{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 5.0,
  "eval_steps": 20,
  "global_step": 3160,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.015822784810126583,
      "grad_norm": 0.9025397896766663,
      "learning_rate": 0.0002,
      "loss": 1.7955,
      "step": 10
    },
    {
      "epoch": 0.03164556962025317,
      "grad_norm": 0.6101710200309753,
      "learning_rate": 0.0002,
      "loss": 1.2957,
      "step": 20
    },
    {
      "epoch": 0.03164556962025317,
      "eval_loss": 1.1671996116638184,
      "eval_runtime": 131.8343,
      "eval_samples_per_second": 4.794,
      "eval_steps_per_second": 0.599,
      "step": 20
    },
    {
      "epoch": 0.04746835443037975,
      "grad_norm": 0.261136531829834,
      "learning_rate": 0.0002,
      "loss": 1.1238,
      "step": 30
    },
    {
      "epoch": 0.06329113924050633,
      "grad_norm": 0.2999696731567383,
      "learning_rate": 0.0002,
      "loss": 1.0512,
      "step": 40
    },
    {
      "epoch": 0.06329113924050633,
      "eval_loss": 0.9885051846504211,
      "eval_runtime": 131.9371,
      "eval_samples_per_second": 4.79,
      "eval_steps_per_second": 0.599,
      "step": 40
    },
    {
      "epoch": 0.07911392405063292,
      "grad_norm": 0.2779262959957123,
      "learning_rate": 0.0002,
      "loss": 0.9873,
      "step": 50
    },
    {
      "epoch": 0.0949367088607595,
      "grad_norm": 0.28684648871421814,
      "learning_rate": 0.0002,
      "loss": 0.9509,
      "step": 60
    },
    {
      "epoch": 0.0949367088607595,
      "eval_loss": 0.9145669341087341,
      "eval_runtime": 131.853,
      "eval_samples_per_second": 4.793,
      "eval_steps_per_second": 0.599,
      "step": 60
    },
    {
      "epoch": 0.11075949367088607,
      "grad_norm": 0.2835591435432434,
      "learning_rate": 0.0002,
      "loss": 0.9087,
      "step": 70
    },
    {
      "epoch": 0.12658227848101267,
      "grad_norm": 0.3226328194141388,
      "learning_rate": 0.0002,
      "loss": 0.8869,
      "step": 80
    },
    {
      "epoch": 0.12658227848101267,
      "eval_loss": 0.8785317540168762,
      "eval_runtime": 131.9051,
      "eval_samples_per_second": 4.791,
      "eval_steps_per_second": 0.599,
      "step": 80
    },
    {
      "epoch": 0.14240506329113925,
      "grad_norm": 0.2908966839313507,
      "learning_rate": 0.0002,
      "loss": 0.8778,
      "step": 90
    },
    {
      "epoch": 0.15822784810126583,
      "grad_norm": 0.3126741349697113,
      "learning_rate": 0.0002,
      "loss": 0.8783,
      "step": 100
    },
    {
      "epoch": 0.15822784810126583,
      "eval_loss": 0.8529007434844971,
      "eval_runtime": 131.9026,
      "eval_samples_per_second": 4.791,
      "eval_steps_per_second": 0.599,
      "step": 100
    },
    {
      "epoch": 0.17405063291139242,
      "grad_norm": 0.29166632890701294,
      "learning_rate": 0.0002,
      "loss": 0.861,
      "step": 110
    },
    {
      "epoch": 0.189873417721519,
      "grad_norm": 0.2920988202095032,
      "learning_rate": 0.0002,
      "loss": 0.8409,
      "step": 120
    },
    {
      "epoch": 0.189873417721519,
      "eval_loss": 0.8351882696151733,
      "eval_runtime": 131.9306,
      "eval_samples_per_second": 4.79,
      "eval_steps_per_second": 0.599,
      "step": 120
    },
    {
      "epoch": 0.20569620253164558,
      "grad_norm": 0.2958802878856659,
      "learning_rate": 0.0002,
      "loss": 0.8452,
      "step": 130
    },
    {
      "epoch": 0.22151898734177214,
      "grad_norm": 0.3161458969116211,
      "learning_rate": 0.0002,
      "loss": 0.813,
      "step": 140
    },
    {
      "epoch": 0.22151898734177214,
      "eval_loss": 0.8209991455078125,
      "eval_runtime": 131.909,
      "eval_samples_per_second": 4.791,
      "eval_steps_per_second": 0.599,
      "step": 140
    },
    {
      "epoch": 0.23734177215189872,
      "grad_norm": 0.30558162927627563,
      "learning_rate": 0.0002,
      "loss": 0.802,
      "step": 150
    },
    {
      "epoch": 0.25316455696202533,
      "grad_norm": 0.3051740825176239,
      "learning_rate": 0.0002,
      "loss": 0.8221,
      "step": 160
    },
    {
      "epoch": 0.25316455696202533,
      "eval_loss": 0.8096217513084412,
      "eval_runtime": 131.8619,
      "eval_samples_per_second": 4.793,
      "eval_steps_per_second": 0.599,
      "step": 160
    },
    {
      "epoch": 0.2689873417721519,
      "grad_norm": 0.2719588577747345,
      "learning_rate": 0.0002,
      "loss": 0.8163,
      "step": 170
    },
    {
      "epoch": 0.2848101265822785,
      "grad_norm": 0.26874399185180664,
      "learning_rate": 0.0002,
      "loss": 0.8202,
      "step": 180
    },
    {
      "epoch": 0.2848101265822785,
      "eval_loss": 0.7993021011352539,
      "eval_runtime": 131.9133,
      "eval_samples_per_second": 4.791,
      "eval_steps_per_second": 0.599,
      "step": 180
    },
    {
      "epoch": 0.30063291139240506,
      "grad_norm": 0.29227378964424133,
      "learning_rate": 0.0002,
      "loss": 0.8266,
      "step": 190
    },
    {
      "epoch": 0.31645569620253167,
      "grad_norm": 0.2868894934654236,
      "learning_rate": 0.0002,
      "loss": 0.8003,
      "step": 200
    },
    {
      "epoch": 0.31645569620253167,
      "eval_loss": 0.7905535101890564,
      "eval_runtime": 131.9228,
      "eval_samples_per_second": 4.791,
      "eval_steps_per_second": 0.599,
      "step": 200
    },
    {
      "epoch": 0.3322784810126582,
      "grad_norm": 0.28897446393966675,
      "learning_rate": 0.0002,
      "loss": 0.7938,
      "step": 210
    },
    {
      "epoch": 0.34810126582278483,
      "grad_norm": 0.28241968154907227,
      "learning_rate": 0.0002,
      "loss": 0.7995,
      "step": 220
    },
    {
      "epoch": 0.34810126582278483,
      "eval_loss": 0.7820630073547363,
      "eval_runtime": 131.8475,
      "eval_samples_per_second": 4.793,
      "eval_steps_per_second": 0.599,
      "step": 220
    },
    {
      "epoch": 0.3639240506329114,
      "grad_norm": 0.2948983907699585,
      "learning_rate": 0.0002,
      "loss": 0.7737,
      "step": 230
    },
    {
      "epoch": 0.379746835443038,
      "grad_norm": 0.29324445128440857,
      "learning_rate": 0.0002,
      "loss": 0.7757,
      "step": 240
    },
    {
      "epoch": 0.379746835443038,
      "eval_loss": 0.7757007479667664,
      "eval_runtime": 131.8815,
      "eval_samples_per_second": 4.792,
      "eval_steps_per_second": 0.599,
      "step": 240
    },
    {
      "epoch": 0.39556962025316456,
      "grad_norm": 0.2765977680683136,
      "learning_rate": 0.0002,
      "loss": 0.782,
      "step": 250
    },
    {
      "epoch": 0.41139240506329117,
      "grad_norm": 0.28577566146850586,
      "learning_rate": 0.0002,
      "loss": 0.79,
      "step": 260
    },
    {
      "epoch": 0.41139240506329117,
      "eval_loss": 0.7689570188522339,
      "eval_runtime": 131.8449,
      "eval_samples_per_second": 4.794,
      "eval_steps_per_second": 0.599,
      "step": 260
    },
    {
      "epoch": 0.4272151898734177,
      "grad_norm": 0.28461453318595886,
      "learning_rate": 0.0002,
      "loss": 0.7711,
      "step": 270
    },
    {
      "epoch": 0.4430379746835443,
      "grad_norm": 0.28736287355422974,
      "learning_rate": 0.0002,
      "loss": 0.7785,
      "step": 280
    },
    {
      "epoch": 0.4430379746835443,
      "eval_loss": 0.7642198204994202,
      "eval_runtime": 131.8434,
      "eval_samples_per_second": 4.794,
      "eval_steps_per_second": 0.599,
      "step": 280
    },
    {
      "epoch": 0.4588607594936709,
      "grad_norm": 0.27680790424346924,
      "learning_rate": 0.0002,
      "loss": 0.7647,
      "step": 290
    },
    {
      "epoch": 0.47468354430379744,
      "grad_norm": 0.2566572427749634,
      "learning_rate": 0.0002,
      "loss": 0.7826,
      "step": 300
    },
    {
      "epoch": 0.47468354430379744,
      "eval_loss": 0.758773922920227,
      "eval_runtime": 131.8642,
      "eval_samples_per_second": 4.793,
      "eval_steps_per_second": 0.599,
      "step": 300
    },
    {
      "epoch": 0.49050632911392406,
      "grad_norm": 0.2631087899208069,
      "learning_rate": 0.0002,
      "loss": 0.7637,
      "step": 310
    },
    {
      "epoch": 0.5063291139240507,
      "grad_norm": 0.2839374542236328,
      "learning_rate": 0.0002,
      "loss": 0.7842,
      "step": 320
    },
    {
      "epoch": 0.5063291139240507,
      "eval_loss": 0.7543513774871826,
      "eval_runtime": 131.8699,
      "eval_samples_per_second": 4.793,
      "eval_steps_per_second": 0.599,
      "step": 320
    },
    {
      "epoch": 0.5221518987341772,
      "grad_norm": 0.2678869068622589,
      "learning_rate": 0.0002,
      "loss": 0.7483,
      "step": 330
    },
    {
      "epoch": 0.5379746835443038,
      "grad_norm": 0.2868534028530121,
      "learning_rate": 0.0002,
      "loss": 0.7548,
      "step": 340
    },
    {
      "epoch": 0.5379746835443038,
      "eval_loss": 0.7495388984680176,
      "eval_runtime": 131.8745,
      "eval_samples_per_second": 4.792,
      "eval_steps_per_second": 0.599,
      "step": 340
    },
    {
      "epoch": 0.5537974683544303,
      "grad_norm": 0.29033464193344116,
      "learning_rate": 0.0002,
      "loss": 0.7689,
      "step": 350
    },
    {
      "epoch": 0.569620253164557,
      "grad_norm": 0.2854663133621216,
      "learning_rate": 0.0002,
      "loss": 0.7538,
      "step": 360
    },
    {
      "epoch": 0.569620253164557,
      "eval_loss": 0.7454854846000671,
      "eval_runtime": 131.8755,
      "eval_samples_per_second": 4.792,
      "eval_steps_per_second": 0.599,
      "step": 360
    },
    {
      "epoch": 0.5854430379746836,
      "grad_norm": 0.26429274678230286,
      "learning_rate": 0.0002,
      "loss": 0.7435,
      "step": 370
    },
    {
      "epoch": 0.6012658227848101,
      "grad_norm": 0.2575242817401886,
      "learning_rate": 0.0002,
      "loss": 0.7767,
      "step": 380
    },
    {
      "epoch": 0.6012658227848101,
      "eval_loss": 0.7419107556343079,
      "eval_runtime": 131.8915,
      "eval_samples_per_second": 4.792,
      "eval_steps_per_second": 0.599,
      "step": 380
    },
    {
      "epoch": 0.6170886075949367,
      "grad_norm": 0.27361172437667847,
      "learning_rate": 0.0002,
      "loss": 0.761,
      "step": 390
    },
    {
      "epoch": 0.6329113924050633,
      "grad_norm": 0.270486980676651,
      "learning_rate": 0.0002,
      "loss": 0.7571,
      "step": 400
    },
