{
  "best_global_step": 210,
  "best_metric": 0.34583956003189087,
  "best_model_checkpoint": "saves/prefix-tuning/llama-3-8b-instruct/train_wsc_42_1760466772/checkpoint-210",
  "epoch": 30.0,
  "eval_steps": 42,
  "global_step": 840,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.17857142857142858,
      "grad_norm": 175.3666534423828,
      "learning_rate": 3.1746031746031746e-06,
      "loss": 11.3214,
      "num_input_tokens_seen": 9216,
      "step": 5
    },
    {
      "epoch": 0.35714285714285715,
      "grad_norm": 145.5986785888672,
      "learning_rate": 7.142857142857143e-06,
      "loss": 6.8093,
      "num_input_tokens_seen": 18112,
      "step": 10
    },
    {
      "epoch": 0.5357142857142857,
      "grad_norm": 58.49702453613281,
      "learning_rate": 1.1111111111111112e-05,
      "loss": 2.1384,
      "num_input_tokens_seen": 26624,
      "step": 15
    },
    {
      "epoch": 0.7142857142857143,
      "grad_norm": 57.37192153930664,
      "learning_rate": 1.5079365079365079e-05,
      "loss": 0.7391,
      "num_input_tokens_seen": 35456,
      "step": 20
    },
    {
      "epoch": 0.8928571428571429,
      "grad_norm": 20.545907974243164,
      "learning_rate": 1.9047619047619046e-05,
      "loss": 0.5556,
      "num_input_tokens_seen": 44096,
      "step": 25
    },
    {
      "epoch": 1.0714285714285714,
      "grad_norm": 25.21103286743164,
      "learning_rate": 2.3015873015873015e-05,
      "loss": 0.5487,
      "num_input_tokens_seen": 52192,
      "step": 30
    },
    {
      "epoch": 1.25,
      "grad_norm": 8.62804126739502,
      "learning_rate": 2.6984126984126984e-05,
      "loss": 0.4509,
      "num_input_tokens_seen": 60896,
      "step": 35
    },
    {
      "epoch": 1.4285714285714286,
      "grad_norm": 23.560665130615234,
      "learning_rate": 3.095238095238095e-05,
      "loss": 0.4818,
      "num_input_tokens_seen": 69920,
      "step": 40
    },
    {
      "epoch": 1.5,
      "eval_loss": 0.3624400794506073,
      "eval_runtime": 1.118,
      "eval_samples_per_second": 99.287,
      "eval_steps_per_second": 12.523,
      "num_input_tokens_seen": 73824,
      "step": 42
    },
    {
      "epoch": 1.6071428571428572,
      "grad_norm": 13.207117080688477,
      "learning_rate": 3.492063492063492e-05,
      "loss": 0.4103,
      "num_input_tokens_seen": 78560,
      "step": 45
    },
    {
      "epoch": 1.7857142857142856,
      "grad_norm": 11.927769660949707,
      "learning_rate": 3.888888888888889e-05,
      "loss": 0.4405,
      "num_input_tokens_seen": 87136,
      "step": 50
    },
    {
      "epoch": 1.9642857142857144,
      "grad_norm": 7.089166641235352,
      "learning_rate": 4.2857142857142856e-05,
      "loss": 0.408,
      "num_input_tokens_seen": 96480,
      "step": 55
    },
    {
      "epoch": 2.142857142857143,
      "grad_norm": 2.959946393966675,
      "learning_rate": 4.682539682539683e-05,
      "loss": 0.3548,
      "num_input_tokens_seen": 104248,
      "step": 60
    },
    {
      "epoch": 2.3214285714285716,
      "grad_norm": 1.298177719116211,
      "learning_rate": 5.0793650793650794e-05,
      "loss": 0.377,
      "num_input_tokens_seen": 113272,
      "step": 65
    },
    {
      "epoch": 2.5,
      "grad_norm": 8.108621597290039,
      "learning_rate": 5.4761904761904766e-05,
      "loss": 0.3922,
      "num_input_tokens_seen": 122168,
      "step": 70
    },
    {
      "epoch": 2.678571428571429,
      "grad_norm": 2.4724233150482178,
      "learning_rate": 5.873015873015873e-05,
      "loss": 0.4379,
      "num_input_tokens_seen": 131576,
      "step": 75
    },
    {
      "epoch": 2.857142857142857,
      "grad_norm": 4.108138084411621,
      "learning_rate": 6.26984126984127e-05,
      "loss": 0.3847,
      "num_input_tokens_seen": 139832,
      "step": 80
    },
    {
      "epoch": 3.0,
      "eval_loss": 0.3848850131034851,
      "eval_runtime": 1.1397,
      "eval_samples_per_second": 97.398,
      "eval_steps_per_second": 12.284,
      "num_input_tokens_seen": 146552,
      "step": 84
    },
    {
      "epoch": 3.0357142857142856,
      "grad_norm": 0.5216990113258362,
      "learning_rate": 6.666666666666667e-05,
      "loss": 0.3722,
      "num_input_tokens_seen": 147832,
      "step": 85
    },
    {
      "epoch": 3.2142857142857144,
      "grad_norm": 0.9133898615837097,
      "learning_rate": 7.063492063492065e-05,
      "loss": 0.4223,
      "num_input_tokens_seen": 157176,
      "step": 90
    },
    {
      "epoch": 3.392857142857143,
      "grad_norm": 4.493597030639648,
      "learning_rate": 7.460317460317461e-05,
      "loss": 0.388,
      "num_input_tokens_seen": 166392,
      "step": 95
    },
    {
      "epoch": 3.571428571428571,
      "grad_norm": 0.2193046361207962,
      "learning_rate": 7.857142857142858e-05,
      "loss": 0.3655,
      "num_input_tokens_seen": 174904,
      "step": 100
    },
    {
      "epoch": 3.75,
      "grad_norm": 3.4654486179351807,
      "learning_rate": 8.253968253968255e-05,
      "loss": 0.4285,
      "num_input_tokens_seen": 184312,
      "step": 105
    },
    {
