{
  "best_metric": 0.41214704513549805,
  "best_model_checkpoint": "mikhail-panzo/fil_b32_le5_s4000/checkpoint-2500",
  "epoch": 266.6666666666667,
  "eval_steps": 500,
  "global_step": 3000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 4.444444444444445,
      "grad_norm": 4.288165092468262,
      "learning_rate": 2.5000000000000004e-07,
      "loss": 0.7989,
      "step": 50
    },
    {
      "epoch": 8.88888888888889,
      "grad_norm": 2.4493463039398193,
      "learning_rate": 5.000000000000001e-07,
      "loss": 0.7701,
      "step": 100
    },
    {
      "epoch": 13.333333333333334,
      "grad_norm": 1.3798907995224,
      "learning_rate": 7.5e-07,
      "loss": 0.7374,
      "step": 150
    },
    {
      "epoch": 17.77777777777778,
      "grad_norm": 2.6173877716064453,
      "learning_rate": 1.0000000000000002e-06,
      "loss": 0.7101,
      "step": 200
    },
    {
      "epoch": 22.22222222222222,
      "grad_norm": 1.3756701946258545,
      "learning_rate": 1.25e-06,
      "loss": 0.6887,
      "step": 250
    },
    {
      "epoch": 26.666666666666668,
      "grad_norm": 1.0456843376159668,
      "learning_rate": 1.5e-06,
      "loss": 0.672,
      "step": 300
    },
    {
      "epoch": 31.11111111111111,
      "grad_norm": 0.9518398642539978,
      "learning_rate": 1.75e-06,
      "loss": 0.6583,
      "step": 350
    },
    {
      "epoch": 35.55555555555556,
      "grad_norm": 1.305978775024414,
      "learning_rate": 2.0000000000000003e-06,
      "loss": 0.6258,
      "step": 400
    },
    {
      "epoch": 40.0,
      "grad_norm": 0.8857623338699341,
      "learning_rate": 2.25e-06,
      "loss": 0.5835,
      "step": 450
    },
    {
      "epoch": 44.44444444444444,
      "grad_norm": 1.052182912826538,
      "learning_rate": 2.5e-06,
      "loss": 0.5539,
      "step": 500
    },
    {
      "epoch": 44.44444444444444,
      "eval_loss": 0.48496538400650024,
      "eval_runtime": 9.4574,
      "eval_samples_per_second": 16.812,
      "eval_steps_per_second": 2.115,
      "step": 500
    },
    {
      "epoch": 48.888888888888886,
      "grad_norm": 1.249458909034729,
      "learning_rate": 2.7500000000000004e-06,
      "loss": 0.5387,
      "step": 550
    },
    {
      "epoch": 53.333333333333336,
      "grad_norm": 0.801243245601654,
      "learning_rate": 3e-06,
      "loss": 0.5344,
      "step": 600
    },
    {
      "epoch": 57.77777777777778,
      "grad_norm": 1.1922223567962646,
      "learning_rate": 3.2500000000000002e-06,
      "loss": 0.5259,
      "step": 650
    },
    {
      "epoch": 62.22222222222222,
      "grad_norm": 0.9428519606590271,
      "learning_rate": 3.5e-06,
      "loss": 0.5179,
      "step": 700
    },
    {
      "epoch": 66.66666666666667,
      "grad_norm": 0.755260705947876,
      "learning_rate": 3.7500000000000005e-06,
      "loss": 0.5082,
      "step": 750
    },
    {
      "epoch": 71.11111111111111,
      "grad_norm": 1.8587590456008911,
      "learning_rate": 4.000000000000001e-06,
      "loss": 0.5096,
      "step": 800
    },
    {
      "epoch": 75.55555555555556,
      "grad_norm": 0.6854829788208008,
      "learning_rate": 4.25e-06,
      "loss": 0.4952,
      "step": 850
    },
    {
      "epoch": 80.0,
      "grad_norm": 0.9088854789733887,
      "learning_rate": 4.5e-06,
      "loss": 0.4941,
      "step": 900
    },
    {
      "epoch": 84.44444444444444,
      "grad_norm": 0.8932746648788452,
      "learning_rate": 4.75e-06,
      "loss": 0.4936,
      "step": 950
    },
    {
      "epoch": 88.88888888888889,
      "grad_norm": 0.9161014556884766,
      "learning_rate": 5e-06,
      "loss": 0.4867,
      "step": 1000
    },
    {
      "epoch": 88.88888888888889,
      "eval_loss": 0.4438903331756592,
      "eval_runtime": 9.2831,
      "eval_samples_per_second": 17.128,