    {
      "epoch": 0.6329113924050633,
      "eval_loss": 0.7377169132232666,
      "eval_runtime": 131.8585,
      "eval_samples_per_second": 4.793,
      "eval_steps_per_second": 0.599,
      "step": 400
    },
    {
      "epoch": 0.6487341772151899,
      "grad_norm": 0.2674134075641632,
      "learning_rate": 0.0002,
      "loss": 0.7528,
      "step": 410
    },
    {
      "epoch": 0.6645569620253164,
      "grad_norm": 0.27510151267051697,
      "learning_rate": 0.0002,
      "loss": 0.731,
      "step": 420
    },
    {
      "epoch": 0.6645569620253164,
      "eval_loss": 0.7345684170722961,
      "eval_runtime": 131.8396,
      "eval_samples_per_second": 4.794,
      "eval_steps_per_second": 0.599,
      "step": 420
    },
    {
      "epoch": 0.680379746835443,
      "grad_norm": 0.2929084599018097,
      "learning_rate": 0.0002,
      "loss": 0.7448,
      "step": 430
    },
    {
      "epoch": 0.6962025316455697,
      "grad_norm": 0.2811911106109619,
      "learning_rate": 0.0002,
      "loss": 0.7395,
      "step": 440
    },
    {
      "epoch": 0.6962025316455697,
      "eval_loss": 0.7302244901657104,
      "eval_runtime": 131.824,
      "eval_samples_per_second": 4.794,
      "eval_steps_per_second": 0.599,
      "step": 440
    },
    {
      "epoch": 0.7120253164556962,
      "grad_norm": 0.2596493661403656,
      "learning_rate": 0.0002,
      "loss": 0.7314,
      "step": 450
    },
    {
      "epoch": 0.7278481012658228,
      "grad_norm": 0.273561030626297,
      "learning_rate": 0.0002,
      "loss": 0.7189,
      "step": 460
    },
    {
      "epoch": 0.7278481012658228,
      "eval_loss": 0.7286897301673889,
      "eval_runtime": 131.8949,
      "eval_samples_per_second": 4.792,
      "eval_steps_per_second": 0.599,
      "step": 460
    },
    {
      "epoch": 0.7436708860759493,
      "grad_norm": 0.2678873538970947,
      "learning_rate": 0.0002,
      "loss": 0.7144,
      "step": 470
    },
    {
      "epoch": 0.759493670886076,
      "grad_norm": 0.28556978702545166,
      "learning_rate": 0.0002,
      "loss": 0.7284,
      "step": 480
    },
    {
      "epoch": 0.759493670886076,
      "eval_loss": 0.7248781323432922,
      "eval_runtime": 131.8044,
      "eval_samples_per_second": 4.795,
      "eval_steps_per_second": 0.599,
      "step": 480
    },
    {
      "epoch": 0.7753164556962026,
      "grad_norm": 0.30613693594932556,
      "learning_rate": 0.0002,
      "loss": 0.7556,
      "step": 490
    },
    {
      "epoch": 0.7911392405063291,
      "grad_norm": 0.27024537324905396,
      "learning_rate": 0.0002,
      "loss": 0.7395,
      "step": 500
    },
    {
      "epoch": 0.7911392405063291,
      "eval_loss": 0.7223949432373047,
      "eval_runtime": 131.9503,
      "eval_samples_per_second": 4.79,
      "eval_steps_per_second": 0.599,
      "step": 500
    },
    {
      "epoch": 0.8069620253164557,
      "grad_norm": 0.26208561658859253,
      "learning_rate": 0.0002,
      "loss": 0.7265,
      "step": 510
    },
    {
      "epoch": 0.8227848101265823,
      "grad_norm": 0.259666383266449,
      "learning_rate": 0.0002,
      "loss": 0.71,
      "step": 520
    },
    {
      "epoch": 0.8227848101265823,
      "eval_loss": 0.7190249562263489,
      "eval_runtime": 131.8787,
      "eval_samples_per_second": 4.792,
      "eval_steps_per_second": 0.599,
      "step": 520
    },
    {
      "epoch": 0.8386075949367089,
      "grad_norm": 0.2681174576282501,
      "learning_rate": 0.0002,
      "loss": 0.7422,
      "step": 530
    },
    {
      "epoch": 0.8544303797468354,
      "grad_norm": 0.26732560992240906,
      "learning_rate": 0.0002,
      "loss": 0.7269,
      "step": 540
    },
    {
      "epoch": 0.8544303797468354,
      "eval_loss": 0.7150493860244751,
      "eval_runtime": 131.8437,
      "eval_samples_per_second": 4.794,
      "eval_steps_per_second": 0.599,
      "step": 540
    },
    {
      "epoch": 0.870253164556962,
      "grad_norm": 0.24266886711120605,
      "learning_rate": 0.0002,
      "loss": 0.7102,
      "step": 550
    },
    {
      "epoch": 0.8860759493670886,
      "grad_norm": 0.25400373339653015,
      "learning_rate": 0.0002,
      "loss": 0.7219,
      "step": 560
    },
    {
      "epoch": 0.8860759493670886,
      "eval_loss": 0.7140419483184814,
      "eval_runtime": 131.777,
      "eval_samples_per_second": 4.796,
      "eval_steps_per_second": 0.599,
      "step": 560
    },
    {
      "epoch": 0.9018987341772152,
      "grad_norm": 0.26475510001182556,
      "learning_rate": 0.0002,
      "loss": 0.7116,
      "step": 570
    },
    {
      "epoch": 0.9177215189873418,
      "grad_norm": 0.2724270820617676,
      "learning_rate": 0.0002,
      "loss": 0.7326,
      "step": 580
    },
    {
      "epoch": 0.9177215189873418,
      "eval_loss": 0.7112420797348022,
      "eval_runtime": 131.8663,
      "eval_samples_per_second": 4.793,
      "eval_steps_per_second": 0.599,
      "step": 580
    },
    {
      "epoch": 0.9335443037974683,
      "grad_norm": 0.25785592198371887,
      "learning_rate": 0.0002,
      "loss": 0.7062,
      "step": 590
    },
    {
      "epoch": 0.9493670886075949,
      "grad_norm": 0.25584301352500916,
      "learning_rate": 0.0002,
      "loss": 0.7377,
      "step": 600
    },
    {
      "epoch": 0.9493670886075949,
      "eval_loss": 0.7086200714111328,
      "eval_runtime": 131.8628,
      "eval_samples_per_second": 4.793,
      "eval_steps_per_second": 0.599,
      "step": 600
    },
    {
      "epoch": 0.9651898734177216,
      "grad_norm": 0.2493722140789032,
      "learning_rate": 0.0002,
      "loss": 0.7257,
      "step": 610
    },
    {
      "epoch": 0.9810126582278481,
      "grad_norm": 0.2619902789592743,
      "learning_rate": 0.0002,
      "loss": 0.7002,
      "step": 620
    },
    {
      "epoch": 0.9810126582278481,
      "eval_loss": 0.7058697938919067,
      "eval_runtime": 131.8432,
      "eval_samples_per_second": 4.794,
      "eval_steps_per_second": 0.599,
      "step": 620
    },
    {
      "epoch": 0.9968354430379747,
      "grad_norm": 0.26538994908332825,
      "learning_rate": 0.0002,
      "loss": 0.696,
      "step": 630
    },
    {
      "epoch": 1.0126582278481013,
      "grad_norm": 0.24350906908512115,
      "learning_rate": 0.0002,
      "loss": 0.6878,
      "step": 640
    },
    {
      "epoch": 1.0126582278481013,
      "eval_loss": 0.7054635882377625,
      "eval_runtime": 131.871,
      "eval_samples_per_second": 4.793,
      "eval_steps_per_second": 0.599,
      "step": 640
    },
    {
      "epoch": 1.0284810126582278,
      "grad_norm": 0.27845484018325806,
      "learning_rate": 0.0002,
      "loss": 0.6912,
      "step": 650
    },
    {
      "epoch": 1.0443037974683544,
      "grad_norm": 0.2543031871318817,
      "learning_rate": 0.0002,
      "loss": 0.6853,
      "step": 660
    },
    {
      "epoch": 1.0443037974683544,
      "eval_loss": 0.7026867270469666,
      "eval_runtime": 131.9203,
      "eval_samples_per_second": 4.791,
      "eval_steps_per_second": 0.599,
      "step": 660
    },
    {
      "epoch": 1.0601265822784811,
      "grad_norm": 0.2510567009449005,
      "learning_rate": 0.0002,
      "loss": 0.6854,
      "step": 670
    },
    {
      "epoch": 1.0759493670886076,
      "grad_norm": 0.2633340060710907,
      "learning_rate": 0.0002,
      "loss": 0.6873,
      "step": 680
    },
    {
      "epoch": 1.0759493670886076,
      "eval_loss": 0.700657308101654,
      "eval_runtime": 131.9117,
      "eval_samples_per_second": 4.791,
      "eval_steps_per_second": 0.599,
      "step": 680
    },
    {
      "epoch": 1.0917721518987342,
      "grad_norm": 0.2701621353626251,
      "learning_rate": 0.0002,
      "loss": 0.692,
      "step": 690
    },
    {
      "epoch": 1.1075949367088607,
      "grad_norm": 0.25083285570144653,
      "learning_rate": 0.0002,
      "loss": 0.6711,
      "step": 700
    },
    {
      "epoch": 1.1075949367088607,
      "eval_loss": 0.6996587514877319,
      "eval_runtime": 131.8839,
      "eval_samples_per_second": 4.792,
      "eval_steps_per_second": 0.599,
      "step": 700
    },
    {
      "epoch": 1.1234177215189873,
      "grad_norm": 0.24586257338523865,
      "learning_rate": 0.0002,
      "loss": 0.671,
      "step": 710
    },
    {
      "epoch": 1.139240506329114,
      "grad_norm": 0.2607501447200775,
      "learning_rate": 0.0002,
      "loss": 0.6943,
      "step": 720
    },
    {
      "epoch": 1.139240506329114,
      "eval_loss": 0.6975678205490112,
      "eval_runtime": 131.896,
      "eval_samples_per_second": 4.792,
      "eval_steps_per_second": 0.599,
      "step": 720
    },
    {
      "epoch": 1.1550632911392404,
      "grad_norm": 0.27073556184768677,
      "learning_rate": 0.0002,
      "loss": 0.6771,
      "step": 730
    },
    {
      "epoch": 1.1708860759493671,
      "grad_norm": 0.26840293407440186,
      "learning_rate": 0.0002,
      "loss": 0.6884,
      "step": 740
    },
    {
      "epoch": 1.1708860759493671,
      "eval_loss": 0.6958794593811035,
      "eval_runtime": 131.952,
      "eval_samples_per_second": 4.79,
      "eval_steps_per_second": 0.599,
      "step": 740
    },
    {
      "epoch": 1.1867088607594938,
      "grad_norm": 0.24738739430904388,
      "learning_rate": 0.0002,
      "loss": 0.6788,
      "step": 750
    },
    {
      "epoch": 1.2025316455696202,
      "grad_norm": 0.2699371874332428,
      "learning_rate": 0.0002,
      "loss": 0.6937,
      "step": 760
    },
    {
      "epoch": 1.2025316455696202,
      "eval_loss": 0.6941386461257935,
      "eval_runtime": 131.9073,
      "eval_samples_per_second": 4.791,
      "eval_steps_per_second": 0.599,
      "step": 760
    },
    {
      "epoch": 1.2183544303797469,
      "grad_norm": 0.27836862206459045,
      "learning_rate": 0.0002,
      "loss": 0.6671,
      "step": 770
    },
    {
      "epoch": 1.2341772151898733,
      "grad_norm": 0.2613048553466797,
      "learning_rate": 0.0002,
      "loss": 0.6707,
      "step": 780
    },
    {
      "epoch": 1.2341772151898733,
      "eval_loss": 0.6920880675315857,
      "eval_runtime": 131.8796,
      "eval_samples_per_second": 4.792,
      "eval_steps_per_second": 0.599,
      "step": 780
    },
    {
      "epoch": 1.25,
      "grad_norm": 0.27010253071784973,
      "learning_rate": 0.0002,
      "loss": 0.6812,
      "step": 790
    },
    {
      "epoch": 1.2658227848101267,
      "grad_norm": 0.26843827962875366,
      "learning_rate": 0.0002,
