{
  "best_metric": 0.4094551205635071,
  "best_model_checkpoint": "mikhail-panzo/fil_b64_le5_s8000/checkpoint-3500",
  "epoch": 155.55555555555554,
  "eval_steps": 500,
  "global_step": 3500,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 2.2222222222222223,
      "grad_norm": 5.408687114715576,
      "learning_rate": 2.5000000000000004e-07,
      "loss": 0.8111,
      "step": 50
    },
    {
      "epoch": 4.444444444444445,
      "grad_norm": 3.0843052864074707,
      "learning_rate": 5.000000000000001e-07,
      "loss": 0.7694,
      "step": 100
    },
    {
      "epoch": 6.666666666666667,
      "grad_norm": 5.368913650512695,
      "learning_rate": 7.5e-07,
      "loss": 0.7562,
      "step": 150
    },
    {
      "epoch": 8.88888888888889,
      "grad_norm": 1.200212836265564,
      "learning_rate": 1.0000000000000002e-06,
      "loss": 0.7184,
      "step": 200
    },
    {
      "epoch": 11.11111111111111,
      "grad_norm": 1.2329410314559937,
      "learning_rate": 1.25e-06,
      "loss": 0.7022,
      "step": 250
    },
    {
      "epoch": 13.333333333333334,
      "grad_norm": 1.7799458503723145,
      "learning_rate": 1.5e-06,
      "loss": 0.6823,
      "step": 300
    },
    {
      "epoch": 15.555555555555555,
      "grad_norm": 2.0204672813415527,
      "learning_rate": 1.75e-06,
      "loss": 0.6635,
      "step": 350
    },
    {
      "epoch": 17.77777777777778,
      "grad_norm": 1.8028016090393066,
      "learning_rate": 2.0000000000000003e-06,
      "loss": 0.6499,
      "step": 400
    },
    {
      "epoch": 20.0,
      "grad_norm": 1.5485812425613403,
      "learning_rate": 2.25e-06,
      "loss": 0.6216,
      "step": 450
    },
    {
      "epoch": 22.22222222222222,
      "grad_norm": 1.0476795434951782,
      "learning_rate": 2.5e-06,
      "loss": 0.575,
      "step": 500
    },
    {
      "epoch": 22.22222222222222,
      "eval_loss": 0.49667900800704956,
      "eval_runtime": 8.3566,
      "eval_samples_per_second": 19.027,
      "eval_steps_per_second": 2.393,
      "step": 500
    },
    {
      "epoch": 24.444444444444443,
      "grad_norm": 1.134787678718567,
      "learning_rate": 2.7500000000000004e-06,
      "loss": 0.5464,
      "step": 550
    },
    {
      "epoch": 26.666666666666668,
      "grad_norm": 1.543503999710083,
      "learning_rate": 3e-06,
      "loss": 0.5432,
      "step": 600
    },
    {
      "epoch": 28.88888888888889,
      "grad_norm": 1.0410478115081787,
      "learning_rate": 3.2500000000000002e-06,
      "loss": 0.5355,
      "step": 650
    },
    {
      "epoch": 31.11111111111111,
      "grad_norm": 1.5184495449066162,
      "learning_rate": 3.5e-06,
      "loss": 0.5289,
      "step": 700
    },
    {
      "epoch": 33.333333333333336,
      "grad_norm": 1.490964651107788,
      "learning_rate": 3.7500000000000005e-06,
      "loss": 0.5225,
      "step": 750
    },
    {
      "epoch": 35.55555555555556,
      "grad_norm": 1.2980331182479858,
      "learning_rate": 4.000000000000001e-06,
      "loss": 0.5122,
      "step": 800
    },
    {
      "epoch": 37.77777777777778,
      "grad_norm": 1.2520424127578735,
      "learning_rate": 4.25e-06,
      "loss": 0.51,
      "step": 850
    },
    {
      "epoch": 40.0,
      "grad_norm": 1.2899489402770996,
      "learning_rate": 4.5e-06,
      "loss": 0.5051,
      "step": 900
    },
    {