      "epoch": 3.928571428571429,
      "grad_norm": 0.2327776700258255,
      "learning_rate": 8.650793650793651e-05,
      "loss": 0.3877,
      "num_input_tokens_seen": 193464,
      "step": 110
    },
    {
      "epoch": 4.107142857142857,
      "grad_norm": 0.6309562921524048,
      "learning_rate": 9.047619047619048e-05,
      "loss": 0.3744,
      "num_input_tokens_seen": 201168,
      "step": 115
    },
    {
      "epoch": 4.285714285714286,
      "grad_norm": 0.7549311518669128,
      "learning_rate": 9.444444444444444e-05,
      "loss": 0.4027,
      "num_input_tokens_seen": 210384,
      "step": 120
    },
    {
      "epoch": 4.464285714285714,
      "grad_norm": 0.4532535672187805,
      "learning_rate": 9.841269841269841e-05,
      "loss": 0.3464,
      "num_input_tokens_seen": 219536,
      "step": 125
    },
    {
      "epoch": 4.5,
      "eval_loss": 0.34812429547309875,
      "eval_runtime": 1.0994,
      "eval_samples_per_second": 100.966,
      "eval_steps_per_second": 12.735,
      "num_input_tokens_seen": 221264,
      "step": 126
    },
    {
      "epoch": 4.642857142857143,
      "grad_norm": 0.14779126644134521,
      "learning_rate": 9.999564408362054e-05,
      "loss": 0.333,
      "num_input_tokens_seen": 228432,
      "step": 130
    },
    {
      "epoch": 4.821428571428571,
      "grad_norm": 0.35166576504707336,
      "learning_rate": 9.996902734308346e-05,
      "loss": 0.3557,
      "num_input_tokens_seen": 238032,
      "step": 135
    },
    {
      "epoch": 5.0,
      "grad_norm": 0.6182653307914734,
      "learning_rate": 9.991822668185927e-05,
      "loss": 0.3749,
      "num_input_tokens_seen": 245376,
      "step": 140
    },
    {
      "epoch": 5.178571428571429,
      "grad_norm": 0.5397220253944397,
      "learning_rate": 9.984326668636131e-05,
      "loss": 0.3563,
      "num_input_tokens_seen": 253952,
      "step": 145
    },
    {
      "epoch": 5.357142857142857,
      "grad_norm": 0.11434465646743774,
      "learning_rate": 9.974418363559444e-05,
      "loss": 0.3447,
      "num_input_tokens_seen": 263296,
      "step": 150
    },
    {
      "epoch": 5.535714285714286,
      "grad_norm": 0.9336232542991638,
      "learning_rate": 9.96210254835968e-05,
      "loss": 0.3606,
      "num_input_tokens_seen": 271808,
      "step": 155
    },
    {
      "epoch": 5.714285714285714,
      "grad_norm": 0.10893812775611877,
      "learning_rate": 9.947385183623098e-05,
      "loss": 0.3507,
      "num_input_tokens_seen": 280704,
      "step": 160
    },
    {
      "epoch": 5.892857142857143,
      "grad_norm": 0.15622828900814056,
      "learning_rate": 9.930273392233624e-05,
      "loss": 0.3578,
      "num_input_tokens_seen": 289600,
      "step": 165
    },
    {
      "epoch": 6.0,
      "eval_loss": 0.37425312399864197,
      "eval_runtime": 1.2693,
      "eval_samples_per_second": 87.447,
      "eval_steps_per_second": 11.029,
      "num_input_tokens_seen": 294256,
      "step": 168
    },
    {
      "epoch": 6.071428571428571,
      "grad_norm": 0.267703652381897,
      "learning_rate": 9.910775455925518e-05,
      "loss": 0.3546,
      "num_input_tokens_seen": 297520,
      "step": 170
    },
    {
      "epoch": 6.25,
      "grad_norm": 0.12959004938602448,
      "learning_rate": 9.888900811275204e-05,
      "loss": 0.3507,
      "num_input_tokens_seen": 306672,
      "step": 175
    },
    {
      "epoch": 6.428571428571429,
      "grad_norm": 0.25042709708213806,
      "learning_rate": 9.864660045134165e-05,
      "loss": 0.3929,
      "num_input_tokens_seen": 315824,
      "step": 180
    },
    {
      "epoch": 6.607142857142857,
      "grad_norm": 0.1262078881263733,
      "learning_rate": 9.838064889505141e-05,
      "loss": 0.35,
      "num_input_tokens_seen": 324464,
      "step": 185
    },
    {
      "epoch": 6.785714285714286,
      "grad_norm": 0.2343859225511551,
      "learning_rate": 9.809128215864097e-05,
      "loss": 0.36,
      "num_input_tokens_seen": 332976,
      "step": 190
    },
    {
      "epoch": 6.964285714285714,
      "grad_norm": 0.2336055040359497,
      "learning_rate": 9.777864028930705e-05,
      "loss": 0.3521,
      "num_input_tokens_seen": 341552,
      "step": 195
    },
    {
      "epoch": 7.142857142857143,
      "grad_norm": 0.34513550996780396,
      "learning_rate": 9.744287459890368e-05,
      "loss": 0.352,
      "num_input_tokens_seen": 349584,
      "step": 200
    },
    {
      "epoch": 7.321428571428571,
      "grad_norm": 0.5254360437393188,
      "learning_rate": 9.708414759071059e-05,
      "loss": 0.3669,
      "num_input_tokens_seen": 358672,
      "step": 205
    },
    {
      "epoch": 7.5,
      "grad_norm": 0.04271647334098816,
      "learning_rate": 9.670263288078502e-05,
      "loss": 0.3492,
      "num_input_tokens_seen": 368144,
      "step": 210
    },
    {
      "epoch": 7.5,
      "eval_loss": 0.34583956003189087,
      "eval_runtime": 1.1811,
      "eval_samples_per_second": 93.983,
      "eval_steps_per_second": 11.854,
      "num_input_tokens_seen": 368144,
      "step": 210
    },
    {
      "epoch": 7.678571428571429,