      "loss": 0.6806,
      "step": 800
    },
    {
      "epoch": 1.2658227848101267,
      "eval_loss": 0.6902583241462708,
      "eval_runtime": 131.9241,
      "eval_samples_per_second": 4.791,
      "eval_steps_per_second": 0.599,
      "step": 800
    },
    {
      "epoch": 1.2816455696202531,
      "grad_norm": 0.2528655230998993,
      "learning_rate": 0.0002,
      "loss": 0.6741,
      "step": 810
    },
    {
      "epoch": 1.2974683544303798,
      "grad_norm": 0.2584218978881836,
      "learning_rate": 0.0002,
      "loss": 0.6743,
      "step": 820
    },
    {
      "epoch": 1.2974683544303798,
      "eval_loss": 0.6891089081764221,
      "eval_runtime": 131.8965,
      "eval_samples_per_second": 4.792,
      "eval_steps_per_second": 0.599,
      "step": 820
    },
    {
      "epoch": 1.3132911392405062,
      "grad_norm": 0.2561744153499603,
      "learning_rate": 0.0002,
      "loss": 0.6854,
      "step": 830
    },
    {
      "epoch": 1.3291139240506329,
      "grad_norm": 0.25455209612846375,
      "learning_rate": 0.0002,
      "loss": 0.6724,
      "step": 840
    },
    {
      "epoch": 1.3291139240506329,
      "eval_loss": 0.6864309906959534,
      "eval_runtime": 131.9472,
      "eval_samples_per_second": 4.79,
      "eval_steps_per_second": 0.599,
      "step": 840
    },
    {
      "epoch": 1.3449367088607596,
      "grad_norm": 0.26583439111709595,
      "learning_rate": 0.0002,
      "loss": 0.6801,
      "step": 850
    },
    {
      "epoch": 1.360759493670886,
      "grad_norm": 0.25029808282852173,
      "learning_rate": 0.0002,
      "loss": 0.6853,
      "step": 860
    },
    {
      "epoch": 1.360759493670886,
      "eval_loss": 0.6851058006286621,
      "eval_runtime": 131.9451,
      "eval_samples_per_second": 4.79,
      "eval_steps_per_second": 0.599,
      "step": 860
    },
    {
      "epoch": 1.3765822784810127,
      "grad_norm": 0.24356305599212646,
      "learning_rate": 0.0002,
      "loss": 0.6746,
      "step": 870
    },
    {
      "epoch": 1.3924050632911391,
      "grad_norm": 0.25391361117362976,
      "learning_rate": 0.0002,
      "loss": 0.6764,
      "step": 880
    },
    {
      "epoch": 1.3924050632911391,
      "eval_loss": 0.6833056807518005,
      "eval_runtime": 131.8759,
      "eval_samples_per_second": 4.792,
      "eval_steps_per_second": 0.599,
      "step": 880
    },
    {
      "epoch": 1.4082278481012658,
      "grad_norm": 0.25311189889907837,
      "learning_rate": 0.0002,
      "loss": 0.6638,
      "step": 890
    },
    {
      "epoch": 1.4240506329113924,
      "grad_norm": 0.25085288286209106,
      "learning_rate": 0.0002,
      "loss": 0.6634,
      "step": 900
    },
    {
      "epoch": 1.4240506329113924,
      "eval_loss": 0.6816738247871399,
      "eval_runtime": 131.9001,
      "eval_samples_per_second": 4.792,
      "eval_steps_per_second": 0.599,
      "step": 900
    },
    {
      "epoch": 1.439873417721519,
      "grad_norm": 0.25155821442604065,
      "learning_rate": 0.0002,
      "loss": 0.6776,
      "step": 910
    },
    {
      "epoch": 1.4556962025316456,
      "grad_norm": 0.24532318115234375,
      "learning_rate": 0.0002,
      "loss": 0.6724,
      "step": 920
    },
    {
      "epoch": 1.4556962025316456,
      "eval_loss": 0.6804719567298889,
      "eval_runtime": 131.8692,
      "eval_samples_per_second": 4.793,
      "eval_steps_per_second": 0.599,
      "step": 920
    },
    {
      "epoch": 1.4715189873417722,
      "grad_norm": 0.25289642810821533,
      "learning_rate": 0.0002,
      "loss": 0.6702,
      "step": 930
    },
    {
      "epoch": 1.4873417721518987,
      "grad_norm": 0.24714156985282898,
      "learning_rate": 0.0002,
      "loss": 0.681,
      "step": 940
    },
    {
      "epoch": 1.4873417721518987,
      "eval_loss": 0.6781874895095825,
      "eval_runtime": 131.9219,
      "eval_samples_per_second": 4.791,
      "eval_steps_per_second": 0.599,
      "step": 940
    },
    {
      "epoch": 1.5031645569620253,
      "grad_norm": 0.2605419158935547,
      "learning_rate": 0.0002,
      "loss": 0.6473,
      "step": 950
    },
    {
      "epoch": 1.518987341772152,
      "grad_norm": 0.24513210356235504,
      "learning_rate": 0.0002,
      "loss": 0.6576,
      "step": 960
    },
    {
      "epoch": 1.518987341772152,
      "eval_loss": 0.6770685911178589,
      "eval_runtime": 131.8952,
      "eval_samples_per_second": 4.792,
      "eval_steps_per_second": 0.599,
      "step": 960
    },
    {
      "epoch": 1.5348101265822784,
      "grad_norm": 0.2506636679172516,
      "learning_rate": 0.0002,
      "loss": 0.6685,
      "step": 970
    },
    {
      "epoch": 1.5506329113924051,
      "grad_norm": 0.24845752120018005,
      "learning_rate": 0.0002,
      "loss": 0.6727,
      "step": 980
    },
    {
      "epoch": 1.5506329113924051,
      "eval_loss": 0.6757538914680481,
      "eval_runtime": 131.9095,
      "eval_samples_per_second": 4.791,
      "eval_steps_per_second": 0.599,
      "step": 980
    },
    {
      "epoch": 1.5664556962025316,
      "grad_norm": 0.2520063817501068,
      "learning_rate": 0.0002,
      "loss": 0.6665,
      "step": 990
    },
    {
      "epoch": 1.5822784810126582,
      "grad_norm": 0.2575024664402008,
      "learning_rate": 0.0002,
      "loss": 0.6517,
      "step": 1000
    },
    {
      "epoch": 1.5822784810126582,
      "eval_loss": 0.6734818816184998,
      "eval_runtime": 131.8976,
      "eval_samples_per_second": 4.792,
      "eval_steps_per_second": 0.599,
      "step": 1000
    },
    {
      "epoch": 1.5981012658227849,
      "grad_norm": 0.2626543939113617,
      "learning_rate": 0.0002,
      "loss": 0.6728,
      "step": 1010
    },
    {
      "epoch": 1.6139240506329116,
      "grad_norm": 0.2601113021373749,
      "learning_rate": 0.0002,
      "loss": 0.6766,
      "step": 1020
    },
    {
      "epoch": 1.6139240506329116,
      "eval_loss": 0.6730306148529053,
      "eval_runtime": 131.9033,
      "eval_samples_per_second": 4.791,
      "eval_steps_per_second": 0.599,
      "step": 1020
    },
    {
      "epoch": 1.629746835443038,
      "grad_norm": 0.24218393862247467,
      "learning_rate": 0.0002,
      "loss": 0.6643,
      "step": 1030
    },
    {
      "epoch": 1.6455696202531644,
      "grad_norm": 0.2555696964263916,
      "learning_rate": 0.0002,
      "loss": 0.6685,
      "step": 1040
    },
    {
      "epoch": 1.6455696202531644,
      "eval_loss": 0.6716721057891846,
      "eval_runtime": 131.8952,
      "eval_samples_per_second": 4.792,
      "eval_steps_per_second": 0.599,
      "step": 1040
    },
    {
      "epoch": 1.6613924050632911,
      "grad_norm": 0.2642650902271271,
      "learning_rate": 0.0002,
      "loss": 0.6647,
      "step": 1050
    },
    {
      "epoch": 1.6772151898734178,
      "grad_norm": 0.26051345467567444,
      "learning_rate": 0.0002,
      "loss": 0.6469,
      "step": 1060
    },
    {
      "epoch": 1.6772151898734178,
      "eval_loss": 0.6705776453018188,
      "eval_runtime": 131.902,
      "eval_samples_per_second": 4.791,
      "eval_steps_per_second": 0.599,
      "step": 1060
    },
    {
      "epoch": 1.6930379746835444,
      "grad_norm": 0.23965594172477722,
      "learning_rate": 0.0002,
      "loss": 0.6639,
      "step": 1070
    },
    {
      "epoch": 1.7088607594936709,
      "grad_norm": 0.25241518020629883,
      "learning_rate": 0.0002,
      "loss": 0.6587,
      "step": 1080
    },
    {
      "epoch": 1.7088607594936709,
      "eval_loss": 0.6684098839759827,
      "eval_runtime": 131.8565,
      "eval_samples_per_second": 4.793,
      "eval_steps_per_second": 0.599,
      "step": 1080
    },
    {
      "epoch": 1.7246835443037973,
      "grad_norm": 0.2560000717639923,
      "learning_rate": 0.0002,
      "loss": 0.6661,
      "step": 1090
    },
    {
      "epoch": 1.740506329113924,
      "grad_norm": 0.2620084285736084,
      "learning_rate": 0.0002,
      "loss": 0.6542,
      "step": 1100
    },
    {
      "epoch": 1.740506329113924,
      "eval_loss": 0.6675167679786682,
      "eval_runtime": 131.867,
      "eval_samples_per_second": 4.793,
      "eval_steps_per_second": 0.599,
      "step": 1100
    },
    {
      "epoch": 1.7563291139240507,
      "grad_norm": 0.2625848650932312,
      "learning_rate": 0.0002,
      "loss": 0.6826,
      "step": 1110
    },
    {
      "epoch": 1.7721518987341773,
      "grad_norm": 0.26932087540626526,
      "learning_rate": 0.0002,
      "loss": 0.6701,
      "step": 1120
    },
    {
      "epoch": 1.7721518987341773,
      "eval_loss": 0.6655858755111694,
      "eval_runtime": 131.8969,
      "eval_samples_per_second": 4.792,
      "eval_steps_per_second": 0.599,
      "step": 1120
    },
    {
      "epoch": 1.7879746835443038,
      "grad_norm": 0.2615478038787842,
      "learning_rate": 0.0002,
      "loss": 0.6547,
      "step": 1130
    },
    {
      "epoch": 1.8037974683544302,
      "grad_norm": 0.25854143500328064,
      "learning_rate": 0.0002,
      "loss": 0.6585,
      "step": 1140
    },
    {
      "epoch": 1.8037974683544302,
      "eval_loss": 0.6637856960296631,
      "eval_runtime": 131.8751,
      "eval_samples_per_second": 4.792,
      "eval_steps_per_second": 0.599,
      "step": 1140
    },
    {
      "epoch": 1.8196202531645569,
      "grad_norm": 0.24300739169120789,
      "learning_rate": 0.0002,
      "loss": 0.652,
      "step": 1150
    },
    {
      "epoch": 1.8354430379746836,
      "grad_norm": 0.25698763132095337,
      "learning_rate": 0.0002,
      "loss": 0.6494,
      "step": 1160
    },
    {
      "epoch": 1.8354430379746836,
      "eval_loss": 0.6645755767822266,
      "eval_runtime": 131.9045,
      "eval_samples_per_second": 4.791,
      "eval_steps_per_second": 0.599,
      "step": 1160
    },
    {
      "epoch": 1.8512658227848102,
      "grad_norm": 0.27428802847862244,
      "learning_rate": 0.0002,
      "loss": 0.6717,
      "step": 1170
    },
    {
      "epoch": 1.8670886075949367,
      "grad_norm": 0.2548980414867401,
      "learning_rate": 0.0002,
      "loss": 0.656,
      "step": 1180
    },
    {
      "epoch": 1.8670886075949367,
      "eval_loss": 0.6619593501091003,
      "eval_runtime": 131.8391,
      "eval_samples_per_second": 4.794,
      "eval_steps_per_second": 0.599,
      "step": 1180
    },
    {
      "epoch": 1.8829113924050633,
      "grad_norm": 0.25784459710121155,
      "learning_rate": 0.0002,
      "loss": 0.6556,
      "step": 1190
    },
