grad_norm": 2.224520502627956,
      "learning_rate": 4.809261313433176e-06,
      "loss": 0.6852016448974609,
      "step": 270
    },
    {
      "epoch": 0.984182776801406,
      "grad_norm": 1.8445152026218052,
      "learning_rate": 4.492117506643256e-06,
      "loss": 0.6881031036376953,
      "step": 280
    },
    {
      "epoch": 1.0175746924428823,
      "grad_norm": 1.8230936110227824,
      "learning_rate": 4.17702704859633e-06,
      "loss": 0.6024819374084472,
      "step": 290
    },
    {
      "epoch": 1.0527240773286468,
      "grad_norm": 1.9744350878935748,
      "learning_rate": 3.865263837496072e-06,
      "loss": 0.5643129348754883,
      "step": 300
    },
    {
      "epoch": 1.0878734622144113,
      "grad_norm": 1.9618138490487484,
      "learning_rate": 3.5580883196189265e-06,
      "loss": 0.5441983699798584,
      "step": 310
    },
    {
      "epoch": 1.1230228471001757,
      "grad_norm": 1.9115536703867921,
      "learning_rate": 3.256742393379909e-06,
      "loss": 0.5266704559326172,
      "step": 320
    },
    {
      "epoch": 1.1581722319859402,
      "grad_norm": 2.1347500782458733,
      "learning_rate": 2.9624443883866403e-06,
      "loss": 0.5213542938232422,
      "step": 330
    },
    {
      "epoch": 1.1933216168717047,
      "grad_norm": 1.851494415342933,
      "learning_rate": 2.6763841397811576e-06,
      "loss": 0.5169600486755371,
      "step": 340
    },
    {
      "epoch": 1.2284710017574691,
      "grad_norm": 2.2004480026852096,
      "learning_rate": 2.3997181777836955e-06,
      "loss": 0.5154135704040528,
      "step": 350
    },
    {
      "epoch": 1.2636203866432338,
      "grad_norm": 2.064960097474409,
      "learning_rate": 2.1335650518869555e-06,
      "loss": 0.5326839447021484,
      "step": 360
    },
    {
      "epoch": 1.2987697715289983,
      "grad_norm": 2.2791411946769733,
      "learning_rate": 1.8790008086049534e-06,
      "loss": 0.5299165725708008,
      "step": 370
    },
    {
      "epoch": 1.3339191564147628,
      "grad_norm": 2.1747163997571644,
      "learning_rate": 1.6370546410597066e-06,
      "loss": 0.5298036098480224,
      "step": 380
    },
    {
      "epoch": 1.3690685413005272,
      "grad_norm": 2.1670480068866476,
      "learning_rate": 1.4087047279943267e-06,
      "loss": 0.5323153972625733,
      "step": 390
    },
    {
      "epoch": 1.4042179261862917,
      "grad_norm": 2.178341675043311,
      "learning_rate": 1.1948742790352342e-06,
      "loss": 0.5313343048095703,
      "step": 400
    },
    {
      "epoch": 1.4393673110720562,
      "grad_norm": 2.2167608532468885,
      "learning_rate": 9.964278021923107e-07,
      "loss": 0.5045088768005371,
      "step": 410
    },
    {
      "epoch": 1.4745166959578206,
      "grad_norm": 2.1066242945859392,
      "learning_rate": 8.141676086873574e-07,
      "loss": 0.5003239631652832,
      "step": 420
    },
    {
      "epoch": 1.5096660808435853,
      "grad_norm": 2.1073700284966517,
      "learning_rate": 6.488305692417074e-07,
      "loss": 0.5294024467468261,
      "step": 430
    },
    {
      "epoch": 1.5448154657293496,
      "grad_norm": 2.004653967796236,
      "learning_rate": 5.010851349371704e-07,
      "loss": 0.5233042240142822,
      "step": 440
    },
    {
      "epoch": 1.5799648506151143,
      "grad_norm": 2.2332707746551135,
      "learning_rate": 3.715286346948671e-07,
      "loss": 0.49408349990844724,
      "step": 450
    },
    {
      "epoch": 1.6151142355008787,
      "grad_norm": 2.191790920883347,
      "learning_rate": 2.6068486029813154e-07,
      "loss": 0.5186878204345703,
      "step": 460
    },
    {
      "epoch": 1.6502636203866432,
      "grad_norm": 2.090930593332735,
      "learning_rate": 1.69001948723142e-07,
      "loss": 0.4858062744140625,
      "step": 470
    },
    {
      "epoch": 1.685413005272408,
      "grad_norm": 2.1621725110130487,
      "learning_rate": 9.685057033892998e-08,
      "loss": 0.5107901096343994,
      "step": 480
    },
    {
      "epoch": 1.7205623901581721,
      "grad_norm": 1.8955043517645798,
      "learning_rate": 4.45224303018027e-08,
      "loss": 0.5104897499084473,
      "step": 490
    },
    {
      "epoch": 1.7557117750439368,
      "grad_norm": 2.2608050891582776,
      "learning_rate": 1.2229089202987487e-08,
      "loss": 0.5028838157653809,
      "step": 500
    },
    {
      "epoch": 1.790861159929701,
      "grad_norm": 2.217597617676314,
      "learning_rate": 1.011077375662195e-10,
      "loss": 0.5305490970611573,
      "step": 510
    },
    {
      "epoch": 1.790861159929701,
      "step": 510,
      "total_flos": 60959187533824.0,
      "train_loss": 0.6313069558611103,
      "train_runtime": 1292.1645,
      "train_samples_per_second": 6.315,
      "train_steps_per_second": 0.395
    }
  ],
  "logging_steps": 10,
  "max_steps": 510,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 2,
  "save_steps": 170,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 60959187533824.0,
  "train_batch_size": 1,
  "trial_name": null,
  "trial_params": null
}