      "grad_norm": 0.2870471179485321,
      "learning_rate": 9.629851511393555e-05,
      "loss": 0.3575,
      "num_input_tokens_seen": 376464,
      "step": 215
    },
    {
      "epoch": 7.857142857142857,
      "grad_norm": 0.16233104467391968,
      "learning_rate": 9.587198987435782e-05,
      "loss": 0.347,
      "num_input_tokens_seen": 385616,
      "step": 220
    },
    {
      "epoch": 8.035714285714286,
      "grad_norm": 0.24467554688453674,
      "learning_rate": 9.542326359097619e-05,
      "loss": 0.3491,
      "num_input_tokens_seen": 393384,
      "step": 225
    },
    {
      "epoch": 8.214285714285714,
      "grad_norm": 0.2640959918498993,
      "learning_rate": 9.495255343753657e-05,
      "loss": 0.3562,
      "num_input_tokens_seen": 402280,
      "step": 230
    },
    {
      "epoch": 8.392857142857142,
      "grad_norm": 0.18736229836940765,
      "learning_rate": 9.446008722749905e-05,
      "loss": 0.3491,
      "num_input_tokens_seen": 410280,
      "step": 235
    },
    {
      "epoch": 8.571428571428571,
      "grad_norm": 0.039574023336172104,
      "learning_rate": 9.394610330378124e-05,
      "loss": 0.3404,
      "num_input_tokens_seen": 418856,
      "step": 240
    },
    {
      "epoch": 8.75,
      "grad_norm": 0.12487329542636871,
      "learning_rate": 9.341085042340532e-05,
      "loss": 0.3528,
      "num_input_tokens_seen": 427816,
      "step": 245
    },
    {
      "epoch": 8.928571428571429,
      "grad_norm": 0.08902257680892944,
      "learning_rate": 9.285458763710524e-05,
      "loss": 0.3542,
      "num_input_tokens_seen": 437352,
      "step": 250
    },
    {
      "epoch": 9.0,
      "eval_loss": 0.34945669770240784,
      "eval_runtime": 1.1316,
      "eval_samples_per_second": 98.092,
      "eval_steps_per_second": 12.372,
      "num_input_tokens_seen": 439768,
      "step": 252
    },
    {
      "epoch": 9.107142857142858,
      "grad_norm": 0.12130332738161087,
      "learning_rate": 9.227758416395169e-05,
      "loss": 0.3435,
      "num_input_tokens_seen": 444504,
      "step": 255
    },
    {
      "epoch": 9.285714285714286,
      "grad_norm": 0.14609932899475098,
      "learning_rate": 9.168011926105598e-05,
      "loss": 0.3565,
      "num_input_tokens_seen": 453144,
      "step": 260
    },
    {
      "epoch": 9.464285714285714,
      "grad_norm": 0.025746047496795654,
      "learning_rate": 9.106248208841569e-05,
      "loss": 0.3499,
      "num_input_tokens_seen": 462040,
      "step": 265
    },
    {
      "epoch": 9.642857142857142,
      "grad_norm": 0.10566066205501556,
      "learning_rate": 9.042497156896748e-05,
      "loss": 0.349,
      "num_input_tokens_seen": 471576,
      "step": 270
    },
    {
      "epoch": 9.821428571428571,
      "grad_norm": 0.2756125032901764,
      "learning_rate": 8.976789624391498e-05,
      "loss": 0.3532,
      "num_input_tokens_seen": 480280,
      "step": 275
    },
    {
      "epoch": 10.0,
      "grad_norm": 0.05616496875882149,
      "learning_rate": 8.90915741234015e-05,
      "loss": 0.3506,
      "num_input_tokens_seen": 489096,
      "step": 280
    },
    {
      "epoch": 10.178571428571429,
      "grad_norm": 0.23147229850292206,
      "learning_rate": 8.839633253260006e-05,
      "loss": 0.3475,
      "num_input_tokens_seen": 498952,
      "step": 285
    },
    {
      "epoch": 10.357142857142858,
      "grad_norm": 0.17449620366096497,
      "learning_rate": 8.768250795329518e-05,
      "loss": 0.34,
      "num_input_tokens_seen": 507976,
      "step": 290
    },
    {
      "epoch": 10.5,
      "eval_loss": 0.35035043954849243,
      "eval_runtime": 1.1011,
      "eval_samples_per_second": 100.804,
      "eval_steps_per_second": 12.714,
      "num_input_tokens_seen": 514888,
      "step": 294
    },
    {
      "epoch": 10.535714285714286,
      "grad_norm": 0.09387495368719101,
      "learning_rate": 8.695044586103296e-05,
      "loss": 0.355,
      "num_input_tokens_seen": 516744,
      "step": 295
    },
    {
      "epoch": 10.714285714285714,
      "grad_norm": 0.269589364528656,
      "learning_rate": 8.620050055791851e-05,
      "loss": 0.3472,
      "num_input_tokens_seen": 525960,
      "step": 300
    },
    {
      "epoch": 10.892857142857142,
      "grad_norm": 0.045772165060043335,
      "learning_rate": 8.543303500114141e-05,
      "loss": 0.3496,
      "num_input_tokens_seen": 534344,
      "step": 305
    },
    {
      "epoch": 11.071428571428571,
      "grad_norm": 0.23182272911071777,
      "learning_rate": 8.464842062731235e-05,
      "loss": 0.3481,
      "num_input_tokens_seen": 541856,
      "step": 310
    },
    {
      "epoch": 11.25,
      "grad_norm": 0.34456300735473633,
      "learning_rate": 8.384703717269584e-05,
      "loss": 0.3448,
      "num_input_tokens_seen": 550176,
      "step": 315
    },
    {
      "epoch": 11.428571428571429,
      "grad_norm": 0.04894077032804489,
      "learning_rate": 8.302927248942627e-05,
      "loss": 0.3405,
      "num_input_tokens_seen": 558368,
      "step": 320
    },
    {
      "epoch": 11.607142857142858,