    {
      "epoch": 1.8987341772151898,
      "grad_norm": 0.2482602596282959,
      "learning_rate": 0.0002,
      "loss": 0.6649,
      "step": 1200
    },
    {
      "epoch": 1.8987341772151898,
      "eval_loss": 0.661401093006134,
      "eval_runtime": 131.8649,
      "eval_samples_per_second": 4.793,
      "eval_steps_per_second": 0.599,
      "step": 1200
    },
    {
      "epoch": 1.9145569620253164,
      "grad_norm": 0.23954671621322632,
      "learning_rate": 0.0002,
      "loss": 0.6578,
      "step": 1210
    },
    {
      "epoch": 1.9303797468354431,
      "grad_norm": 0.2435765266418457,
      "learning_rate": 0.0002,
      "loss": 0.6533,
      "step": 1220
    },
    {
      "epoch": 1.9303797468354431,
      "eval_loss": 0.6592634916305542,
      "eval_runtime": 131.8393,
      "eval_samples_per_second": 4.794,
      "eval_steps_per_second": 0.599,
      "step": 1220
    },
    {
      "epoch": 1.9462025316455698,
      "grad_norm": 0.24789728224277496,
      "learning_rate": 0.0002,
      "loss": 0.6694,
      "step": 1230
    },
    {
      "epoch": 1.9620253164556962,
      "grad_norm": 0.25847727060317993,
      "learning_rate": 0.0002,
      "loss": 0.6608,
      "step": 1240
    },
    {
      "epoch": 1.9620253164556962,
      "eval_loss": 0.6584632992744446,
      "eval_runtime": 131.8793,
      "eval_samples_per_second": 4.792,
      "eval_steps_per_second": 0.599,
      "step": 1240
    },
    {
      "epoch": 1.9778481012658227,
      "grad_norm": 0.2391880303621292,
      "learning_rate": 0.0002,
      "loss": 0.6493,
      "step": 1250
    },
    {
      "epoch": 1.9936708860759493,
      "grad_norm": 0.23870311677455902,
      "learning_rate": 0.0002,
      "loss": 0.6418,
      "step": 1260
    },
    {
      "epoch": 1.9936708860759493,
      "eval_loss": 0.6586398482322693,
      "eval_runtime": 131.8943,
      "eval_samples_per_second": 4.792,
      "eval_steps_per_second": 0.599,
      "step": 1260
    },
    {
      "epoch": 2.009493670886076,
      "grad_norm": 0.23965978622436523,
      "learning_rate": 0.0002,
      "loss": 0.6466,
      "step": 1270
    },
    {
      "epoch": 2.0253164556962027,
      "grad_norm": 0.26871880888938904,
      "learning_rate": 0.0002,
      "loss": 0.6221,
      "step": 1280
    },
    {
      "epoch": 2.0253164556962027,
      "eval_loss": 0.6579166650772095,
      "eval_runtime": 132.0132,
      "eval_samples_per_second": 4.787,
      "eval_steps_per_second": 0.598,
      "step": 1280
    },
    {
      "epoch": 2.041139240506329,
      "grad_norm": 0.25253158807754517,
      "learning_rate": 0.0002,
      "loss": 0.633,
      "step": 1290
    },
    {
      "epoch": 2.0569620253164556,
      "grad_norm": 0.2572665214538574,
      "learning_rate": 0.0002,
      "loss": 0.6034,
      "step": 1300
    },
    {
      "epoch": 2.0569620253164556,
      "eval_loss": 0.657181441783905,
      "eval_runtime": 131.9493,
      "eval_samples_per_second": 4.79,
      "eval_steps_per_second": 0.599,
      "step": 1300
    },
    {
      "epoch": 2.0727848101265822,
      "grad_norm": 0.26970574259757996,
      "learning_rate": 0.0002,
      "loss": 0.6159,
      "step": 1310
    },
    {
      "epoch": 2.088607594936709,
      "grad_norm": 0.2501896619796753,
      "learning_rate": 0.0002,
      "loss": 0.6206,
      "step": 1320
    },
    {
      "epoch": 2.088607594936709,
      "eval_loss": 0.657564103603363,
      "eval_runtime": 131.8725,
      "eval_samples_per_second": 4.793,
      "eval_steps_per_second": 0.599,
      "step": 1320
    },
    {
      "epoch": 2.1044303797468356,
      "grad_norm": 0.2759926915168762,
      "learning_rate": 0.0002,
      "loss": 0.6086,
      "step": 1330
    },
    {
      "epoch": 2.1202531645569622,
      "grad_norm": 0.25795313715934753,
      "learning_rate": 0.0002,
      "loss": 0.608,
      "step": 1340
    },
    {
      "epoch": 2.1202531645569622,
      "eval_loss": 0.6556717753410339,
      "eval_runtime": 131.8399,
      "eval_samples_per_second": 4.794,
      "eval_steps_per_second": 0.599,
      "step": 1340
    },
    {
      "epoch": 2.1360759493670884,
      "grad_norm": 0.24557806551456451,
      "learning_rate": 0.0002,
      "loss": 0.6248,
      "step": 1350
    },
    {
      "epoch": 2.151898734177215,
      "grad_norm": 0.2617652714252472,
      "learning_rate": 0.0002,
      "loss": 0.6183,
      "step": 1360
    },
    {
      "epoch": 2.151898734177215,
      "eval_loss": 0.655511736869812,
      "eval_runtime": 131.8152,
      "eval_samples_per_second": 4.795,
      "eval_steps_per_second": 0.599,
      "step": 1360
    },
    {
      "epoch": 2.1677215189873418,
      "grad_norm": 0.2689368724822998,
      "learning_rate": 0.0002,
      "loss": 0.6308,
      "step": 1370
    },
    {
      "epoch": 2.1835443037974684,
      "grad_norm": 0.26333004236221313,
      "learning_rate": 0.0002,
      "loss": 0.6052,
      "step": 1380
    },
    {
      "epoch": 2.1835443037974684,
      "eval_loss": 0.655017614364624,
      "eval_runtime": 131.8216,
      "eval_samples_per_second": 4.794,
      "eval_steps_per_second": 0.599,
      "step": 1380
    },
    {
      "epoch": 2.199367088607595,
      "grad_norm": 0.26235902309417725,
      "learning_rate": 0.0002,
      "loss": 0.6348,
      "step": 1390
    },
    {
      "epoch": 2.2151898734177213,
      "grad_norm": 0.25419360399246216,
      "learning_rate": 0.0002,
      "loss": 0.604,
      "step": 1400
    },
    {
      "epoch": 2.2151898734177213,
      "eval_loss": 0.6535363793373108,
      "eval_runtime": 131.9185,
      "eval_samples_per_second": 4.791,
      "eval_steps_per_second": 0.599,
      "step": 1400
    },
    {
      "epoch": 2.231012658227848,
      "grad_norm": 0.24557967483997345,
      "learning_rate": 0.0002,
      "loss": 0.6171,
      "step": 1410
    },
    {
      "epoch": 2.2468354430379747,
      "grad_norm": 0.2597355246543884,
      "learning_rate": 0.0002,
      "loss": 0.64,
      "step": 1420
    },
    {
      "epoch": 2.2468354430379747,
      "eval_loss": 0.6533951163291931,
      "eval_runtime": 131.8581,
      "eval_samples_per_second": 4.793,
      "eval_steps_per_second": 0.599,
      "step": 1420
    },
    {
      "epoch": 2.2626582278481013,
      "grad_norm": 0.25525757670402527,
      "learning_rate": 0.0002,
      "loss": 0.6087,
      "step": 1430
    },
    {
      "epoch": 2.278481012658228,
      "grad_norm": 0.26448124647140503,
      "learning_rate": 0.0002,
      "loss": 0.6063,
      "step": 1440
    },
    {
      "epoch": 2.278481012658228,
      "eval_loss": 0.6525977253913879,
      "eval_runtime": 131.9317,
      "eval_samples_per_second": 4.79,
      "eval_steps_per_second": 0.599,
      "step": 1440
    },
    {
      "epoch": 2.2943037974683547,
      "grad_norm": 0.24552085995674133,
      "learning_rate": 0.0002,
      "loss": 0.6025,
      "step": 1450
    },
    {
      "epoch": 2.310126582278481,
      "grad_norm": 0.26224812865257263,
      "learning_rate": 0.0002,
      "loss": 0.6107,
      "step": 1460
    },
    {
      "epoch": 2.310126582278481,
      "eval_loss": 0.6506198048591614,
      "eval_runtime": 131.9002,
      "eval_samples_per_second": 4.792,
      "eval_steps_per_second": 0.599,
      "step": 1460
    },
    {
      "epoch": 2.3259493670886076,
      "grad_norm": 0.25793251395225525,
      "learning_rate": 0.0002,
      "loss": 0.6119,
      "step": 1470
    },
    {
      "epoch": 2.3417721518987342,
      "grad_norm": 0.2662043273448944,
      "learning_rate": 0.0002,
      "loss": 0.6219,
      "step": 1480
    },
    {
      "epoch": 2.3417721518987342,
      "eval_loss": 0.6498812437057495,
      "eval_runtime": 131.8871,
      "eval_samples_per_second": 4.792,
      "eval_steps_per_second": 0.599,
      "step": 1480
    },
    {
      "epoch": 2.357594936708861,
      "grad_norm": 0.28206878900527954,
      "learning_rate": 0.0002,
      "loss": 0.6308,
      "step": 1490
    },
    {
      "epoch": 2.3734177215189876,
      "grad_norm": 0.2855586111545563,
      "learning_rate": 0.0002,
      "loss": 0.5985,
      "step": 1500
    },
    {
      "epoch": 2.3734177215189876,
      "eval_loss": 0.6503695845603943,
      "eval_runtime": 131.8626,
      "eval_samples_per_second": 4.793,
      "eval_steps_per_second": 0.599,
      "step": 1500
    },
    {
      "epoch": 2.3892405063291138,
      "grad_norm": 0.271095871925354,
      "learning_rate": 0.0002,
      "loss": 0.6274,
      "step": 1510
    },
    {
      "epoch": 2.4050632911392404,
      "grad_norm": 0.2552088797092438,
      "learning_rate": 0.0002,
      "loss": 0.6069,
      "step": 1520
    },
    {
      "epoch": 2.4050632911392404,
      "eval_loss": 0.6493340134620667,
      "eval_runtime": 131.8519,
      "eval_samples_per_second": 4.793,
      "eval_steps_per_second": 0.599,
      "step": 1520
    },
    {
      "epoch": 2.420886075949367,
      "grad_norm": 0.26949629187583923,
      "learning_rate": 0.0002,
      "loss": 0.6203,
      "step": 1530
    },
    {
      "epoch": 2.4367088607594938,
      "grad_norm": 0.268215537071228,
      "learning_rate": 0.0002,
      "loss": 0.6174,
      "step": 1540
    },
    {
      "epoch": 2.4367088607594938,
      "eval_loss": 0.6472347378730774,
      "eval_runtime": 131.7911,
      "eval_samples_per_second": 4.795,
      "eval_steps_per_second": 0.599,
      "step": 1540
    },
    {
      "epoch": 2.4525316455696204,
      "grad_norm": 0.27880433201789856,
      "learning_rate": 0.0002,
      "loss": 0.6158,
      "step": 1550
    },
    {
      "epoch": 2.4683544303797467,
      "grad_norm": 0.2624712586402893,
      "learning_rate": 0.0002,
      "loss": 0.6198,
      "step": 1560
    },
    {
      "epoch": 2.4683544303797467,
      "eval_loss": 0.6470383405685425,
      "eval_runtime": 131.796,
      "eval_samples_per_second": 4.795,
      "eval_steps_per_second": 0.599,
      "step": 1560
    },
    {
      "epoch": 2.4841772151898733,
      "grad_norm": 0.25361037254333496,
      "learning_rate": 0.0002,
      "loss": 0.6119,
      "step": 1570
    },
    {
      "epoch": 2.5,
      "grad_norm": 0.2555208206176758,
      "learning_rate": 0.0002,
      "loss": 0.5924,
      "step": 1580
    },
    {
      "epoch": 2.5,
      "eval_loss": 0.646508514881134,
      "eval_runtime": 131.8167,
      "eval_samples_per_second": 4.795,
      "eval_steps_per_second": 0.599,
      "step": 1580
    },
    {
      "epoch": 2.5158227848101267,
      "grad_norm": 0.26373204588890076,
