{
  "best_metric": 0.4048522114753723,
  "best_model_checkpoint": "mikhail_panzo/fil_b128_le4_s4000/checkpoint-2000",
  "epoch": 311.1111111111111,
  "eval_steps": 500,
  "global_step": 3500,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 4.444444444444445,
      "grad_norm": 2.152308940887451,
      "learning_rate": 2.5e-06,
      "loss": 0.7738,
      "step": 50
    },
    {
      "epoch": 8.88888888888889,
      "grad_norm": 1.326554536819458,
      "learning_rate": 5e-06,
      "loss": 0.6968,
      "step": 100
    },
    {
      "epoch": 13.333333333333334,
      "grad_norm": 1.4387542009353638,
      "learning_rate": 7.5e-06,
      "loss": 0.636,
      "step": 150
    },
    {
      "epoch": 17.77777777777778,
      "grad_norm": 1.033190131187439,
      "learning_rate": 1e-05,
      "loss": 0.5486,
      "step": 200
    },
    {
      "epoch": 22.22222222222222,
      "grad_norm": 1.4383914470672607,
      "learning_rate": 1.25e-05,
      "loss": 0.516,
      "step": 250
    },
    {
      "epoch": 26.666666666666668,
      "grad_norm": 1.8267806768417358,
      "learning_rate": 1.5e-05,
      "loss": 0.4989,
      "step": 300
    },
    {
      "epoch": 31.11111111111111,
      "grad_norm": 1.2525017261505127,
      "learning_rate": 1.75e-05,
      "loss": 0.4908,
      "step": 350
    },
    {
      "epoch": 35.55555555555556,
      "grad_norm": 1.1662768125534058,
      "learning_rate": 2e-05,
      "loss": 0.4766,
      "step": 400
    },
    {
      "epoch": 40.0,
      "grad_norm": 1.110721230506897,
      "learning_rate": 2.25e-05,
      "loss": 0.4716,
      "step": 450
    },
    {
      "epoch": 44.44444444444444,
      "grad_norm": 1.6693321466445923,
      "learning_rate": 2.5e-05,
      "loss": 0.4635,
      "step": 500
    },
    {
      "epoch": 44.44444444444444,
      "eval_loss": 0.42072969675064087,
      "eval_runtime": 9.5188,
      "eval_samples_per_second": 16.704,
      "eval_steps_per_second": 2.101,
      "step": 500
    },
    {
      "epoch": 48.888888888888886,
      "grad_norm": 1.0238850116729736,
      "learning_rate": 2.7500000000000004e-05,
      "loss": 0.4585,
      "step": 550
    },
    {
      "epoch": 53.333333333333336,
      "grad_norm": 0.8729687929153442,
      "learning_rate": 3e-05,
      "loss": 0.4572,
      "step": 600
    },
    {
      "epoch": 57.77777777777778,
      "grad_norm": 1.4291303157806396,
      "learning_rate": 3.2500000000000004e-05,
      "loss": 0.452,
      "step": 650
    },
    {
      "epoch": 62.22222222222222,
      "grad_norm": 1.3374826908111572,
      "learning_rate": 3.5e-05,
      "loss": 0.447,
      "step": 700
    },
    {
      "epoch": 66.66666666666667,
      "grad_norm": 1.7066700458526611,
      "learning_rate": 3.7500000000000003e-05,
      "loss": 0.4424,
      "step": 750
    },
    {
      "epoch": 71.11111111111111,
      "grad_norm": 2.1874194145202637,
      "learning_rate": 4e-05,
      "loss": 0.4443,
      "step": 800
    },
    {
      "epoch": 75.55555555555556,
      "grad_norm": 1.4660208225250244,
      "learning_rate": 4.25e-05,
      "loss": 0.4346,
      "step": 850
    },
    {
      "epoch": 80.0,
      "grad_norm": 1.2421706914901733,
      "learning_rate": 4.5e-05,
      "loss": 0.4328,
      "step": 900
    },