      "grad_norm": 0.33567559719085693,
      "learning_rate": 8.219552235779578e-05,
      "loss": 0.3466,
      "num_input_tokens_seen": 567392,
      "step": 325
    },
    {
      "epoch": 11.785714285714286,
      "grad_norm": 0.2454521507024765,
      "learning_rate": 8.134619029470534e-05,
      "loss": 0.3594,
      "num_input_tokens_seen": 576864,
      "step": 330
    },
    {
      "epoch": 11.964285714285714,
      "grad_norm": 0.41574931144714355,
      "learning_rate": 8.048168735837121e-05,
      "loss": 0.3578,
      "num_input_tokens_seen": 585376,
      "step": 335
    },
    {
      "epoch": 12.0,
      "eval_loss": 0.3650008738040924,
      "eval_runtime": 1.1362,
      "eval_samples_per_second": 97.697,
      "eval_steps_per_second": 12.322,
      "num_input_tokens_seen": 586448,
      "step": 336
    },
    {
      "epoch": 12.142857142857142,
      "grad_norm": 0.13403870165348053,
      "learning_rate": 7.960243194938192e-05,
      "loss": 0.3703,
      "num_input_tokens_seen": 593296,
      "step": 340
    },
    {
      "epoch": 12.321428571428571,
      "grad_norm": 0.03592513129115105,
      "learning_rate": 7.87088496082013e-05,
      "loss": 0.3383,
      "num_input_tokens_seen": 602000,
      "step": 345
    },
    {
      "epoch": 12.5,
      "grad_norm": 0.30938151478767395,
      "learning_rate": 7.780137280921636e-05,
      "loss": 0.3849,
      "num_input_tokens_seen": 610960,
      "step": 350
    },
    {
      "epoch": 12.678571428571429,
      "grad_norm": 0.047836095094680786,
      "learning_rate": 7.688044075142887e-05,
      "loss": 0.3439,
      "num_input_tokens_seen": 619984,
      "step": 355
    },
    {
      "epoch": 12.857142857142858,
      "grad_norm": 0.08763613551855087,
      "learning_rate": 7.594649914589287e-05,
      "loss": 0.3448,
      "num_input_tokens_seen": 629776,
      "step": 360
    },
    {
      "epoch": 13.035714285714286,
      "grad_norm": 0.0415462851524353,
      "learning_rate": 7.500000000000001e-05,
      "loss": 0.348,
      "num_input_tokens_seen": 637888,
      "step": 365
    },
    {
      "epoch": 13.214285714285714,
      "grad_norm": 0.12538637220859528,
      "learning_rate": 7.404140139871797e-05,
      "loss": 0.345,
      "num_input_tokens_seen": 648128,
      "step": 370
    },
    {
      "epoch": 13.392857142857142,
      "grad_norm": 0.17965561151504517,
      "learning_rate": 7.307116728288727e-05,
      "loss": 0.3462,
      "num_input_tokens_seen": 656768,
      "step": 375
    },
    {
      "epoch": 13.5,
      "eval_loss": 0.3538319170475006,
      "eval_runtime": 1.1141,
      "eval_samples_per_second": 99.628,
      "eval_steps_per_second": 12.566,
      "num_input_tokens_seen": 662016,
      "step": 378
    },
    {
      "epoch": 13.571428571428571,
      "grad_norm": 0.22999395430088043,
      "learning_rate": 7.208976722468392e-05,
      "loss": 0.3606,
      "num_input_tokens_seen": 665472,
      "step": 380
    },
    {
      "epoch": 13.75,
      "grad_norm": 0.0893191397190094,
      "learning_rate": 7.109767620035689e-05,
      "loss": 0.3501,
      "num_input_tokens_seen": 673664,
      "step": 385
    },
    {
      "epoch": 13.928571428571429,
      "grad_norm": 0.09730294346809387,
      "learning_rate": 7.00953743603498e-05,
      "loss": 0.3461,
      "num_input_tokens_seen": 682688,
      "step": 390
    },
    {
      "epoch": 14.107142857142858,
      "grad_norm": 0.05350238084793091,
      "learning_rate": 6.908334679691863e-05,
      "loss": 0.3522,
      "num_input_tokens_seen": 690936,
      "step": 395
    },
    {
      "epoch": 14.285714285714286,
      "grad_norm": 0.04386021941900253,
      "learning_rate": 6.806208330935766e-05,
      "loss": 0.3553,
      "num_input_tokens_seen": 700536,
      "step": 400
    },
    {
      "epoch": 14.464285714285714,
      "grad_norm": 0.0879250317811966,
      "learning_rate": 6.703207816694719e-05,
      "loss": 0.3382,
      "num_input_tokens_seen": 709560,
      "step": 405
    },
    {
      "epoch": 14.642857142857142,
      "grad_norm": 0.31405237317085266,
      "learning_rate": 6.599382986973808e-05,
      "loss": 0.348,
      "num_input_tokens_seen": 718584,
      "step": 410
    },
    {
      "epoch": 14.821428571428571,
      "grad_norm": 0.3567558228969574,
      "learning_rate": 6.494784090728852e-05,
      "loss": 0.3544,
      "num_input_tokens_seen": 727416,
      "step": 415
    },
    {
      "epoch": 15.0,
      "grad_norm": 0.061139896512031555,
      "learning_rate": 6.389461751547008e-05,
      "loss": 0.3506,
      "num_input_tokens_seen": 735680,
      "step": 420
    },
    {
      "epoch": 15.0,
      "eval_loss": 0.35569673776626587,
      "eval_runtime": 1.1923,
      "eval_samples_per_second": 93.099,
      "eval_steps_per_second": 11.742,
      "num_input_tokens_seen": 735680,
      "step": 420
    },
    {
      "epoch": 15.178571428571429,
      "grad_norm": 0.19610817730426788,
      "learning_rate": 6.283466943146053e-05,
      "loss": 0.3486,
      "num_input_tokens_seen": 744832,
      "step": 425
    },
    {
      "epoch": 15.357142857142858,