      "learning_rate": 0.0002,
      "loss": 0.6164,
      "step": 1590
    },
    {
      "epoch": 2.5316455696202533,
      "grad_norm": 0.26238492131233215,
      "learning_rate": 0.0002,
      "loss": 0.6048,
      "step": 1600
    },
    {
      "epoch": 2.5316455696202533,
      "eval_loss": 0.645945131778717,
      "eval_runtime": 131.877,
      "eval_samples_per_second": 4.792,
      "eval_steps_per_second": 0.599,
      "step": 1600
    },
    {
      "epoch": 2.5474683544303796,
      "grad_norm": 0.26891419291496277,
      "learning_rate": 0.0002,
      "loss": 0.6085,
      "step": 1610
    },
    {
      "epoch": 2.5632911392405062,
      "grad_norm": 0.28595566749572754,
      "learning_rate": 0.0002,
      "loss": 0.6146,
      "step": 1620
    },
    {
      "epoch": 2.5632911392405062,
      "eval_loss": 0.6452015042304993,
      "eval_runtime": 131.8123,
      "eval_samples_per_second": 4.795,
      "eval_steps_per_second": 0.599,
      "step": 1620
    },
    {
      "epoch": 2.579113924050633,
      "grad_norm": 0.25937801599502563,
      "learning_rate": 0.0002,
      "loss": 0.6215,
      "step": 1630
    },
    {
      "epoch": 2.5949367088607596,
      "grad_norm": 0.25574633479118347,
      "learning_rate": 0.0002,
      "loss": 0.6207,
      "step": 1640
    },
    {
      "epoch": 2.5949367088607596,
      "eval_loss": 0.6426708698272705,
      "eval_runtime": 131.8011,
      "eval_samples_per_second": 4.795,
      "eval_steps_per_second": 0.599,
      "step": 1640
    },
    {
      "epoch": 2.6107594936708862,
      "grad_norm": 0.24796649813652039,
      "learning_rate": 0.0002,
      "loss": 0.6036,
      "step": 1650
    },
    {
      "epoch": 2.6265822784810124,
      "grad_norm": 0.26026397943496704,
      "learning_rate": 0.0002,
      "loss": 0.6094,
      "step": 1660
    },
    {
      "epoch": 2.6265822784810124,
      "eval_loss": 0.6430174708366394,
      "eval_runtime": 131.8934,
      "eval_samples_per_second": 4.792,
      "eval_steps_per_second": 0.599,
      "step": 1660
    },
    {
      "epoch": 2.642405063291139,
      "grad_norm": 0.24800552427768707,
      "learning_rate": 0.0002,
      "loss": 0.6135,
      "step": 1670
    },
    {
      "epoch": 2.6582278481012658,
      "grad_norm": 0.2579246461391449,
      "learning_rate": 0.0002,
      "loss": 0.6145,
      "step": 1680
    },
    {
      "epoch": 2.6582278481012658,
      "eval_loss": 0.6424046158790588,
      "eval_runtime": 131.8962,
      "eval_samples_per_second": 4.792,
      "eval_steps_per_second": 0.599,
      "step": 1680
    },
    {
      "epoch": 2.6740506329113924,
      "grad_norm": 0.2610401213169098,
      "learning_rate": 0.0002,
      "loss": 0.6115,
      "step": 1690
    },
    {
      "epoch": 2.689873417721519,
      "grad_norm": 0.25508347153663635,
      "learning_rate": 0.0002,
      "loss": 0.605,
      "step": 1700
    },
    {
      "epoch": 2.689873417721519,
      "eval_loss": 0.6421927213668823,
      "eval_runtime": 131.8697,
      "eval_samples_per_second": 4.793,
      "eval_steps_per_second": 0.599,
      "step": 1700
    },
    {
      "epoch": 2.7056962025316453,
      "grad_norm": 0.2684323489665985,
      "learning_rate": 0.0002,
      "loss": 0.6092,
      "step": 1710
    },
    {
      "epoch": 2.721518987341772,
      "grad_norm": 0.2703189551830292,
      "learning_rate": 0.0002,
      "loss": 0.6101,
      "step": 1720
    },
    {
      "epoch": 2.721518987341772,
      "eval_loss": 0.6405203342437744,
      "eval_runtime": 131.899,
      "eval_samples_per_second": 4.792,
      "eval_steps_per_second": 0.599,
      "step": 1720
    },
    {
      "epoch": 2.7373417721518987,
      "grad_norm": 0.2539083659648895,
      "learning_rate": 0.0002,
      "loss": 0.624,
      "step": 1730
    },
    {
      "epoch": 2.7531645569620253,
      "grad_norm": 0.2978338897228241,
      "learning_rate": 0.0002,
      "loss": 0.6005,
      "step": 1740
    },
    {
      "epoch": 2.7531645569620253,
      "eval_loss": 0.6394311189651489,
      "eval_runtime": 131.9208,
      "eval_samples_per_second": 4.791,
      "eval_steps_per_second": 0.599,
      "step": 1740
    },
    {
      "epoch": 2.768987341772152,
      "grad_norm": 0.27297040820121765,
      "learning_rate": 0.0002,
      "loss": 0.612,
      "step": 1750
    },
    {
      "epoch": 2.7848101265822782,
      "grad_norm": 0.2446739226579666,
      "learning_rate": 0.0002,
      "loss": 0.6271,
      "step": 1760
    },
    {
      "epoch": 2.7848101265822782,
      "eval_loss": 0.6381310820579529,
      "eval_runtime": 131.8575,
      "eval_samples_per_second": 4.793,
      "eval_steps_per_second": 0.599,
      "step": 1760
    },
    {
      "epoch": 2.8006329113924053,
      "grad_norm": 0.26931503415107727,
      "learning_rate": 0.0002,
      "loss": 0.603,
      "step": 1770
    },
    {
      "epoch": 2.8164556962025316,
      "grad_norm": 0.2579115927219391,
      "learning_rate": 0.0002,
      "loss": 0.6041,
      "step": 1780
    },
    {
      "epoch": 2.8164556962025316,
      "eval_loss": 0.6374809741973877,
      "eval_runtime": 131.8822,
      "eval_samples_per_second": 4.792,
      "eval_steps_per_second": 0.599,
      "step": 1780
    },
    {
      "epoch": 2.8322784810126582,
      "grad_norm": 0.2737986147403717,
      "learning_rate": 0.0002,
      "loss": 0.6193,
      "step": 1790
    },
    {
      "epoch": 2.848101265822785,
      "grad_norm": 0.2413921058177948,
      "learning_rate": 0.0002,
      "loss": 0.6336,
      "step": 1800
    },
    {
      "epoch": 2.848101265822785,
      "eval_loss": 0.6369552612304688,
      "eval_runtime": 131.8875,
      "eval_samples_per_second": 4.792,
      "eval_steps_per_second": 0.599,
      "step": 1800
    },
    {
      "epoch": 2.8639240506329116,
      "grad_norm": 0.2685437798500061,
      "learning_rate": 0.0002,
      "loss": 0.6085,
      "step": 1810
    },
    {
      "epoch": 2.879746835443038,
      "grad_norm": 0.26042917370796204,
      "learning_rate": 0.0002,
      "loss": 0.6173,
      "step": 1820
    },
    {
      "epoch": 2.879746835443038,
      "eval_loss": 0.6361930966377258,
      "eval_runtime": 131.8683,
      "eval_samples_per_second": 4.793,
      "eval_steps_per_second": 0.599,
      "step": 1820
    },
    {
      "epoch": 2.8955696202531644,
      "grad_norm": 0.24503293633460999,
      "learning_rate": 0.0002,
      "loss": 0.6162,
      "step": 1830
    },
    {
      "epoch": 2.911392405063291,
      "grad_norm": 0.2594417929649353,
      "learning_rate": 0.0002,
      "loss": 0.5924,
      "step": 1840
    },
    {
      "epoch": 2.911392405063291,
      "eval_loss": 0.6372414827346802,
      "eval_runtime": 131.8322,
      "eval_samples_per_second": 4.794,
      "eval_steps_per_second": 0.599,
      "step": 1840
    },
    {
      "epoch": 2.9272151898734178,
      "grad_norm": 0.2502780556678772,
      "learning_rate": 0.0002,
      "loss": 0.5984,
      "step": 1850
    },
    {
      "epoch": 2.9430379746835444,
      "grad_norm": 0.2585500478744507,
      "learning_rate": 0.0002,
      "loss": 0.6187,
      "step": 1860
    },
    {
      "epoch": 2.9430379746835444,
      "eval_loss": 0.6348665952682495,
      "eval_runtime": 131.8489,
      "eval_samples_per_second": 4.793,
      "eval_steps_per_second": 0.599,
      "step": 1860
    },
    {
      "epoch": 2.958860759493671,
      "grad_norm": 0.26718345284461975,
      "learning_rate": 0.0002,
      "loss": 0.6129,
      "step": 1870
    },
    {
      "epoch": 2.9746835443037973,
      "grad_norm": 0.259824275970459,
      "learning_rate": 0.0002,
      "loss": 0.6057,
      "step": 1880
    },
    {
      "epoch": 2.9746835443037973,
      "eval_loss": 0.6338453888893127,
      "eval_runtime": 131.8288,
      "eval_samples_per_second": 4.794,
      "eval_steps_per_second": 0.599,
      "step": 1880
    },
    {
      "epoch": 2.990506329113924,
      "grad_norm": 0.26295337080955505,
      "learning_rate": 0.0002,
      "loss": 0.6173,
      "step": 1890
    },
    {
      "epoch": 3.0063291139240507,
      "grad_norm": 0.26111751794815063,
      "learning_rate": 0.0002,
      "loss": 0.5942,
      "step": 1900
    },
    {
      "epoch": 3.0063291139240507,
      "eval_loss": 0.6363694667816162,
      "eval_runtime": 131.9154,
      "eval_samples_per_second": 4.791,
      "eval_steps_per_second": 0.599,
      "step": 1900
    },
    {
      "epoch": 3.0221518987341773,
      "grad_norm": 0.2783117890357971,
      "learning_rate": 0.0002,
      "loss": 0.569,
      "step": 1910
    },
    {
      "epoch": 3.037974683544304,
      "grad_norm": 0.2551935911178589,
      "learning_rate": 0.0002,
      "loss": 0.5573,
      "step": 1920
    },
    {
      "epoch": 3.037974683544304,
      "eval_loss": 0.6390289068222046,
      "eval_runtime": 131.8595,
      "eval_samples_per_second": 4.793,
      "eval_steps_per_second": 0.599,
      "step": 1920
    },
    {
      "epoch": 3.0537974683544302,
      "grad_norm": 0.27019408345222473,
      "learning_rate": 0.0002,
      "loss": 0.5564,
      "step": 1930
    },
    {
      "epoch": 3.069620253164557,
      "grad_norm": 0.2871812582015991,
      "learning_rate": 0.0002,
      "loss": 0.587,
      "step": 1940
    },
    {
      "epoch": 3.069620253164557,
      "eval_loss": 0.6355338096618652,
      "eval_runtime": 131.8991,
      "eval_samples_per_second": 4.792,
      "eval_steps_per_second": 0.599,
      "step": 1940
    },
    {
      "epoch": 3.0854430379746836,
      "grad_norm": 0.25816774368286133,
      "learning_rate": 0.0002,
      "loss": 0.5632,
      "step": 1950
    },
    {
      "epoch": 3.1012658227848102,
      "grad_norm": 0.265296071767807,
      "learning_rate": 0.0002,
      "loss": 0.5689,
      "step": 1960
    },
    {
      "epoch": 3.1012658227848102,
      "eval_loss": 0.63770592212677,
      "eval_runtime": 131.9096,
      "eval_samples_per_second": 4.791,
      "eval_steps_per_second": 0.599,
      "step": 1960
    },
    {
      "epoch": 3.117088607594937,
      "grad_norm": 0.25830793380737305,
      "learning_rate": 0.0002,
      "loss": 0.5746,
      "step": 1970
    },
    {
      "epoch": 3.132911392405063,
      "grad_norm": 0.2769959270954132,
      "learning_rate": 0.0002,
      "loss": 0.5534,
      "step": 1980
    },
    {
      "epoch": 3.132911392405063,
      "eval_loss": 0.6358143091201782,
      "eval_runtime": 131.8634,
      "eval_samples_per_second": 4.793,
      "eval_steps_per_second": 0.599,