      "epoch": 42.22222222222222,
      "grad_norm": 1.2064433097839355,
      "learning_rate": 4.75e-06,
      "loss": 0.5006,
      "step": 950
    },
    {
      "epoch": 44.44444444444444,
      "grad_norm": 1.1178175210952759,
      "learning_rate": 5e-06,
      "loss": 0.4945,
      "step": 1000
    },
    {
      "epoch": 44.44444444444444,
      "eval_loss": 0.4459993243217468,
      "eval_runtime": 8.2351,
      "eval_samples_per_second": 19.308,
      "eval_steps_per_second": 2.429,
      "step": 1000
    },
    {
      "epoch": 46.666666666666664,
      "grad_norm": 1.2157412767410278,
      "learning_rate": 5.2500000000000006e-06,
      "loss": 0.4976,
      "step": 1050
    },
    {
      "epoch": 48.888888888888886,
      "grad_norm": 1.0625967979431152,
      "learning_rate": 5.500000000000001e-06,
      "loss": 0.485,
      "step": 1100
    },
    {
      "epoch": 51.111111111111114,
      "grad_norm": 1.2712880373001099,
      "learning_rate": 5.75e-06,
      "loss": 0.4835,
      "step": 1150
    },
    {
      "epoch": 53.333333333333336,
      "grad_norm": 1.1251996755599976,
      "learning_rate": 6e-06,
      "loss": 0.482,
      "step": 1200
    },
    {
      "epoch": 55.55555555555556,
      "grad_norm": 1.1670546531677246,
      "learning_rate": 6.25e-06,
      "loss": 0.4847,
      "step": 1250
    },
    {
      "epoch": 57.77777777777778,
      "grad_norm": 1.2118104696273804,
      "learning_rate": 6.5000000000000004e-06,
      "loss": 0.4804,
      "step": 1300
    },
    {
      "epoch": 60.0,
      "grad_norm": 1.352782964706421,
      "learning_rate": 6.750000000000001e-06,
      "loss": 0.4763,
      "step": 1350
    },
    {
      "epoch": 62.22222222222222,
      "grad_norm": 0.9930539727210999,
      "learning_rate": 7e-06,
      "loss": 0.4776,
      "step": 1400
    },
    {
      "epoch": 64.44444444444444,
      "grad_norm": 0.9236252903938293,
      "learning_rate": 7.25e-06,
      "loss": 0.4646,
      "step": 1450
    },
    {
      "epoch": 66.66666666666667,
      "grad_norm": 1.295523762702942,
      "learning_rate": 7.500000000000001e-06,
      "loss": 0.4681,
      "step": 1500
    },
    {
      "epoch": 66.66666666666667,
      "eval_loss": 0.43010735511779785,
      "eval_runtime": 8.1609,
      "eval_samples_per_second": 19.483,
      "eval_steps_per_second": 2.451,
      "step": 1500
    },
    {
      "epoch": 68.88888888888889,
      "grad_norm": 1.0008094310760498,
      "learning_rate": 7.75e-06,
      "loss": 0.4673,
      "step": 1550
    },
    {
      "epoch": 71.11111111111111,
      "grad_norm": 1.275070071220398,
      "learning_rate": 8.000000000000001e-06,
      "loss": 0.4654,
      "step": 1600
    },
    {
      "epoch": 73.33333333333333,
      "grad_norm": 1.0170423984527588,
      "learning_rate": 8.25e-06,
      "loss": 0.4571,
      "step": 1650
    },
    {
      "epoch": 75.55555555555556,
      "grad_norm": 1.0809487104415894,
      "learning_rate": 8.5e-06,
      "loss": 0.4578,
      "step": 1700
    },
    {
      "epoch": 77.77777777777777,
      "grad_norm": 1.9192956686019897,
      "learning_rate": 8.750000000000001e-06,
      "loss": 0.4574,
      "step": 1750
    },
    {
      "epoch": 80.0,
      "grad_norm": 2.315152406692505,
      "learning_rate": 9e-06,
      "loss": 0.4566,
      "step": 1800
    },
    {
      "epoch": 82.22222222222223,
