{
  "best_metric": 1.8369135856628418,
  "best_model_checkpoint": "./llama3-itihasa/checkpoint-2000",
  "epoch": 0.9737098344693281,
  "eval_steps": 500,
  "global_step": 2000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.024342745861733205,
      "grad_norm": 0.798042893409729,
      "learning_rate": 0.00012165450121654502,
      "loss": 2.421,
      "step": 50
    },
    {
      "epoch": 0.04868549172346641,
      "grad_norm": 0.6517059803009033,
      "learning_rate": 0.00024330900243309004,
      "loss": 2.0325,
      "step": 100
    },
    {
      "epoch": 0.0730282375851996,
      "grad_norm": 1.047722339630127,
      "learning_rate": 0.00036496350364963507,
      "loss": 2.0045,
      "step": 150
    },
    {
      "epoch": 0.09737098344693282,
      "grad_norm": 0.724214494228363,
      "learning_rate": 0.00048661800486618007,
      "loss": 1.9665,
      "step": 200
    },
    {
      "epoch": 0.12171372930866602,
      "grad_norm": 0.7182360887527466,
      "learning_rate": 0.0006082725060827251,
      "loss": 1.9531,
      "step": 250
    },
    {
      "epoch": 0.1460564751703992,
      "grad_norm": 0.741189181804657,
      "learning_rate": 0.0007299270072992701,
      "loss": 1.9566,
      "step": 300
    },
    {
      "epoch": 0.17039922103213243,
      "grad_norm": 0.7193257212638855,
      "learning_rate": 0.0008515815085158151,
      "loss": 1.9536,
      "step": 350
    },
    {
      "epoch": 0.19474196689386564,
      "grad_norm": 0.7618579864501953,
      "learning_rate": 0.0009732360097323601,
      "loss": 1.9642,
      "step": 400
    },
    {
      "epoch": 0.21908471275559882,
      "grad_norm": 0.8242681622505188,
      "learning_rate": 0.0009894509061401135,
      "loss": 1.9899,
      "step": 450
    },
    {
      "epoch": 0.24342745861733203,
      "grad_norm": 1.0397591590881348,
      "learning_rate": 0.0009759264268325669,
      "loss": 1.9819,
      "step": 500
    },
    {
      "epoch": 0.24342745861733203,
      "eval_loss": 1.9741202592849731,
      "eval_runtime": 452.9914,
      "eval_samples_per_second": 3.022,
      "eval_steps_per_second": 0.757,
      "step": 500
    },
    {
      "epoch": 0.26777020447906524,
      "grad_norm": 0.904543936252594,
      "learning_rate": 0.0009624019475250203,
      "loss": 1.9513,
      "step": 550
    },
    {
      "epoch": 0.2921129503407984,
      "grad_norm": 0.8555233478546143,
      "learning_rate": 0.0009488774682174737,
      "loss": 1.9543,
      "step": 600
    },
    {
      "epoch": 0.31645569620253167,
      "grad_norm": 0.7415252923965454,
      "learning_rate": 0.000935352988909927,
      "loss": 1.9621,
      "step": 650
    },
    {
      "epoch": 0.34079844206426485,
      "grad_norm": 0.7195931077003479,
      "learning_rate": 0.0009218285096023803,
      "loss": 1.9414,
      "step": 700
    },
    {
      "epoch": 0.36514118792599803,
      "grad_norm": 0.9943011999130249,
      "learning_rate": 0.0009083040302948337,
      "loss": 1.9594,
      "step": 750
    },
    {
      "epoch": 0.3894839337877313,
      "grad_norm": 0.7514945268630981,
      "learning_rate": 0.000894779550987287,
      "loss": 1.9278,
      "step": 800
    },
    {
      "epoch": 0.41382667964946446,
      "grad_norm": 0.7884072661399841,
      "learning_rate": 0.0008812550716797404,
      "loss": 1.9345,
      "step": 850
    },
    {
      "epoch": 0.43816942551119764,
      "grad_norm": 0.7612814903259277,
      "learning_rate": 0.0008677305923721937,
      "loss": 1.9081,
      "step": 900
    },
    {
      "epoch": 0.4625121713729309,
      "grad_norm": 0.7805455327033997,
      "learning_rate": 0.000854206113064647,
      "loss": 1.9188,
      "step": 950
    },
    {
      "epoch": 0.48685491723466406,
      "grad_norm": 0.7169475555419922,
      "learning_rate": 0.0008406816337571004,
      "loss": 1.8903,
      "step": 1000
    },
    {
      "epoch": 0.48685491723466406,
      "eval_loss": 1.919243335723877,
      "eval_runtime": 452.9281,
      "eval_samples_per_second": 3.023,
      "eval_steps_per_second": 0.757,
      "step": 1000
    },
    {