      "grad_norm": 0.05017664283514023,
      "learning_rate": 6.176850964704213e-05,
      "loss": 0.3573,
      "num_input_tokens_seen": 753728,
      "step": 430
    },
    {
      "epoch": 15.535714285714286,
      "grad_norm": 0.047355543822050095,
      "learning_rate": 6.069665416032487e-05,
      "loss": 0.3564,
      "num_input_tokens_seen": 762752,
      "step": 435
    },
    {
      "epoch": 15.714285714285714,
      "grad_norm": 0.16390874981880188,
      "learning_rate": 5.961962172601458e-05,
      "loss": 0.3461,
      "num_input_tokens_seen": 771200,
      "step": 440
    },
    {
      "epoch": 15.892857142857142,
      "grad_norm": 0.19481535255908966,
      "learning_rate": 5.853793360434687e-05,
      "loss": 0.3564,
      "num_input_tokens_seen": 779776,
      "step": 445
    },
    {
      "epoch": 16.071428571428573,
      "grad_norm": 0.04299188032746315,
      "learning_rate": 5.745211330880872e-05,
      "loss": 0.3214,
      "num_input_tokens_seen": 788984,
      "step": 450
    },
    {
      "epoch": 16.25,
      "grad_norm": 0.07021050900220871,
      "learning_rate": 5.636268635276918e-05,
      "loss": 0.355,
      "num_input_tokens_seen": 798456,
      "step": 455
    },
    {
      "epoch": 16.428571428571427,
      "grad_norm": 0.2501037120819092,
      "learning_rate": 5.527017999514239e-05,
      "loss": 0.3489,
      "num_input_tokens_seen": 807352,
      "step": 460
    },
    {
      "epoch": 16.5,
      "eval_loss": 0.35193607211112976,
      "eval_runtime": 1.1402,
      "eval_samples_per_second": 97.356,
      "eval_steps_per_second": 12.279,
      "num_input_tokens_seen": 810232,
      "step": 462
    },
    {
      "epoch": 16.607142857142858,
      "grad_norm": 0.08860552310943604,
      "learning_rate": 5.417512298520585e-05,
      "loss": 0.3506,
      "num_input_tokens_seen": 814712,
      "step": 465
    },
    {
      "epoch": 16.785714285714285,
      "grad_norm": 0.08465074002742767,
      "learning_rate": 5.307804530669716e-05,
      "loss": 0.346,
      "num_input_tokens_seen": 823608,
      "step": 470
    },
    {
      "epoch": 16.964285714285715,
      "grad_norm": 0.14993856847286224,
      "learning_rate": 5.197947792131348e-05,
      "loss": 0.3467,
      "num_input_tokens_seen": 832888,
      "step": 475
    },
    {
      "epoch": 17.142857142857142,
      "grad_norm": 0.17248603701591492,
      "learning_rate": 5.0879952511737696e-05,
      "loss": 0.3452,
      "num_input_tokens_seen": 840464,
      "step": 480
    },
    {
      "epoch": 17.321428571428573,
      "grad_norm": 0.10642991960048676,
      "learning_rate": 4.97800012243155e-05,
      "loss": 0.3402,
      "num_input_tokens_seen": 849488,
      "step": 485
    },
    {
      "epoch": 17.5,
      "grad_norm": 0.07554838806390762,
      "learning_rate": 4.86801564115082e-05,
      "loss": 0.3428,
      "num_input_tokens_seen": 858448,
      "step": 490
    },
    {
      "epoch": 17.678571428571427,
      "grad_norm": 0.1515018343925476,
      "learning_rate": 4.758095037424567e-05,
      "loss": 0.352,
      "num_input_tokens_seen": 867280,
      "step": 495
    },
    {
      "epoch": 17.857142857142858,
      "grad_norm": 0.2787249684333801,
      "learning_rate": 4.648291510430438e-05,
      "loss": 0.3528,
      "num_input_tokens_seen": 876880,
      "step": 500
    },
    {
      "epoch": 18.0,
      "eval_loss": 0.3557737469673157,
      "eval_runtime": 1.1986,
      "eval_samples_per_second": 92.609,
      "eval_steps_per_second": 11.68,
      "num_input_tokens_seen": 882920,
      "step": 504
    },
    {
      "epoch": 18.035714285714285,
      "grad_norm": 0.21712534129619598,
      "learning_rate": 4.5386582026834906e-05,
      "loss": 0.3418,
      "num_input_tokens_seen": 885480,
      "step": 505
    },
    {
      "epoch": 18.214285714285715,
      "grad_norm": 0.05514955148100853,
      "learning_rate": 4.4292481743163755e-05,
      "loss": 0.3441,
      "num_input_tokens_seen": 893864,
      "step": 510
    },
    {
      "epoch": 18.392857142857142,
      "grad_norm": 0.06789080053567886,
      "learning_rate": 4.3201143773993865e-05,
      "loss": 0.3524,
      "num_input_tokens_seen": 901928,
      "step": 515
    },
    {
      "epoch": 18.571428571428573,
      "grad_norm": 0.11630789190530777,
      "learning_rate": 4.2113096303128125e-05,
      "loss": 0.3495,
      "num_input_tokens_seen": 910696,
      "step": 520
    },
    {
      "epoch": 18.75,
      "grad_norm": 0.13940396904945374,
      "learning_rate": 4.102886592183996e-05,
      "loss": 0.3378,
      "num_input_tokens_seen": 920104,
      "step": 525
    },
    {
      "epoch": 18.928571428571427,
      "grad_norm": 0.051015470176935196,
      "learning_rate": 3.9948977374014544e-05,
      "loss": 0.3511,
      "num_input_tokens_seen": 928936,
      "step": 530
    },
    {
      "epoch": 19.107142857142858,
      "grad_norm": 0.09734626859426498,
      "learning_rate": 3.887395330218429e-05,
      "loss": 0.3465,
      "num_input_tokens_seen": 937840,
      "step": 535
    },
    {
      "epoch": 19.285714285714285,