      "step": 1980
    },
    {
      "epoch": 3.1487341772151898,
      "grad_norm": 0.2655363082885742,
      "learning_rate": 0.0002,
      "loss": 0.5811,
      "step": 1990
    },
    {
      "epoch": 3.1645569620253164,
      "grad_norm": 0.25941500067710876,
      "learning_rate": 0.0002,
      "loss": 0.5682,
      "step": 2000
    },
    {
      "epoch": 3.1645569620253164,
      "eval_loss": 0.636451780796051,
      "eval_runtime": 131.857,
      "eval_samples_per_second": 4.793,
      "eval_steps_per_second": 0.599,
      "step": 2000
    },
    {
      "epoch": 3.180379746835443,
      "grad_norm": 0.28629186749458313,
      "learning_rate": 0.0002,
      "loss": 0.5729,
      "step": 2010
    },
    {
      "epoch": 3.1962025316455698,
      "grad_norm": 0.2634666860103607,
      "learning_rate": 0.0002,
      "loss": 0.5693,
      "step": 2020
    },
    {
      "epoch": 3.1962025316455698,
      "eval_loss": 0.6351698040962219,
      "eval_runtime": 131.7971,
      "eval_samples_per_second": 4.795,
      "eval_steps_per_second": 0.599,
      "step": 2020
    },
    {
      "epoch": 3.212025316455696,
      "grad_norm": 0.27958911657333374,
      "learning_rate": 0.0002,
      "loss": 0.5697,
      "step": 2030
    },
    {
      "epoch": 3.2278481012658227,
      "grad_norm": 0.2637881636619568,
      "learning_rate": 0.0002,
      "loss": 0.5654,
      "step": 2040
    },
    {
      "epoch": 3.2278481012658227,
      "eval_loss": 0.6344249844551086,
      "eval_runtime": 131.7917,
      "eval_samples_per_second": 4.795,
      "eval_steps_per_second": 0.599,
      "step": 2040
    },
    {
      "epoch": 3.2436708860759493,
      "grad_norm": 0.2683382034301758,
      "learning_rate": 0.0002,
      "loss": 0.5623,
      "step": 2050
    },
    {
      "epoch": 3.259493670886076,
      "grad_norm": 0.2672869563102722,
      "learning_rate": 0.0002,
      "loss": 0.5466,
      "step": 2060
    },
    {
      "epoch": 3.259493670886076,
      "eval_loss": 0.6347548365592957,
      "eval_runtime": 131.8564,
      "eval_samples_per_second": 4.793,
      "eval_steps_per_second": 0.599,
      "step": 2060
    },
    {
      "epoch": 3.2753164556962027,
      "grad_norm": 0.26369979977607727,
      "learning_rate": 0.0002,
      "loss": 0.5832,
      "step": 2070
    },
    {
      "epoch": 3.291139240506329,
      "grad_norm": 0.2876640856266022,
      "learning_rate": 0.0002,
      "loss": 0.5782,
      "step": 2080
    },
    {
      "epoch": 3.291139240506329,
      "eval_loss": 0.6328052282333374,
      "eval_runtime": 131.9245,
      "eval_samples_per_second": 4.791,
      "eval_steps_per_second": 0.599,
      "step": 2080
    },
    {
      "epoch": 3.3069620253164556,
      "grad_norm": 0.3062177002429962,
      "learning_rate": 0.0002,
      "loss": 0.5718,
      "step": 2090
    },
    {
      "epoch": 3.3227848101265822,
      "grad_norm": 0.2935482859611511,
      "learning_rate": 0.0002,
      "loss": 0.5838,
      "step": 2100
    },
    {
      "epoch": 3.3227848101265822,
      "eval_loss": 0.6349889039993286,
      "eval_runtime": 131.87,
      "eval_samples_per_second": 4.793,
      "eval_steps_per_second": 0.599,
      "step": 2100
    },
    {
      "epoch": 3.338607594936709,
      "grad_norm": 0.25793778896331787,
      "learning_rate": 0.0002,
      "loss": 0.5703,
      "step": 2110
    },
    {
      "epoch": 3.3544303797468356,
      "grad_norm": 0.27156421542167664,
      "learning_rate": 0.0002,
      "loss": 0.5745,
      "step": 2120
    },
    {
      "epoch": 3.3544303797468356,
      "eval_loss": 0.6326487064361572,
      "eval_runtime": 131.838,
      "eval_samples_per_second": 4.794,
      "eval_steps_per_second": 0.599,
      "step": 2120
    },
    {
      "epoch": 3.370253164556962,
      "grad_norm": 0.2985718548297882,
      "learning_rate": 0.0002,
      "loss": 0.5719,
      "step": 2130
    },
    {
      "epoch": 3.3860759493670884,
      "grad_norm": 0.28314507007598877,
      "learning_rate": 0.0002,
      "loss": 0.5708,
      "step": 2140
    },
    {
      "epoch": 3.3860759493670884,
      "eval_loss": 0.6317163109779358,
      "eval_runtime": 131.8745,
      "eval_samples_per_second": 4.792,
      "eval_steps_per_second": 0.599,
      "step": 2140
    },
    {
      "epoch": 3.401898734177215,
      "grad_norm": 0.2604416608810425,
      "learning_rate": 0.0002,
      "loss": 0.5801,
      "step": 2150
    },
    {
      "epoch": 3.4177215189873418,
      "grad_norm": 0.2912323772907257,
      "learning_rate": 0.0002,
      "loss": 0.5891,
      "step": 2160
    },
    {
      "epoch": 3.4177215189873418,
      "eval_loss": 0.6304855346679688,
      "eval_runtime": 131.8583,
      "eval_samples_per_second": 4.793,
      "eval_steps_per_second": 0.599,
      "step": 2160
    },
    {
      "epoch": 3.4335443037974684,
      "grad_norm": 0.3034590780735016,
      "learning_rate": 0.0002,
      "loss": 0.5694,
      "step": 2170
    },
    {
      "epoch": 3.449367088607595,
      "grad_norm": 0.263943612575531,
      "learning_rate": 0.0002,
      "loss": 0.5818,
      "step": 2180
    },
    {
      "epoch": 3.449367088607595,
      "eval_loss": 0.6299254894256592,
      "eval_runtime": 131.8245,
      "eval_samples_per_second": 4.794,
      "eval_steps_per_second": 0.599,
      "step": 2180
    },
    {
      "epoch": 3.4651898734177213,
      "grad_norm": 0.2781783640384674,
      "learning_rate": 0.0002,
      "loss": 0.5624,
      "step": 2190
    },
    {
      "epoch": 3.481012658227848,
      "grad_norm": 0.2678270637989044,
      "learning_rate": 0.0002,
      "loss": 0.5656,
      "step": 2200
    },
    {
      "epoch": 3.481012658227848,
      "eval_loss": 0.6314519643783569,
      "eval_runtime": 131.8584,
      "eval_samples_per_second": 4.793,
      "eval_steps_per_second": 0.599,
      "step": 2200
    },
    {
      "epoch": 3.4968354430379747,
      "grad_norm": 0.2663576006889343,
      "learning_rate": 0.0002,
      "loss": 0.5723,
      "step": 2210
    },
    {
      "epoch": 3.5126582278481013,
      "grad_norm": 0.2613811194896698,
      "learning_rate": 0.0002,
      "loss": 0.5767,
      "step": 2220
    },
    {
      "epoch": 3.5126582278481013,
      "eval_loss": 0.6281495690345764,
      "eval_runtime": 131.8917,
      "eval_samples_per_second": 4.792,
      "eval_steps_per_second": 0.599,
      "step": 2220
    },
    {
      "epoch": 3.528481012658228,
      "grad_norm": 0.30084243416786194,
      "learning_rate": 0.0002,
      "loss": 0.5851,
      "step": 2230
    },
    {
      "epoch": 3.5443037974683547,
      "grad_norm": 0.2818966507911682,
      "learning_rate": 0.0002,
      "loss": 0.5661,
      "step": 2240
    },
    {
      "epoch": 3.5443037974683547,
      "eval_loss": 0.629593014717102,
      "eval_runtime": 131.8896,
      "eval_samples_per_second": 4.792,
      "eval_steps_per_second": 0.599,
      "step": 2240
    },
    {
      "epoch": 3.560126582278481,
      "grad_norm": 0.2574240565299988,
      "learning_rate": 0.0002,
      "loss": 0.5692,
      "step": 2250
    },
    {
      "epoch": 3.5759493670886076,
      "grad_norm": 0.29194533824920654,
      "learning_rate": 0.0002,
      "loss": 0.5749,
      "step": 2260
    },
    {
      "epoch": 3.5759493670886076,
      "eval_loss": 0.6296534538269043,
      "eval_runtime": 131.8964,
      "eval_samples_per_second": 4.792,
      "eval_steps_per_second": 0.599,
      "step": 2260
    },
    {
      "epoch": 3.5917721518987342,
      "grad_norm": 0.2794256806373596,
      "learning_rate": 0.0002,
      "loss": 0.5779,
      "step": 2270
    },
    {
      "epoch": 3.607594936708861,
      "grad_norm": 0.2534295320510864,
      "learning_rate": 0.0002,
      "loss": 0.5632,
      "step": 2280
    },
    {
      "epoch": 3.607594936708861,
      "eval_loss": 0.6269585490226746,
      "eval_runtime": 131.8528,
      "eval_samples_per_second": 4.793,
      "eval_steps_per_second": 0.599,
      "step": 2280
    },
    {
      "epoch": 3.6234177215189876,
      "grad_norm": 0.26690003275871277,
      "learning_rate": 0.0002,
      "loss": 0.5751,
      "step": 2290
    },
    {
      "epoch": 3.6392405063291138,
      "grad_norm": 0.27994346618652344,
      "learning_rate": 0.0002,
      "loss": 0.5784,
      "step": 2300
    },
    {
      "epoch": 3.6392405063291138,
      "eval_loss": 0.6280952095985413,
      "eval_runtime": 131.8912,
      "eval_samples_per_second": 4.792,
      "eval_steps_per_second": 0.599,
      "step": 2300
    },
    {
      "epoch": 3.6550632911392404,
      "grad_norm": 0.26842448115348816,
      "learning_rate": 0.0002,
      "loss": 0.5717,
      "step": 2310
    },
    {
      "epoch": 3.670886075949367,
      "grad_norm": 0.28392571210861206,
      "learning_rate": 0.0002,
      "loss": 0.5746,
      "step": 2320
    },
    {
      "epoch": 3.670886075949367,
      "eval_loss": 0.6269270181655884,
      "eval_runtime": 131.9086,
      "eval_samples_per_second": 4.791,
      "eval_steps_per_second": 0.599,
      "step": 2320
    },
    {
      "epoch": 3.6867088607594938,
      "grad_norm": 0.27782168984413147,
      "learning_rate": 0.0002,
      "loss": 0.5872,
      "step": 2330
    },
    {
      "epoch": 3.7025316455696204,
      "grad_norm": 0.27614086866378784,
      "learning_rate": 0.0002,
      "loss": 0.5641,
      "step": 2340
    },
    {
      "epoch": 3.7025316455696204,
      "eval_loss": 0.6254413723945618,
      "eval_runtime": 131.8607,
      "eval_samples_per_second": 4.793,
      "eval_steps_per_second": 0.599,
      "step": 2340
    },
    {
      "epoch": 3.7183544303797467,
      "grad_norm": 0.26656702160835266,
      "learning_rate": 0.0002,
      "loss": 0.5663,
      "step": 2350
    },
    {
      "epoch": 3.7341772151898733,
      "grad_norm": 0.26700153946876526,
      "learning_rate": 0.0002,
      "loss": 0.5733,
      "step": 2360
    },
    {
      "epoch": 3.7341772151898733,
      "eval_loss": 0.6256644129753113,
      "eval_runtime": 131.9246,
      "eval_samples_per_second": 4.791,
      "eval_steps_per_second": 0.599,
      "step": 2360
    },
    {
      "epoch": 3.75,
      "grad_norm": 0.27719756960868835,
      "learning_rate": 0.0002,
      "loss": 0.5841,
      "step": 2370
    },
    {
      "epoch": 3.7658227848101267,
      "grad_norm": 0.27030646800994873,
      "learning_rate": 0.0002,
      "loss": 0.5432,
      "step": 2380
    },
    {
      "epoch": 3.7658227848101267,
      "eval_loss": 0.6274837255477905,