      "epoch": 0.5111976630963972,
      "grad_norm": 0.7700228095054626,
      "learning_rate": 0.0008271571544495538,
      "loss": 1.8996,
      "step": 1050
    },
    {
      "epoch": 0.5355404089581305,
      "grad_norm": 0.8165948390960693,
      "learning_rate": 0.000813632675142007,
      "loss": 1.9178,
      "step": 1100
    },
    {
      "epoch": 0.5598831548198637,
      "grad_norm": 0.6771554350852966,
      "learning_rate": 0.0008001081958344604,
      "loss": 1.9012,
      "step": 1150
    },
    {
      "epoch": 0.5842259006815969,
      "grad_norm": 0.7270589470863342,
      "learning_rate": 0.0007865837165269137,
      "loss": 1.8971,
      "step": 1200
    },
    {
      "epoch": 0.6085686465433301,
      "grad_norm": 0.7590579390525818,
      "learning_rate": 0.0007730592372193671,
      "loss": 1.8842,
      "step": 1250
    },
    {
      "epoch": 0.6329113924050633,
      "grad_norm": 0.642905592918396,
      "learning_rate": 0.0007595347579118204,
      "loss": 1.8661,
      "step": 1300
    },
    {
      "epoch": 0.6572541382667965,
      "grad_norm": 0.7350940704345703,
      "learning_rate": 0.0007460102786042737,
      "loss": 1.8792,
      "step": 1350
    },
    {
      "epoch": 0.6815968841285297,
      "grad_norm": 0.7721772193908691,
      "learning_rate": 0.0007324857992967271,
      "loss": 1.8627,
      "step": 1400
    },
    {
      "epoch": 0.7059396299902629,
      "grad_norm": 0.7251712083816528,
      "learning_rate": 0.0007189613199891805,
      "loss": 1.8715,
      "step": 1450
    },
    {
      "epoch": 0.7302823758519961,
      "grad_norm": 0.6874309778213501,
      "learning_rate": 0.0007054368406816337,
      "loss": 1.8491,
      "step": 1500
    },
    {
      "epoch": 0.7302823758519961,
      "eval_loss": 1.8761951923370361,
      "eval_runtime": 452.8961,
      "eval_samples_per_second": 3.023,
      "eval_steps_per_second": 0.757,
      "step": 1500
    },
    {
      "epoch": 0.7546251217137293,
      "grad_norm": 0.7240548729896545,
      "learning_rate": 0.000691912361374087,
      "loss": 1.8585,
      "step": 1550
    },
    {
      "epoch": 0.7789678675754625,
      "grad_norm": 0.6931944489479065,
      "learning_rate": 0.0006783878820665405,
      "loss": 1.8623,
      "step": 1600
    },
    {
      "epoch": 0.8033106134371957,
      "grad_norm": 0.710396409034729,
      "learning_rate": 0.0006648634027589938,
      "loss": 1.8467,
      "step": 1650
    },
    {
      "epoch": 0.8276533592989289,
      "grad_norm": 0.7441169619560242,
      "learning_rate": 0.0006513389234514471,
      "loss": 1.857,
      "step": 1700
    },
    {
      "epoch": 0.8519961051606622,
      "grad_norm": 0.752155065536499,
      "learning_rate": 0.0006378144441439005,
      "loss": 1.8404,
      "step": 1750
    },
    {
      "epoch": 0.8763388510223953,
      "grad_norm": 0.6920920014381409,
      "learning_rate": 0.0006242899648363538,
      "loss": 1.8331,
      "step": 1800
    },
    {
      "epoch": 0.9006815968841285,
      "grad_norm": 0.6669381260871887,
      "learning_rate": 0.0006107654855288071,
      "loss": 1.8466,
      "step": 1850
    },
    {
      "epoch": 0.9250243427458618,
      "grad_norm": 0.6877081394195557,
      "learning_rate": 0.0005972410062212606,
      "loss": 1.829,
      "step": 1900
    },
    {
      "epoch": 0.9493670886075949,
      "grad_norm": 0.7491869926452637,
      "learning_rate": 0.0005837165269137138,
      "loss": 1.8331,
      "step": 1950
    },
    {
      "epoch": 0.9737098344693281,
      "grad_norm": 0.7513737082481384,
      "learning_rate": 0.0005701920476061671,
      "loss": 1.8106,
      "step": 2000
    },
    {
      "epoch": 0.9737098344693281,
      "eval_loss": 1.8369135856628418,
      "eval_runtime": 452.8951,
      "eval_samples_per_second": 3.023,
      "eval_steps_per_second": 0.757,
      "step": 2000
    }
  ],
  "logging_steps": 50,
  "max_steps": 4108,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 2,
  "save_steps": 1000,
  "total_flos": 3.6955207237632e+17,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}