      "grad_norm": 0.11513973772525787,
      "learning_rate": 3.780431399458114e-05,
      "loss": 0.3478,
      "num_input_tokens_seen": 947248,
      "step": 540
    },
    {
      "epoch": 19.464285714285715,
      "grad_norm": 0.11863286793231964,
      "learning_rate": 3.6740577133328524e-05,
      "loss": 0.3408,
      "num_input_tokens_seen": 955568,
      "step": 545
    },
    {
      "epoch": 19.5,
      "eval_loss": 0.3516767919063568,
      "eval_runtime": 1.1222,
      "eval_samples_per_second": 98.916,
      "eval_steps_per_second": 12.476,
      "num_input_tokens_seen": 957488,
      "step": 546
    },
    {
      "epoch": 19.642857142857142,
      "grad_norm": 0.04675190523266792,
      "learning_rate": 3.568325754389438e-05,
      "loss": 0.3459,
      "num_input_tokens_seen": 964400,
      "step": 550
    },
    {
      "epoch": 19.821428571428573,
      "grad_norm": 0.13177751004695892,
      "learning_rate": 3.4632866945926855e-05,
      "loss": 0.3473,
      "num_input_tokens_seen": 972720,
      "step": 555
    },
    {
      "epoch": 20.0,
      "grad_norm": 0.05624629184603691,
      "learning_rate": 3.3589913705593235e-05,
      "loss": 0.3508,
      "num_input_tokens_seen": 980760,
      "step": 560
    },
    {
      "epoch": 20.178571428571427,
      "grad_norm": 0.04487062245607376,
      "learning_rate": 3.255490258954167e-05,
      "loss": 0.3414,
      "num_input_tokens_seen": 989464,
      "step": 565
    },
    {
      "epoch": 20.357142857142858,
      "grad_norm": 0.3037759065628052,
      "learning_rate": 3.152833452060522e-05,
      "loss": 0.3518,
      "num_input_tokens_seen": 998680,
      "step": 570
    },
    {
      "epoch": 20.535714285714285,
      "grad_norm": 0.34507808089256287,
      "learning_rate": 3.0510706335366035e-05,
      "loss": 0.3479,
      "num_input_tokens_seen": 1007768,
      "step": 575
    },
    {
      "epoch": 20.714285714285715,
      "grad_norm": 0.32024624943733215,
      "learning_rate": 2.9502510543697325e-05,
      "loss": 0.3549,
      "num_input_tokens_seen": 1016536,
      "step": 580
    },
    {
      "epoch": 20.892857142857142,
      "grad_norm": 0.0999528020620346,
      "learning_rate": 2.850423509039928e-05,
      "loss": 0.3469,
      "num_input_tokens_seen": 1025048,
      "step": 585
    },
    {
      "epoch": 21.0,
      "eval_loss": 0.35418227314949036,
      "eval_runtime": 1.282,
      "eval_samples_per_second": 86.583,
      "eval_steps_per_second": 10.92,
      "num_input_tokens_seen": 1029792,
      "step": 588
    },
    {
      "epoch": 21.071428571428573,
      "grad_norm": 0.10501652210950851,
      "learning_rate": 2.751636311904444e-05,
      "loss": 0.343,
      "num_input_tokens_seen": 1032864,
      "step": 590
    },
    {
      "epoch": 21.25,
      "grad_norm": 0.04894215986132622,
      "learning_rate": 2.6539372738146695e-05,
      "loss": 0.3426,
      "num_input_tokens_seen": 1041248,
      "step": 595
    },
    {
      "epoch": 21.428571428571427,
      "grad_norm": 0.11162778735160828,
      "learning_rate": 2.5573736789767232e-05,
      "loss": 0.3358,
      "num_input_tokens_seen": 1050720,
      "step": 600
    },
    {
      "epoch": 21.607142857142858,
      "grad_norm": 0.11263741552829742,
      "learning_rate": 2.4619922620669218e-05,
      "loss": 0.3549,
      "num_input_tokens_seen": 1060064,
      "step": 605
    },
    {
      "epoch": 21.785714285714285,
      "grad_norm": 0.2273244857788086,
      "learning_rate": 2.3678391856132204e-05,
      "loss": 0.3524,
      "num_input_tokens_seen": 1068256,
      "step": 610
    },
    {
      "epoch": 21.964285714285715,
      "grad_norm": 0.23367320001125336,
      "learning_rate": 2.2749600176535534e-05,
      "loss": 0.3474,
      "num_input_tokens_seen": 1077024,
      "step": 615
    },
    {
      "epoch": 22.142857142857142,
      "grad_norm": 0.17969422042369843,
      "learning_rate": 2.1833997096818898e-05,
      "loss": 0.3425,
      "num_input_tokens_seen": 1086328,
      "step": 620
    },
    {
      "epoch": 22.321428571428573,
      "grad_norm": 0.20163026452064514,
      "learning_rate": 2.0932025748927013e-05,
      "loss": 0.3438,
      "num_input_tokens_seen": 1094520,
      "step": 625
    },
    {
      "epoch": 22.5,
      "grad_norm": 0.23486794531345367,
      "learning_rate": 2.0044122667343297e-05,
      "loss": 0.3488,
      "num_input_tokens_seen": 1103160,
      "step": 630
    },
    {
      "epoch": 22.5,
      "eval_loss": 0.35537073016166687,
      "eval_runtime": 1.1326,
      "eval_samples_per_second": 98.004,
      "eval_steps_per_second": 12.361,
      "num_input_tokens_seen": 1103160,
      "step": 630
    },
    {
      "epoch": 22.678571428571427,
      "grad_norm": 0.04255451634526253,
      "learning_rate": 1.917071757781679e-05,
      "loss": 0.3465,
      "num_input_tokens_seen": 1112376,
      "step": 635
    },
    {
      "epoch": 22.857142857142858,
      "grad_norm": 0.07782665640115738,
      "learning_rate": 1.831223318938419e-05,
      "loss": 0.3449,
      "num_input_tokens_seen": 1121464,
      "step": 640