      "grad_norm": 1.0425012111663818,
      "learning_rate": 9.250000000000001e-06,
      "loss": 0.4583,
      "step": 1850
    },
    {
      "epoch": 84.44444444444444,
      "grad_norm": 0.896725058555603,
      "learning_rate": 9.5e-06,
      "loss": 0.4542,
      "step": 1900
    },
    {
      "epoch": 86.66666666666667,
      "grad_norm": 1.2378209829330444,
      "learning_rate": 9.75e-06,
      "loss": 0.455,
      "step": 1950
    },
    {
      "epoch": 88.88888888888889,
      "grad_norm": 1.078316330909729,
      "learning_rate": 1e-05,
      "loss": 0.4514,
      "step": 2000
    },
    {
      "epoch": 88.88888888888889,
      "eval_loss": 0.4194343686103821,
      "eval_runtime": 8.1871,
      "eval_samples_per_second": 19.421,
      "eval_steps_per_second": 2.443,
      "step": 2000
    },
    {
      "epoch": 91.11111111111111,
      "grad_norm": 1.1909762620925903,
      "learning_rate": 9.916666666666668e-06,
      "loss": 0.456,
      "step": 2050
    },
    {
      "epoch": 93.33333333333333,
      "grad_norm": 1.0388458967208862,
      "learning_rate": 9.833333333333333e-06,
      "loss": 0.4443,
      "step": 2100
    },
    {
      "epoch": 95.55555555555556,
      "grad_norm": 1.16915762424469,
      "learning_rate": 9.75e-06,
      "loss": 0.449,
      "step": 2150
    },
    {
      "epoch": 97.77777777777777,
      "grad_norm": 1.2234934568405151,
      "learning_rate": 9.666666666666667e-06,
      "loss": 0.4438,
      "step": 2200
    },
    {
      "epoch": 100.0,
      "grad_norm": 1.3398857116699219,
      "learning_rate": 9.583333333333335e-06,
      "loss": 0.4429,
      "step": 2250
    },
    {
      "epoch": 102.22222222222223,
      "grad_norm": 1.2885675430297852,
      "learning_rate": 9.5e-06,
      "loss": 0.4477,
      "step": 2300
    },
    {
      "epoch": 104.44444444444444,
      "grad_norm": 0.962570071220398,
      "learning_rate": 9.416666666666667e-06,
      "loss": 0.4439,
      "step": 2350
    },
    {
      "epoch": 106.66666666666667,
      "grad_norm": 1.100001573562622,
      "learning_rate": 9.333333333333334e-06,
      "loss": 0.4419,
      "step": 2400
    },
    {
      "epoch": 108.88888888888889,
      "grad_norm": 1.30157470703125,
      "learning_rate": 9.250000000000001e-06,
      "loss": 0.436,
      "step": 2450
    },
    {
      "epoch": 111.11111111111111,
      "grad_norm": 1.1278440952301025,
      "learning_rate": 9.166666666666666e-06,
      "loss": 0.439,
      "step": 2500
    },
    {
      "epoch": 111.11111111111111,
      "eval_loss": 0.4126753807067871,
      "eval_runtime": 7.3748,
      "eval_samples_per_second": 21.56,
      "eval_steps_per_second": 2.712,
      "step": 2500
    },
    {
      "epoch": 113.33333333333333,
      "grad_norm": 1.9389395713806152,
      "learning_rate": 9.083333333333333e-06,
      "loss": 0.4383,
      "step": 2550
    },
    {
      "epoch": 115.55555555555556,
      "grad_norm": 0.9682930707931519,
      "learning_rate": 9e-06,
      "loss": 0.4339,
      "step": 2600
    },
    {
      "epoch": 117.77777777777777,
      "grad_norm": 1.722495198249817,
      "learning_rate": 8.916666666666667e-06,
      "loss": 0.4344,
      "step": 2650
    },
    {
      "epoch": 120.0,
      "grad_norm": 1.1434502601623535,
      "learning_rate": 8.833333333333334e-06,
      "loss": 0.4363,