    {
      "epoch": 84.44444444444444,
      "grad_norm": 0.8712289929389954,
      "learning_rate": 4.75e-05,
      "loss": 0.4335,
      "step": 950
    },
    {
      "epoch": 88.88888888888889,
      "grad_norm": 1.161516547203064,
      "learning_rate": 5e-05,
      "loss": 0.4317,
      "step": 1000
    },
    {
      "epoch": 88.88888888888889,
      "eval_loss": 0.40811634063720703,
      "eval_runtime": 9.5158,
      "eval_samples_per_second": 16.709,
      "eval_steps_per_second": 2.102,
      "step": 1000
    },
    {
      "epoch": 93.33333333333333,
      "grad_norm": 4.811057090759277,
      "learning_rate": 5.25e-05,
      "loss": 0.4311,
      "step": 1050
    },
    {
      "epoch": 97.77777777777777,
      "grad_norm": 1.5318878889083862,
      "learning_rate": 5.500000000000001e-05,
      "loss": 0.4253,
      "step": 1100
    },
    {
      "epoch": 102.22222222222223,
      "grad_norm": 1.7489895820617676,
      "learning_rate": 5.7499999999999995e-05,
      "loss": 0.4229,
      "step": 1150
    },
    {
      "epoch": 106.66666666666667,
      "grad_norm": 1.162777304649353,
      "learning_rate": 6e-05,
      "loss": 0.4214,
      "step": 1200
    },
    {
      "epoch": 111.11111111111111,
      "grad_norm": 1.2216763496398926,
      "learning_rate": 6.25e-05,
      "loss": 0.4168,
      "step": 1250
    },
    {
      "epoch": 115.55555555555556,
      "grad_norm": 1.368053674697876,
      "learning_rate": 6.500000000000001e-05,
      "loss": 0.4197,
      "step": 1300
    },
    {
      "epoch": 120.0,
      "grad_norm": 1.3240705728530884,
      "learning_rate": 6.750000000000001e-05,
      "loss": 0.4143,
      "step": 1350
    },
    {
      "epoch": 124.44444444444444,
      "grad_norm": 3.4118428230285645,
      "learning_rate": 7e-05,
      "loss": 0.4159,
      "step": 1400
    },
    {
      "epoch": 128.88888888888889,
      "grad_norm": 2.0438427925109863,
      "learning_rate": 7.25e-05,
      "loss": 0.4135,
      "step": 1450
    },
    {
      "epoch": 133.33333333333334,
      "grad_norm": 3.8993186950683594,
      "learning_rate": 7.500000000000001e-05,
      "loss": 0.412,
      "step": 1500
    },
    {
      "epoch": 133.33333333333334,
      "eval_loss": 0.40509456396102905,
      "eval_runtime": 8.6764,
      "eval_samples_per_second": 18.326,
      "eval_steps_per_second": 2.305,
      "step": 1500
    },
    {
      "epoch": 137.77777777777777,
      "grad_norm": 1.3933427333831787,
      "learning_rate": 7.75e-05,
      "loss": 0.4121,
      "step": 1550
    },
    {
      "epoch": 142.22222222222223,
      "grad_norm": 2.8709917068481445,
      "learning_rate": 8e-05,
      "loss": 0.4095,
      "step": 1600
    },
    {
      "epoch": 146.66666666666666,
      "grad_norm": 1.1129873991012573,
      "learning_rate": 8.25e-05,
      "loss": 0.4096,
      "step": 1650
    },
    {
      "epoch": 151.11111111111111,
      "grad_norm": 1.5557284355163574,
      "learning_rate": 8.5e-05,
      "loss": 0.4058,
      "step": 1700
    },
    {
      "epoch": 155.55555555555554,
      "grad_norm": 2.134979009628296,
      "learning_rate": 8.75e-05,
      "loss": 0.4045,
      "step": 1750
    },
    {
      "epoch": 160.0,
      "grad_norm": 2.0640597343444824,
      "learning_rate": 9e-05,