      "eval_runtime": 131.8987,
      "eval_samples_per_second": 4.792,
      "eval_steps_per_second": 0.599,
      "step": 2380
    },
    {
      "epoch": 3.7816455696202533,
      "grad_norm": 0.2677309811115265,
      "learning_rate": 0.0002,
      "loss": 0.5806,
      "step": 2390
    },
    {
      "epoch": 3.7974683544303796,
      "grad_norm": 0.2937711775302887,
      "learning_rate": 0.0002,
      "loss": 0.567,
      "step": 2400
    },
    {
      "epoch": 3.7974683544303796,
      "eval_loss": 0.6241787672042847,
      "eval_runtime": 131.9033,
      "eval_samples_per_second": 4.791,
      "eval_steps_per_second": 0.599,
      "step": 2400
    },
    {
      "epoch": 3.8132911392405062,
      "grad_norm": 0.2765547037124634,
      "learning_rate": 0.0002,
      "loss": 0.5721,
      "step": 2410
    },
    {
      "epoch": 3.829113924050633,
      "grad_norm": 0.26176583766937256,
      "learning_rate": 0.0002,
      "loss": 0.5747,
      "step": 2420
    },
    {
      "epoch": 3.829113924050633,
      "eval_loss": 0.6235907673835754,
      "eval_runtime": 131.8649,
      "eval_samples_per_second": 4.793,
      "eval_steps_per_second": 0.599,
      "step": 2420
    },
    {
      "epoch": 3.8449367088607596,
      "grad_norm": 0.26931002736091614,
      "learning_rate": 0.0002,
      "loss": 0.5693,
      "step": 2430
    },
    {
      "epoch": 3.8607594936708862,
      "grad_norm": 0.278001606464386,
      "learning_rate": 0.0002,
      "loss": 0.566,
      "step": 2440
    },
    {
      "epoch": 3.8607594936708862,
      "eval_loss": 0.6221994161605835,
      "eval_runtime": 131.8728,
      "eval_samples_per_second": 4.792,
      "eval_steps_per_second": 0.599,
      "step": 2440
    },
    {
      "epoch": 3.8765822784810124,
      "grad_norm": 0.26464810967445374,
      "learning_rate": 0.0002,
      "loss": 0.5728,
      "step": 2450
    },
    {
      "epoch": 3.892405063291139,
      "grad_norm": 0.2813071310520172,
      "learning_rate": 0.0002,
      "loss": 0.5783,
      "step": 2460
    },
    {
      "epoch": 3.892405063291139,
      "eval_loss": 0.6222373247146606,
      "eval_runtime": 131.919,
      "eval_samples_per_second": 4.791,
      "eval_steps_per_second": 0.599,
      "step": 2460
    },
    {
      "epoch": 3.9082278481012658,
      "grad_norm": 0.2681663930416107,
      "learning_rate": 0.0002,
      "loss": 0.5788,
      "step": 2470
    },
    {
      "epoch": 3.9240506329113924,
      "grad_norm": 0.2852930426597595,
      "learning_rate": 0.0002,
      "loss": 0.5602,
      "step": 2480
    },
    {
      "epoch": 3.9240506329113924,
      "eval_loss": 0.6225759983062744,
      "eval_runtime": 131.8675,
      "eval_samples_per_second": 4.793,
      "eval_steps_per_second": 0.599,
      "step": 2480
    },
    {
      "epoch": 3.939873417721519,
      "grad_norm": 0.25480765104293823,
      "learning_rate": 0.0002,
      "loss": 0.5751,
      "step": 2490
    },
    {
      "epoch": 3.9556962025316453,
      "grad_norm": 0.26707321405410767,
      "learning_rate": 0.0002,
      "loss": 0.5748,
      "step": 2500
    },
    {
      "epoch": 3.9556962025316453,
      "eval_loss": 0.620729386806488,
      "eval_runtime": 131.9156,
      "eval_samples_per_second": 4.791,
      "eval_steps_per_second": 0.599,
      "step": 2500
    },
    {
      "epoch": 3.971518987341772,
      "grad_norm": 0.264605849981308,
      "learning_rate": 0.0002,
      "loss": 0.5689,
      "step": 2510
    },
    {
      "epoch": 3.9873417721518987,
      "grad_norm": 0.2832660377025604,
      "learning_rate": 0.0002,
      "loss": 0.5753,
      "step": 2520
    },
    {
      "epoch": 3.9873417721518987,
      "eval_loss": 0.6208144426345825,
      "eval_runtime": 131.8513,
      "eval_samples_per_second": 4.793,
      "eval_steps_per_second": 0.599,
      "step": 2520
    },
    {
      "epoch": 4.003164556962025,
      "grad_norm": 0.26203468441963196,
      "learning_rate": 0.0002,
      "loss": 0.5592,
      "step": 2530
    },
    {
      "epoch": 4.018987341772152,
      "grad_norm": 0.2708684802055359,
      "learning_rate": 0.0002,
      "loss": 0.5384,
      "step": 2540
    },
    {
      "epoch": 4.018987341772152,
      "eval_loss": 0.6223664879798889,
      "eval_runtime": 131.9321,
      "eval_samples_per_second": 4.79,
      "eval_steps_per_second": 0.599,
      "step": 2540
    },
    {
      "epoch": 4.034810126582278,
      "grad_norm": 0.2994883358478546,
      "learning_rate": 0.0002,
      "loss": 0.5291,
      "step": 2550
    },
    {
      "epoch": 4.050632911392405,
      "grad_norm": 0.28201737999916077,
      "learning_rate": 0.0002,
      "loss": 0.526,
      "step": 2560
    },
    {
      "epoch": 4.050632911392405,
      "eval_loss": 0.6275315284729004,
      "eval_runtime": 131.8771,
      "eval_samples_per_second": 4.792,
      "eval_steps_per_second": 0.599,
      "step": 2560
    },
    {
      "epoch": 4.0664556962025316,
      "grad_norm": 0.28186795115470886,
      "learning_rate": 0.0002,
      "loss": 0.5137,
      "step": 2570
    },
    {
      "epoch": 4.082278481012658,
      "grad_norm": 0.27791962027549744,
      "learning_rate": 0.0002,
      "loss": 0.5263,
      "step": 2580
    },
    {
      "epoch": 4.082278481012658,
      "eval_loss": 0.6268836855888367,
      "eval_runtime": 131.888,
      "eval_samples_per_second": 4.792,
      "eval_steps_per_second": 0.599,
      "step": 2580
    },
    {
      "epoch": 4.098101265822785,
      "grad_norm": 0.26893365383148193,
      "learning_rate": 0.0002,
      "loss": 0.5209,
      "step": 2590
    },
    {
      "epoch": 4.113924050632911,
      "grad_norm": 0.2870092988014221,
      "learning_rate": 0.0002,
      "loss": 0.5249,
      "step": 2600
    },
    {
      "epoch": 4.113924050632911,
      "eval_loss": 0.625966489315033,
      "eval_runtime": 131.8722,
      "eval_samples_per_second": 4.793,
      "eval_steps_per_second": 0.599,
      "step": 2600
    },
    {
      "epoch": 4.129746835443038,
      "grad_norm": 0.2935843765735626,
      "learning_rate": 0.0002,
      "loss": 0.522,
      "step": 2610
    },
    {
      "epoch": 4.1455696202531644,
      "grad_norm": 0.31569287180900574,
      "learning_rate": 0.0002,
      "loss": 0.5253,
      "step": 2620
    },
    {
      "epoch": 4.1455696202531644,
      "eval_loss": 0.6253317594528198,
      "eval_runtime": 131.9017,
      "eval_samples_per_second": 4.791,
      "eval_steps_per_second": 0.599,
      "step": 2620
    },
    {
      "epoch": 4.1613924050632916,
      "grad_norm": 0.28812527656555176,
      "learning_rate": 0.0002,
      "loss": 0.5209,
      "step": 2630
    },
    {
      "epoch": 4.177215189873418,
      "grad_norm": 0.28006505966186523,
      "learning_rate": 0.0002,
      "loss": 0.5231,
      "step": 2640
    },
    {
      "epoch": 4.177215189873418,
      "eval_loss": 0.6253841519355774,
      "eval_runtime": 132.022,
      "eval_samples_per_second": 4.787,
      "eval_steps_per_second": 0.598,
      "step": 2640
    },
    {
      "epoch": 4.193037974683544,
      "grad_norm": 0.27704644203186035,
      "learning_rate": 0.0002,
      "loss": 0.5351,
      "step": 2650
    },
    {
      "epoch": 4.208860759493671,
      "grad_norm": 0.2679150402545929,
      "learning_rate": 0.0002,
      "loss": 0.5355,
      "step": 2660
    },
    {
      "epoch": 4.208860759493671,
      "eval_loss": 0.6239191889762878,
      "eval_runtime": 131.8844,
      "eval_samples_per_second": 4.792,
      "eval_steps_per_second": 0.599,
      "step": 2660
    },
    {
      "epoch": 4.224683544303797,
      "grad_norm": 0.29690659046173096,
      "learning_rate": 0.0002,
      "loss": 0.5328,
      "step": 2670
    },
    {
      "epoch": 4.2405063291139244,
      "grad_norm": 0.2684895098209381,
      "learning_rate": 0.0002,
      "loss": 0.5213,
      "step": 2680
    },
    {
      "epoch": 4.2405063291139244,
      "eval_loss": 0.626095175743103,
      "eval_runtime": 131.9589,
      "eval_samples_per_second": 4.789,
      "eval_steps_per_second": 0.599,
      "step": 2680
    },
    {
      "epoch": 4.256329113924051,
      "grad_norm": 0.2884593904018402,
      "learning_rate": 0.0002,
      "loss": 0.5345,
      "step": 2690
    },
    {
      "epoch": 4.272151898734177,
      "grad_norm": 0.30268314480781555,
      "learning_rate": 0.0002,
      "loss": 0.5443,
      "step": 2700
    },
    {
      "epoch": 4.272151898734177,
      "eval_loss": 0.6236627101898193,
      "eval_runtime": 131.9384,
      "eval_samples_per_second": 4.79,
      "eval_steps_per_second": 0.599,
      "step": 2700
    },
    {
      "epoch": 4.287974683544304,
      "grad_norm": 0.2858387231826782,
      "learning_rate": 0.0002,
      "loss": 0.5287,
      "step": 2710
    },
    {
      "epoch": 4.30379746835443,
      "grad_norm": 0.3007354438304901,
      "learning_rate": 0.0002,
      "loss": 0.5333,
      "step": 2720
    },
    {
      "epoch": 4.30379746835443,
      "eval_loss": 0.624500036239624,
      "eval_runtime": 131.9182,
      "eval_samples_per_second": 4.791,
      "eval_steps_per_second": 0.599,
      "step": 2720
    },
    {
      "epoch": 4.319620253164557,
      "grad_norm": 0.28278785943984985,
      "learning_rate": 0.0002,
      "loss": 0.5344,
      "step": 2730
    },
    {
      "epoch": 4.3354430379746836,
      "grad_norm": 0.30831849575042725,
      "learning_rate": 0.0002,
      "loss": 0.5422,
      "step": 2740
    },
    {
      "epoch": 4.3354430379746836,
      "eval_loss": 0.623884379863739,
      "eval_runtime": 131.8758,
      "eval_samples_per_second": 4.792,
      "eval_steps_per_second": 0.599,
      "step": 2740
    },
    {
      "epoch": 4.35126582278481,
      "grad_norm": 0.27501291036605835,
      "learning_rate": 0.0002,
      "loss": 0.5347,
      "step": 2750
    },
    {
      "epoch": 4.367088607594937,
      "grad_norm": 0.29319167137145996,
      "learning_rate": 0.0002,
      "loss": 0.5324,
      "step": 2760
    },
    {
      "epoch": 4.367088607594937,
      "eval_loss": 0.6249279379844666,
      "eval_runtime": 131.9025,
      "eval_samples_per_second": 4.791,
      "eval_steps_per_second": 0.599,
      "step": 2760
    },
    {
      "epoch": 4.382911392405063,
      "grad_norm": 0.28570714592933655,
      "learning_rate": 0.0002,
      "loss": 0.5378,
      "step": 2770
    },
    {
      "epoch": 4.39873417721519,
      "grad_norm": 0.28097033500671387,
      "learning_rate": 0.0002,
      "loss": 0.5402,