    },
    {
      "epoch": 23.035714285714285,
      "grad_norm": 0.11463826894760132,
      "learning_rate": 1.746908498978791e-05,
      "loss": 0.3526,
      "num_input_tokens_seen": 1130152,
      "step": 645
    },
    {
      "epoch": 23.214285714285715,
      "grad_norm": 0.07137856632471085,
      "learning_rate": 1.6641681044389014e-05,
      "loss": 0.3495,
      "num_input_tokens_seen": 1138664,
      "step": 650
    },
    {
      "epoch": 23.392857142857142,
      "grad_norm": 0.08257535845041275,
      "learning_rate": 1.5830421798672568e-05,
      "loss": 0.3447,
      "num_input_tokens_seen": 1146216,
      "step": 655
    },
    {
      "epoch": 23.571428571428573,
      "grad_norm": 0.0766848772764206,
      "learning_rate": 1.5035699884440697e-05,
      "loss": 0.3509,
      "num_input_tokens_seen": 1155496,
      "step": 660
    },
    {
      "epoch": 23.75,
      "grad_norm": 0.08535895496606827,
      "learning_rate": 1.4257899929787294e-05,
      "loss": 0.3515,
      "num_input_tokens_seen": 1164840,
      "step": 665
    },
    {
      "epoch": 23.928571428571427,
      "grad_norm": 0.09435199946165085,
      "learning_rate": 1.3497398372946501e-05,
      "loss": 0.3402,
      "num_input_tokens_seen": 1173736,
      "step": 670
    },
    {
      "epoch": 24.0,
      "eval_loss": 0.35293668508529663,
      "eval_runtime": 1.2061,
      "eval_samples_per_second": 92.034,
      "eval_steps_per_second": 11.608,
      "num_input_tokens_seen": 1176968,
      "step": 672
    },
    {
      "epoch": 24.107142857142858,
      "grad_norm": 0.05331406742334366,
      "learning_rate": 1.2754563280104714e-05,
      "loss": 0.3404,
      "num_input_tokens_seen": 1182344,
      "step": 675
    },
    {
      "epoch": 24.285714285714285,
      "grad_norm": 0.14441683888435364,
      "learning_rate": 1.202975416726464e-05,
      "loss": 0.357,
      "num_input_tokens_seen": 1191944,
      "step": 680
    },
    {
      "epoch": 24.464285714285715,
      "grad_norm": 0.06084274500608444,
      "learning_rate": 1.1323321826247346e-05,
      "loss": 0.3355,
      "num_input_tokens_seen": 1200520,
      "step": 685
    },
    {
      "epoch": 24.642857142857142,
      "grad_norm": 0.13699252903461456,
      "learning_rate": 1.0635608154916648e-05,
      "loss": 0.3433,
      "num_input_tokens_seen": 1209288,
      "step": 690
    },
    {
      "epoch": 24.821428571428573,
      "grad_norm": 0.14677266776561737,
      "learning_rate": 9.966945991708005e-06,
      "loss": 0.3554,
      "num_input_tokens_seen": 1217544,
      "step": 695
    },
    {
      "epoch": 25.0,
      "grad_norm": 0.23254753649234772,
      "learning_rate": 9.317658954541992e-06,
      "loss": 0.3439,
      "num_input_tokens_seen": 1224848,
      "step": 700
    },
    {
      "epoch": 25.178571428571427,
      "grad_norm": 0.1202574372291565,
      "learning_rate": 8.688061284200266e-06,
      "loss": 0.3371,
      "num_input_tokens_seen": 1234064,
      "step": 705
    },
    {
      "epoch": 25.357142857142858,
      "grad_norm": 0.055635951459407806,
      "learning_rate": 8.07845769223981e-06,
      "loss": 0.3422,
      "num_input_tokens_seen": 1242448,
      "step": 710
    },
    {
      "epoch": 25.5,
      "eval_loss": 0.35582345724105835,
      "eval_runtime": 1.0722,
      "eval_samples_per_second": 103.524,
      "eval_steps_per_second": 13.057,
      "num_input_tokens_seen": 1250064,
      "step": 714
    },
    {
      "epoch": 25.535714285714285,
      "grad_norm": 0.29359593987464905,
      "learning_rate": 7.489143213519301e-06,
      "loss": 0.352,
      "num_input_tokens_seen": 1251536,
      "step": 715
    },
    {
      "epoch": 25.714285714285715,
      "grad_norm": 0.08035387843847275,
      "learning_rate": 6.920403063408526e-06,
      "loss": 0.3426,
      "num_input_tokens_seen": 1259792,
      "step": 720
    },
    {
      "epoch": 25.892857142857142,
      "grad_norm": 0.21466954052448273,
      "learning_rate": 6.372512499750471e-06,
      "loss": 0.3496,
      "num_input_tokens_seen": 1268560,
      "step": 725
    },
    {
      "epoch": 26.071428571428573,
      "grad_norm": 0.15136410295963287,
      "learning_rate": 5.845736689642472e-06,
      "loss": 0.3447,
      "num_input_tokens_seen": 1276648,
      "step": 730
    },
    {
      "epoch": 26.25,
      "grad_norm": 0.13561971485614777,
      "learning_rate": 5.3403305811010885e-06,
      "loss": 0.3406,
      "num_input_tokens_seen": 1287144,
      "step": 735
    },
    {
      "epoch": 26.428571428571427,
      "grad_norm": 0.07142584770917892,
      "learning_rate": 4.8565387796728865e-06,
      "loss": 0.3451,
      "num_input_tokens_seen": 1294504,
      "step": 740
    },
    {
      "epoch": 26.607142857142858,
      "grad_norm": 0.2093130499124527,
      "learning_rate": 4.394595430050613e-06,
      "loss": 0.3485,
      "num_input_tokens_seen": 1303336,
      "step": 745
    },
    {
      "epoch": 26.785714285714285,
      "grad_norm": 0.07084821909666061,
      "learning_rate": 3.954724102752316e-06,
      "loss": 0.3451,