      "eval_steps_per_second": 2.154,
      "step": 1000
    },
    {
      "epoch": 93.33333333333333,
      "grad_norm": 0.6858564615249634,
      "learning_rate": 5.2500000000000006e-06,
      "loss": 0.486,
      "step": 1050
    },
    {
      "epoch": 97.77777777777777,
      "grad_norm": 1.2040996551513672,
      "learning_rate": 5.500000000000001e-06,
      "loss": 0.4789,
      "step": 1100
    },
    {
      "epoch": 102.22222222222223,
      "grad_norm": 1.0067086219787598,
      "learning_rate": 5.75e-06,
      "loss": 0.477,
      "step": 1150
    },
    {
      "epoch": 106.66666666666667,
      "grad_norm": 1.4230589866638184,
      "learning_rate": 6e-06,
      "loss": 0.473,
      "step": 1200
    },
    {
      "epoch": 111.11111111111111,
      "grad_norm": 0.916904091835022,
      "learning_rate": 6.25e-06,
      "loss": 0.4698,
      "step": 1250
    },
    {
      "epoch": 115.55555555555556,
      "grad_norm": 0.9870907664299011,
      "learning_rate": 6.5000000000000004e-06,
      "loss": 0.4693,
      "step": 1300
    },
    {
      "epoch": 120.0,
      "grad_norm": 1.0495694875717163,
      "learning_rate": 6.750000000000001e-06,
      "loss": 0.4651,
      "step": 1350
    },
    {
      "epoch": 124.44444444444444,
      "grad_norm": 1.6615877151489258,
      "learning_rate": 7e-06,
      "loss": 0.4648,
      "step": 1400
    },
    {
      "epoch": 128.88888888888889,
      "grad_norm": 0.89169842004776,
      "learning_rate": 7.25e-06,
      "loss": 0.4643,
      "step": 1450
    },
    {
      "epoch": 133.33333333333334,
      "grad_norm": 0.8503329157829285,
      "learning_rate": 7.500000000000001e-06,
      "loss": 0.4595,
      "step": 1500
    },
    {
      "epoch": 133.33333333333334,
      "eval_loss": 0.4245499074459076,
      "eval_runtime": 8.4972,
      "eval_samples_per_second": 18.712,
      "eval_steps_per_second": 2.354,
      "step": 1500
    },
    {
      "epoch": 137.77777777777777,
      "grad_norm": 0.9252403974533081,
      "learning_rate": 7.75e-06,
      "loss": 0.4617,
      "step": 1550
    },
    {
      "epoch": 142.22222222222223,
      "grad_norm": 0.8366873860359192,
      "learning_rate": 8.000000000000001e-06,
      "loss": 0.4557,
      "step": 1600
    },
    {
      "epoch": 146.66666666666666,
      "grad_norm": 0.7907727360725403,
      "learning_rate": 8.25e-06,
      "loss": 0.4565,
      "step": 1650
    },
    {
      "epoch": 151.11111111111111,
      "grad_norm": 0.8415725231170654,
      "learning_rate": 8.5e-06,
      "loss": 0.4544,
      "step": 1700
    },
    {
      "epoch": 155.55555555555554,
      "grad_norm": 0.8423948884010315,
      "learning_rate": 8.750000000000001e-06,
      "loss": 0.4528,
      "step": 1750
    },
    {
      "epoch": 160.0,
      "grad_norm": 0.8236134648323059,
      "learning_rate": 9e-06,
      "loss": 0.4555,
      "step": 1800
    },
    {
      "epoch": 164.44444444444446,
      "grad_norm": 0.8070024251937866,
      "learning_rate": 9.250000000000001e-06,
      "loss": 0.4474,
      "step": 1850
    },
    {
      "epoch": 168.88888888888889,
      "grad_norm": 1.041704535484314,
      "learning_rate": 9.5e-06,
      "loss": 0.4478,
      "step": 1900
    },
    {
      "epoch": 173.33333333333334,
      "grad_norm": 1.021083116531372,
      "learning_rate": 9.75e-06,
      "loss": 0.4433,
      "step": 1950
    },
    {
      "epoch": 177.77777777777777,
      "grad_norm": 1.097479224205017,
      "learning_rate": 1e-05,
      "loss": 0.4395,
      "step": 2000
    },
    {
      "epoch": 177.77777777777777,
      "eval_loss": 0.41554170846939087,
      "eval_runtime": 9.3691,
      "eval_samples_per_second": 16.971,
      "eval_steps_per_second": 2.135,
      "step": 2000
    },
    {
      "epoch": 182.22222222222223,
      "grad_norm": 1.0658737421035767,
      "learning_rate": 9.75e-06,