      "step": 2700
    },
    {
      "epoch": 122.22222222222223,
      "grad_norm": 1.3238564729690552,
      "learning_rate": 8.751666666666668e-06,
      "loss": 0.4402,
      "step": 2750
    },
    {
      "epoch": 124.44444444444444,
      "grad_norm": 1.2620904445648193,
      "learning_rate": 8.668333333333335e-06,
      "loss": 0.4321,
      "step": 2800
    },
    {
      "epoch": 126.66666666666667,
      "grad_norm": 1.0108044147491455,
      "learning_rate": 8.585000000000002e-06,
      "loss": 0.4353,
      "step": 2850
    },
    {
      "epoch": 128.88888888888889,
      "grad_norm": 1.6491698026657104,
      "learning_rate": 8.501666666666667e-06,
      "loss": 0.441,
      "step": 2900
    },
    {
      "epoch": 131.11111111111111,
      "grad_norm": 1.0939834117889404,
      "learning_rate": 8.418333333333334e-06,
      "loss": 0.4342,
      "step": 2950
    },
    {
      "epoch": 133.33333333333334,
      "grad_norm": 1.105880856513977,
      "learning_rate": 8.335e-06,
      "loss": 0.4304,
      "step": 3000
    },
    {
      "epoch": 133.33333333333334,
      "eval_loss": 0.41190987825393677,
      "eval_runtime": 8.1339,
      "eval_samples_per_second": 19.548,
      "eval_steps_per_second": 2.459,
      "step": 3000
    },
    {
      "epoch": 135.55555555555554,
      "grad_norm": 1.1262001991271973,
      "learning_rate": 8.251666666666668e-06,
      "loss": 0.4291,
      "step": 3050
    },
    {
      "epoch": 137.77777777777777,
      "grad_norm": 1.0499992370605469,
      "learning_rate": 8.168333333333334e-06,
      "loss": 0.4274,
      "step": 3100
    },
    {
      "epoch": 140.0,
      "grad_norm": 1.4579908847808838,
      "learning_rate": 8.085000000000001e-06,
      "loss": 0.4279,
      "step": 3150
    },
    {
      "epoch": 142.22222222222223,
      "grad_norm": 1.0599946975708008,
      "learning_rate": 8.001666666666668e-06,
      "loss": 0.4284,
      "step": 3200
    },
    {
      "epoch": 144.44444444444446,
      "grad_norm": 1.041557788848877,
      "learning_rate": 7.918333333333333e-06,
      "loss": 0.4276,
      "step": 3250
    },
    {
      "epoch": 146.66666666666666,
      "grad_norm": 1.1548891067504883,
      "learning_rate": 7.835e-06,
      "loss": 0.4264,
      "step": 3300
    },
    {
      "epoch": 148.88888888888889,
      "grad_norm": 0.968100368976593,
      "learning_rate": 7.751666666666667e-06,
      "loss": 0.4239,
      "step": 3350
    },
    {
      "epoch": 151.11111111111111,
      "grad_norm": 0.9444773197174072,
      "learning_rate": 7.668333333333334e-06,
      "loss": 0.4255,
      "step": 3400
    },
    {
      "epoch": 153.33333333333334,
      "grad_norm": 1.0166351795196533,
      "learning_rate": 7.585e-06,
      "loss": 0.4245,
      "step": 3450
    },
    {
      "epoch": 155.55555555555554,
      "grad_norm": 0.8712626695632935,
      "learning_rate": 7.501666666666667e-06,
      "loss": 0.4261,
      "step": 3500
    },
    {
      "epoch": 155.55555555555554,
      "eval_loss": 0.4094551205635071,
      "eval_runtime": 8.3177,
      "eval_samples_per_second": 19.116,
      "eval_steps_per_second": 2.405,
      "step": 3500
    }
  ],
  "logging_steps": 50,
  "max_steps": 8000,
  "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": 4.822102582474954e+16,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": null
}