      "num_input_tokens_seen": 1312296,
      "step": 750
    },
    {
      "epoch": 26.964285714285715,
      "grad_norm": 0.1166943907737732,
      "learning_rate": 3.537137685918074e-06,
      "loss": 0.3474,
      "num_input_tokens_seen": 1319912,
      "step": 755
    },
    {
      "epoch": 27.0,
      "eval_loss": 0.3538145124912262,
      "eval_runtime": 1.1107,
      "eval_samples_per_second": 99.936,
      "eval_steps_per_second": 12.605,
      "num_input_tokens_seen": 1321408,
      "step": 756
    },
    {
      "epoch": 27.142857142857142,
      "grad_norm": 0.21413478255271912,
      "learning_rate": 3.1420382822767323e-06,
      "loss": 0.3382,
      "num_input_tokens_seen": 1329024,
      "step": 760
    },
    {
      "epoch": 27.321428571428573,
      "grad_norm": 0.08071761578321457,
      "learning_rate": 2.7696171113326396e-06,
      "loss": 0.3531,
      "num_input_tokens_seen": 1337984,
      "step": 765
    },
    {
      "epoch": 27.5,
      "grad_norm": 0.20051950216293335,
      "learning_rate": 2.420054416819556e-06,
      "loss": 0.3338,
      "num_input_tokens_seen": 1346112,
      "step": 770
    },
    {
      "epoch": 27.678571428571427,
      "grad_norm": 0.05100846663117409,
      "learning_rate": 2.093519379466602e-06,
      "loss": 0.342,
      "num_input_tokens_seen": 1355584,
      "step": 775
    },
    {
      "epoch": 27.857142857142858,
      "grad_norm": 0.11974883079528809,
      "learning_rate": 1.7901700351184659e-06,
      "loss": 0.3414,
      "num_input_tokens_seen": 1364864,
      "step": 780
    },
    {
      "epoch": 28.035714285714285,
      "grad_norm": 0.0698491781949997,
      "learning_rate": 1.5101531982495308e-06,
      "loss": 0.3553,
      "num_input_tokens_seen": 1373400,
      "step": 785
    },
    {
      "epoch": 28.214285714285715,
      "grad_norm": 0.18634414672851562,
      "learning_rate": 1.2536043909088191e-06,
      "loss": 0.349,
      "num_input_tokens_seen": 1382744,
      "step": 790
    },
    {
      "epoch": 28.392857142857142,
      "grad_norm": 0.05743042752146721,
      "learning_rate": 1.0206477771303236e-06,
      "loss": 0.3409,
      "num_input_tokens_seen": 1390424,
      "step": 795
    },
    {
      "epoch": 28.5,
      "eval_loss": 0.35238173604011536,
      "eval_runtime": 1.1779,
      "eval_samples_per_second": 94.233,
      "eval_steps_per_second": 11.885,
      "num_input_tokens_seen": 1394904,
      "step": 798
    },
    {
      "epoch": 28.571428571428573,
      "grad_norm": 0.0811547264456749,
      "learning_rate": 8.113961028402894e-07,
      "loss": 0.3513,
      "num_input_tokens_seen": 1399256,
      "step": 800
    },
    {
      "epoch": 28.75,
      "grad_norm": 0.06173446401953697,
      "learning_rate": 6.259506412906402e-07,
      "loss": 0.35,
      "num_input_tokens_seen": 1407896,
      "step": 805
    },
    {
      "epoch": 28.928571428571427,
      "grad_norm": 0.05312683433294296,
      "learning_rate": 4.6440114404492363e-07,
      "loss": 0.338,
      "num_input_tokens_seen": 1416472,
      "step": 810
    },
    {
      "epoch": 29.107142857142858,
      "grad_norm": 0.07188601791858673,
      "learning_rate": 3.268257975405697e-07,
      "loss": 0.3438,
      "num_input_tokens_seen": 1425120,
      "step": 815
    },
    {
      "epoch": 29.285714285714285,
      "grad_norm": 0.07120722532272339,
      "learning_rate": 2.1329118524827662e-07,
      "loss": 0.3419,
      "num_input_tokens_seen": 1433248,
      "step": 820
    },
    {
      "epoch": 29.464285714285715,
      "grad_norm": 0.12804925441741943,
      "learning_rate": 1.238522554470989e-07,
      "loss": 0.3417,
      "num_input_tokens_seen": 1442592,
      "step": 825
    },
    {
      "epoch": 29.642857142857142,
      "grad_norm": 0.26469874382019043,
      "learning_rate": 5.855229463068712e-08,
      "loss": 0.3451,
      "num_input_tokens_seen": 1451360,
      "step": 830
    },
    {
      "epoch": 29.821428571428573,
      "grad_norm": 0.21112701296806335,
      "learning_rate": 1.742290655755707e-08,
      "loss": 0.3525,
      "num_input_tokens_seen": 1459808,
      "step": 835
    },
    {
      "epoch": 30.0,
      "grad_norm": 0.2481113076210022,
      "learning_rate": 4.839969555581192e-10,
      "loss": 0.3344,
      "num_input_tokens_seen": 1468632,
      "step": 840
    },
    {
      "epoch": 30.0,
      "eval_loss": 0.3523610234260559,
      "eval_runtime": 1.2257,
      "eval_samples_per_second": 90.557,
      "eval_steps_per_second": 11.422,
      "num_input_tokens_seen": 1468632,
      "step": 840
    },
    {
      "epoch": 30.0,
      "num_input_tokens_seen": 1468632,
      "step": 840,
      "total_flos": 6.613183518533222e+16,
      "train_loss": 0.47382661629290806,
      "train_runtime": 311.6966,
      "train_samples_per_second": 42.638,
      "train_steps_per_second": 2.695
    }
  ],
  "logging_steps": 5,
  "max_steps": 840,
  "num_input_tokens_seen": 1468632,
  "num_train_epochs": 30,
  "save_steps": 42,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 6.613183518533222e+16,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}