      "loss": 0.4403,
      "step": 2050
    },
    {
      "epoch": 186.66666666666666,
      "grad_norm": 0.7741203308105469,
      "learning_rate": 9.5e-06,
      "loss": 0.4371,
      "step": 2100
    },
    {
      "epoch": 191.11111111111111,
      "grad_norm": 0.8409827947616577,
      "learning_rate": 9.250000000000001e-06,
      "loss": 0.4383,
      "step": 2150
    },
    {
      "epoch": 195.55555555555554,
      "grad_norm": 0.7340779900550842,
      "learning_rate": 9e-06,
      "loss": 0.4367,
      "step": 2200
    },
    {
      "epoch": 200.0,
      "grad_norm": 1.0130568742752075,
      "learning_rate": 8.750000000000001e-06,
      "loss": 0.4364,
      "step": 2250
    },
    {
      "epoch": 204.44444444444446,
      "grad_norm": 0.7156779170036316,
      "learning_rate": 8.5e-06,
      "loss": 0.4337,
      "step": 2300
    },
    {
      "epoch": 208.88888888888889,
      "grad_norm": 1.1859618425369263,
      "learning_rate": 8.25e-06,
      "loss": 0.4342,
      "step": 2350
    },
    {
      "epoch": 213.33333333333334,
      "grad_norm": 0.6573166847229004,
      "learning_rate": 8.000000000000001e-06,
      "loss": 0.432,
      "step": 2400
    },
    {
      "epoch": 217.77777777777777,
      "grad_norm": 1.0668668746948242,
      "learning_rate": 7.75e-06,
      "loss": 0.4338,
      "step": 2450
    },
    {
      "epoch": 222.22222222222223,
      "grad_norm": 0.8622764945030212,
      "learning_rate": 7.500000000000001e-06,
      "loss": 0.4327,
      "step": 2500
    },
    {
      "epoch": 222.22222222222223,
      "eval_loss": 0.41214704513549805,
      "eval_runtime": 9.4587,
      "eval_samples_per_second": 16.81,
      "eval_steps_per_second": 2.114,
      "step": 2500
    },
    {
      "epoch": 226.66666666666666,
      "grad_norm": 1.0643348693847656,
      "learning_rate": 7.25e-06,
      "loss": 0.4333,
      "step": 2550
    },
    {
      "epoch": 231.11111111111111,
      "grad_norm": 0.6673319339752197,
      "learning_rate": 7e-06,
      "loss": 0.4306,
      "step": 2600
    },
    {
      "epoch": 235.55555555555554,
      "grad_norm": 0.7380855679512024,
      "learning_rate": 6.750000000000001e-06,
      "loss": 0.4295,
      "step": 2650
    },
    {
      "epoch": 240.0,
      "grad_norm": 0.984495222568512,
      "learning_rate": 6.5000000000000004e-06,
      "loss": 0.4291,
      "step": 2700
    },
    {
      "epoch": 244.44444444444446,
      "grad_norm": 0.8747392892837524,
      "learning_rate": 6.25e-06,
      "loss": 0.4268,
      "step": 2750
    },
    {
      "epoch": 248.88888888888889,
      "grad_norm": 0.781096339225769,
      "learning_rate": 6e-06,
      "loss": 0.4262,
      "step": 2800
    },
    {
      "epoch": 253.33333333333334,
      "grad_norm": 1.004494309425354,
      "learning_rate": 5.75e-06,
      "loss": 0.4244,
      "step": 2850
    },
    {
      "epoch": 257.77777777777777,
      "grad_norm": 0.8275569677352905,
      "learning_rate": 5.500000000000001e-06,
      "loss": 0.4265,
      "step": 2900
    },
    {
      "epoch": 262.22222222222223,
      "grad_norm": 0.6887015104293823,
      "learning_rate": 5.2500000000000006e-06,
      "loss": 0.426,
      "step": 2950
    },
    {
      "epoch": 266.6666666666667,
      "grad_norm": 0.8565031886100769,
      "learning_rate": 5e-06,
      "loss": 0.4279,
      "step": 3000
    },
    {
      "epoch": 266.6666666666667,
      "eval_loss": 0.41272813081741333,
      "eval_runtime": 9.7505,
      "eval_samples_per_second": 16.307,
      "eval_steps_per_second": 2.051,
      "step": 3000
    }
  ],
  "logging_steps": 50,
  "max_steps": 4000,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 364,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 8.24943252918312e+16,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": null
}