      "loss": 0.4103,
      "step": 1800
    },
    {
      "epoch": 164.44444444444446,
      "grad_norm": 1.2578678131103516,
      "learning_rate": 9.250000000000001e-05,
      "loss": 0.4048,
      "step": 1850
    },
    {
      "epoch": 168.88888888888889,
      "grad_norm": 1.7405868768692017,
      "learning_rate": 9.5e-05,
      "loss": 0.4019,
      "step": 1900
    },
    {
      "epoch": 173.33333333333334,
      "grad_norm": 1.7727621793746948,
      "learning_rate": 9.75e-05,
      "loss": 0.3982,
      "step": 1950
    },
    {
      "epoch": 177.77777777777777,
      "grad_norm": 2.25417160987854,
      "learning_rate": 0.0001,
      "loss": 0.395,
      "step": 2000
    },
    {
      "epoch": 177.77777777777777,
      "eval_loss": 0.4048522114753723,
      "eval_runtime": 8.8879,
      "eval_samples_per_second": 17.89,
      "eval_steps_per_second": 2.25,
      "step": 2000
    },
    {
      "epoch": 182.22222222222223,
      "grad_norm": 1.0432671308517456,
      "learning_rate": 9.75e-05,
      "loss": 0.3938,
      "step": 2050
    },
    {
      "epoch": 186.66666666666666,
      "grad_norm": 2.726656675338745,
      "learning_rate": 9.5e-05,
      "loss": 0.3911,
      "step": 2100
    },
    {
      "epoch": 191.11111111111111,
      "grad_norm": 1.70256769657135,
      "learning_rate": 9.250000000000001e-05,
      "loss": 0.3923,
      "step": 2150
    },
    {
      "epoch": 195.55555555555554,
      "grad_norm": 1.177212119102478,
      "learning_rate": 9e-05,
      "loss": 0.3897,
      "step": 2200
    },
    {
      "epoch": 200.0,
      "grad_norm": 1.0524529218673706,
      "learning_rate": 8.75e-05,
      "loss": 0.3892,
      "step": 2250
    },
    {
      "epoch": 204.44444444444446,
      "grad_norm": 2.8829586505889893,
      "learning_rate": 8.5e-05,
      "loss": 0.3873,
      "step": 2300
    },
    {
      "epoch": 208.88888888888889,
      "grad_norm": 1.4545131921768188,
      "learning_rate": 8.25e-05,
      "loss": 0.3866,
      "step": 2350
    },
    {
      "epoch": 213.33333333333334,
      "grad_norm": 1.0848562717437744,
      "learning_rate": 8e-05,
      "loss": 0.3846,
      "step": 2400
    },
    {
      "epoch": 217.77777777777777,
      "grad_norm": 4.756465435028076,
      "learning_rate": 7.75e-05,
      "loss": 0.3862,
      "step": 2450
    },
    {
      "epoch": 222.22222222222223,
      "grad_norm": 2.198255777359009,
      "learning_rate": 7.500000000000001e-05,
      "loss": 0.3848,
      "step": 2500
    },
    {
      "epoch": 222.22222222222223,
      "eval_loss": 0.4062703251838684,
      "eval_runtime": 9.7951,
      "eval_samples_per_second": 16.233,
      "eval_steps_per_second": 2.042,
      "step": 2500
    },
    {
      "epoch": 226.66666666666666,
      "grad_norm": 0.992829442024231,
      "learning_rate": 7.25e-05,
      "loss": 0.3829,
      "step": 2550
    },
    {
      "epoch": 231.11111111111111,
      "grad_norm": 1.3153003454208374,
      "learning_rate": 7e-05,
      "loss": 0.3804,
      "step": 2600
    },
    {
      "epoch": 235.55555555555554,
      "grad_norm": 1.8451892137527466,
      "learning_rate": 6.750000000000001e-05,
      "loss": 0.3791,
      "step": 2650
    },
    {
      "epoch": 240.0,
      "grad_norm": 2.005200147628784,