      "step": 2780
    },
    {
      "epoch": 4.39873417721519,
      "eval_loss": 0.6221233606338501,
      "eval_runtime": 131.8934,
      "eval_samples_per_second": 4.792,
      "eval_steps_per_second": 0.599,
      "step": 2780
    },
    {
      "epoch": 4.4145569620253164,
      "grad_norm": 0.29954394698143005,
      "learning_rate": 0.0002,
      "loss": 0.5336,
      "step": 2790
    },
    {
      "epoch": 4.430379746835443,
      "grad_norm": 0.3126460313796997,
      "learning_rate": 0.0002,
      "loss": 0.5249,
      "step": 2800
    },
    {
      "epoch": 4.430379746835443,
      "eval_loss": 0.6213365197181702,
      "eval_runtime": 131.8655,
      "eval_samples_per_second": 4.793,
      "eval_steps_per_second": 0.599,
      "step": 2800
    },
    {
      "epoch": 4.44620253164557,
      "grad_norm": 0.31044065952301025,
      "learning_rate": 0.0002,
      "loss": 0.5292,
      "step": 2810
    },
    {
      "epoch": 4.462025316455696,
      "grad_norm": 0.29306626319885254,
      "learning_rate": 0.0002,
      "loss": 0.534,
      "step": 2820
    },
    {
      "epoch": 4.462025316455696,
      "eval_loss": 0.6223319172859192,
      "eval_runtime": 131.8462,
      "eval_samples_per_second": 4.793,
      "eval_steps_per_second": 0.599,
      "step": 2820
    },
    {
      "epoch": 4.477848101265823,
      "grad_norm": 0.28371503949165344,
      "learning_rate": 0.0002,
      "loss": 0.5307,
      "step": 2830
    },
    {
      "epoch": 4.493670886075949,
      "grad_norm": 0.3208947479724884,
      "learning_rate": 0.0002,
      "loss": 0.5291,
      "step": 2840
    },
    {
      "epoch": 4.493670886075949,
      "eval_loss": 0.621949315071106,
      "eval_runtime": 131.8523,
      "eval_samples_per_second": 4.793,
      "eval_steps_per_second": 0.599,
      "step": 2840
    },
    {
      "epoch": 4.509493670886076,
      "grad_norm": 0.2842102646827698,
      "learning_rate": 0.0002,
      "loss": 0.5213,
      "step": 2850
    },
    {
      "epoch": 4.525316455696203,
      "grad_norm": 0.28963133692741394,
      "learning_rate": 0.0002,
      "loss": 0.5337,
      "step": 2860
    },
    {
      "epoch": 4.525316455696203,
      "eval_loss": 0.6212794184684753,
      "eval_runtime": 131.8857,
      "eval_samples_per_second": 4.792,
      "eval_steps_per_second": 0.599,
      "step": 2860
    },
    {
      "epoch": 4.541139240506329,
      "grad_norm": 0.2925531268119812,
      "learning_rate": 0.0002,
      "loss": 0.5479,
      "step": 2870
    },
    {
      "epoch": 4.556962025316456,
      "grad_norm": 0.3018389642238617,
      "learning_rate": 0.0002,
      "loss": 0.5444,
      "step": 2880
    },
    {
      "epoch": 4.556962025316456,
      "eval_loss": 0.6203261017799377,
      "eval_runtime": 131.869,
      "eval_samples_per_second": 4.793,
      "eval_steps_per_second": 0.599,
      "step": 2880
    },
    {
      "epoch": 4.572784810126582,
      "grad_norm": 0.3160776197910309,
      "learning_rate": 0.0002,
      "loss": 0.5338,
      "step": 2890
    },
    {
      "epoch": 4.588607594936709,
      "grad_norm": 0.28098049759864807,
      "learning_rate": 0.0002,
      "loss": 0.5415,
      "step": 2900
    },
    {
      "epoch": 4.588607594936709,
      "eval_loss": 0.6205925941467285,
      "eval_runtime": 131.8978,
      "eval_samples_per_second": 4.792,
      "eval_steps_per_second": 0.599,
      "step": 2900
    },
    {
      "epoch": 4.6044303797468356,
      "grad_norm": 0.2892148196697235,
      "learning_rate": 0.0002,
      "loss": 0.5357,
      "step": 2910
    },
    {
      "epoch": 4.620253164556962,
      "grad_norm": 0.2922069728374481,
      "learning_rate": 0.0002,
      "loss": 0.5341,
      "step": 2920
    },
    {
      "epoch": 4.620253164556962,
      "eval_loss": 0.6200461387634277,
      "eval_runtime": 131.8759,
      "eval_samples_per_second": 4.792,
      "eval_steps_per_second": 0.599,
      "step": 2920
    },
    {
      "epoch": 4.636075949367089,
      "grad_norm": 0.2866216003894806,
      "learning_rate": 0.0002,
      "loss": 0.5295,
      "step": 2930
    },
    {
      "epoch": 4.651898734177215,
      "grad_norm": 0.30505985021591187,
      "learning_rate": 0.0002,
      "loss": 0.5309,
      "step": 2940
    },
    {
      "epoch": 4.651898734177215,
      "eval_loss": 0.6205445528030396,
      "eval_runtime": 131.8601,
      "eval_samples_per_second": 4.793,
      "eval_steps_per_second": 0.599,
      "step": 2940
    },
    {
      "epoch": 4.667721518987342,
      "grad_norm": 0.2945426106452942,
      "learning_rate": 0.0002,
      "loss": 0.5444,
      "step": 2950
    },
    {
      "epoch": 4.6835443037974684,
      "grad_norm": 0.3077249526977539,
      "learning_rate": 0.0002,
      "loss": 0.536,
      "step": 2960
    },
    {
      "epoch": 4.6835443037974684,
      "eval_loss": 0.6187692284584045,
      "eval_runtime": 131.8257,
      "eval_samples_per_second": 4.794,
      "eval_steps_per_second": 0.599,
      "step": 2960
    },
    {
      "epoch": 4.699367088607595,
      "grad_norm": 0.29698699712753296,
      "learning_rate": 0.0002,
      "loss": 0.5379,
      "step": 2970
    },
    {
      "epoch": 4.715189873417722,
      "grad_norm": 0.2893849313259125,
      "learning_rate": 0.0002,
      "loss": 0.545,
      "step": 2980
    },
    {
      "epoch": 4.715189873417722,
      "eval_loss": 0.6184265613555908,
      "eval_runtime": 131.8651,
      "eval_samples_per_second": 4.793,
      "eval_steps_per_second": 0.599,
      "step": 2980
    },
    {
      "epoch": 4.731012658227848,
      "grad_norm": 0.3033924400806427,
      "learning_rate": 0.0002,
      "loss": 0.5414,
      "step": 2990
    },
    {
      "epoch": 4.746835443037975,
      "grad_norm": 0.2796303629875183,
      "learning_rate": 0.0002,
      "loss": 0.5487,
      "step": 3000
    },
    {
      "epoch": 4.746835443037975,
      "eval_loss": 0.6157076954841614,
      "eval_runtime": 131.8771,
      "eval_samples_per_second": 4.792,
      "eval_steps_per_second": 0.599,
      "step": 3000
    },
    {
      "epoch": 4.762658227848101,
      "grad_norm": 0.30320191383361816,
      "learning_rate": 0.0002,
      "loss": 0.5375,
      "step": 3010
    },
    {
      "epoch": 4.7784810126582276,
      "grad_norm": 0.29859721660614014,
      "learning_rate": 0.0002,
      "loss": 0.5292,
      "step": 3020
    },
    {
      "epoch": 4.7784810126582276,
      "eval_loss": 0.6158034205436707,
      "eval_runtime": 131.8997,
      "eval_samples_per_second": 4.792,
      "eval_steps_per_second": 0.599,
      "step": 3020
    },
    {
      "epoch": 4.794303797468355,
      "grad_norm": 0.3109189569950104,
      "learning_rate": 0.0002,
      "loss": 0.5423,
      "step": 3030
    },
    {
      "epoch": 4.810126582278481,
      "grad_norm": 0.27828630805015564,
      "learning_rate": 0.0002,
      "loss": 0.5341,
      "step": 3040
    },
    {
      "epoch": 4.810126582278481,
      "eval_loss": 0.6152633428573608,
      "eval_runtime": 131.9276,
      "eval_samples_per_second": 4.791,
      "eval_steps_per_second": 0.599,
      "step": 3040
    },
    {
      "epoch": 4.825949367088608,
      "grad_norm": 0.2859291434288025,
      "learning_rate": 0.0002,
      "loss": 0.5421,
      "step": 3050
    },
    {
      "epoch": 4.841772151898734,
      "grad_norm": 0.2746394872665405,
      "learning_rate": 0.0002,
      "loss": 0.546,
      "step": 3060
    },
    {
      "epoch": 4.841772151898734,
      "eval_loss": 0.6155226230621338,
      "eval_runtime": 131.9573,
      "eval_samples_per_second": 4.789,
      "eval_steps_per_second": 0.599,
      "step": 3060
    },
    {
      "epoch": 4.8575949367088604,
      "grad_norm": 0.2710588574409485,
      "learning_rate": 0.0002,
      "loss": 0.5502,
      "step": 3070
    },
    {
      "epoch": 4.8734177215189876,
      "grad_norm": 0.29007774591445923,
      "learning_rate": 0.0002,
      "loss": 0.5401,
      "step": 3080
    },
    {
      "epoch": 4.8734177215189876,
      "eval_loss": 0.6148353219032288,
      "eval_runtime": 131.9271,
      "eval_samples_per_second": 4.791,
      "eval_steps_per_second": 0.599,
      "step": 3080
    },
    {
      "epoch": 4.889240506329114,
      "grad_norm": 0.27158820629119873,
      "learning_rate": 0.0002,
      "loss": 0.5488,
      "step": 3090
    },
    {
      "epoch": 4.905063291139241,
      "grad_norm": 0.2922937572002411,
      "learning_rate": 0.0002,
      "loss": 0.5388,
      "step": 3100
    },
    {
      "epoch": 4.905063291139241,
      "eval_loss": 0.6152787804603577,
      "eval_runtime": 131.8914,
      "eval_samples_per_second": 4.792,
      "eval_steps_per_second": 0.599,
      "step": 3100
    },
    {
      "epoch": 4.920886075949367,
      "grad_norm": 0.28401508927345276,
      "learning_rate": 0.0002,
      "loss": 0.5305,
      "step": 3110
    },
    {
      "epoch": 4.936708860759493,
      "grad_norm": 0.29028162360191345,
      "learning_rate": 0.0002,
      "loss": 0.5353,
      "step": 3120
    },
    {
      "epoch": 4.936708860759493,
      "eval_loss": 0.6147912740707397,
      "eval_runtime": 131.9091,
      "eval_samples_per_second": 4.791,
      "eval_steps_per_second": 0.599,
      "step": 3120
    },
    {
      "epoch": 4.9525316455696204,
      "grad_norm": 0.2920195162296295,
      "learning_rate": 0.0002,
      "loss": 0.5384,
      "step": 3130
    },
    {
      "epoch": 4.968354430379747,
      "grad_norm": 0.33459678292274475,
      "learning_rate": 0.0002,
      "loss": 0.5349,
      "step": 3140
    },
    {
      "epoch": 4.968354430379747,
      "eval_loss": 0.6154718399047852,
      "eval_runtime": 131.9168,
      "eval_samples_per_second": 4.791,
      "eval_steps_per_second": 0.599,
      "step": 3140
    },
    {
      "epoch": 4.984177215189874,
      "grad_norm": 0.2918411195278168,
      "learning_rate": 0.0002,
      "loss": 0.5383,
      "step": 3150
    },
    {
      "epoch": 5.0,
      "grad_norm": 0.2778303623199463,
      "learning_rate": 0.0002,
      "loss": 0.5371,
      "step": 3160
    },
    {
      "epoch": 5.0,
      "eval_loss": 0.6133139133453369,
      "eval_runtime": 131.8981,
      "eval_samples_per_second": 4.792,
      "eval_steps_per_second": 0.599,
      "step": 3160
    }
  ],
  "logging_steps": 10,
  "max_steps": 3160,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 5,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 1.108664495678423e+18,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}