      "learning_rate": 6.500000000000001e-05,
      "loss": 0.3776,
      "step": 2700
    },
    {
      "epoch": 244.44444444444446,
      "grad_norm": 1.134822964668274,
      "learning_rate": 6.25e-05,
      "loss": 0.3754,
      "step": 2750
    },
    {
      "epoch": 248.88888888888889,
      "grad_norm": 1.4820879697799683,
      "learning_rate": 6e-05,
      "loss": 0.3746,
      "step": 2800
    },
    {
      "epoch": 253.33333333333334,
      "grad_norm": 0.989895761013031,
      "learning_rate": 5.7499999999999995e-05,
      "loss": 0.3735,
      "step": 2850
    },
    {
      "epoch": 257.77777777777777,
      "grad_norm": 1.1705851554870605,
      "learning_rate": 5.500000000000001e-05,
      "loss": 0.3729,
      "step": 2900
    },
    {
      "epoch": 262.22222222222223,
      "grad_norm": 0.8756380677223206,
      "learning_rate": 5.25e-05,
      "loss": 0.3719,
      "step": 2950
    },
    {
      "epoch": 266.6666666666667,
      "grad_norm": 0.8016765713691711,
      "learning_rate": 5e-05,
      "loss": 0.3738,
      "step": 3000
    },
    {
      "epoch": 266.6666666666667,
      "eval_loss": 0.40633583068847656,
      "eval_runtime": 9.5581,
      "eval_samples_per_second": 16.635,
      "eval_steps_per_second": 2.092,
      "step": 3000
    },
    {
      "epoch": 271.1111111111111,
      "grad_norm": 0.6258949637413025,
      "learning_rate": 4.75e-05,
      "loss": 0.3702,
      "step": 3050
    },
    {
      "epoch": 275.55555555555554,
      "grad_norm": 0.9251272678375244,
      "learning_rate": 4.5e-05,
      "loss": 0.3678,
      "step": 3100
    },
    {
      "epoch": 280.0,
      "grad_norm": 0.8492236733436584,
      "learning_rate": 4.25e-05,
      "loss": 0.3699,
      "step": 3150
    },
    {
      "epoch": 284.44444444444446,
      "grad_norm": 0.8289771676063538,
      "learning_rate": 4e-05,
      "loss": 0.3643,
      "step": 3200
    },
    {
      "epoch": 288.8888888888889,
      "grad_norm": 1.0957742929458618,
      "learning_rate": 3.7500000000000003e-05,
      "loss": 0.3665,
      "step": 3250
    },
    {
      "epoch": 293.3333333333333,
      "grad_norm": 0.7582021355628967,
      "learning_rate": 3.5e-05,
      "loss": 0.3622,
      "step": 3300
    },
    {
      "epoch": 297.77777777777777,
      "grad_norm": 0.7244244813919067,
      "learning_rate": 3.2500000000000004e-05,
      "loss": 0.363,
      "step": 3350
    },
    {
      "epoch": 302.22222222222223,
      "grad_norm": 1.2840015888214111,
      "learning_rate": 3e-05,
      "loss": 0.3622,
      "step": 3400
    },
    {
      "epoch": 306.6666666666667,
      "grad_norm": 0.6718984246253967,
      "learning_rate": 2.7500000000000004e-05,
      "loss": 0.365,
      "step": 3450
    },
    {
      "epoch": 311.1111111111111,
      "grad_norm": 0.6533474326133728,
      "learning_rate": 2.5e-05,
      "loss": 0.3618,
      "step": 3500
    },
    {
      "epoch": 311.1111111111111,
      "eval_loss": 0.4071924090385437,
      "eval_runtime": 9.5143,
      "eval_samples_per_second": 16.712,
      "eval_steps_per_second": 2.102,
      "step": 3500
    }
  ],
  "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": 9.618990397807824e+16,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": null
}