{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 10.0,
  "eval_steps": 500,
  "global_step": 29280,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.17,
      "grad_norm": 1.2976419925689697,
      "learning_rate": 6.25e-05,
      "loss": 6.1408,
      "step": 500
    },
    {
      "epoch": 0.34,
      "grad_norm": 0.9176012277603149,
      "learning_rate": 0.000125,
      "loss": 3.6186,
      "step": 1000
    },
    {
      "epoch": 0.51,
      "grad_norm": 1.0713545083999634,
      "learning_rate": 0.0001875,
      "loss": 3.337,
      "step": 1500
    },
    {
      "epoch": 0.68,
      "grad_norm": 0.9314835667610168,
      "learning_rate": 0.00025,
      "loss": 3.1944,
      "step": 2000
    },
    {
      "epoch": 0.85,
      "grad_norm": 1.087267279624939,
      "learning_rate": 0.0003125,
      "loss": 3.0603,
      "step": 2500
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.4366931588461262,
      "eval_loss": 3.025468111038208,
      "eval_runtime": 3.375,
      "eval_samples_per_second": 1330.682,
      "eval_steps_per_second": 10.667,
      "step": 2928
    },
    {
      "epoch": 1.02,
      "grad_norm": 0.9429690837860107,
      "learning_rate": 0.000375,
      "loss": 2.9476,
      "step": 3000
    },
    {
      "epoch": 1.2,
      "grad_norm": 0.950432300567627,
      "learning_rate": 0.00043750000000000006,
      "loss": 2.8889,
      "step": 3500
    },
    {
      "epoch": 1.37,
      "grad_norm": 0.9229320287704468,
      "learning_rate": 0.0005,
      "loss": 2.8218,
      "step": 4000
    },
    {
      "epoch": 1.54,
      "grad_norm": 0.844063937664032,
      "learning_rate": 0.0005625000000000001,
      "loss": 2.777,
      "step": 4500
    },
    {
      "epoch": 1.71,
      "grad_norm": 0.8429157137870789,
      "learning_rate": 0.000625,
      "loss": 2.7377,
      "step": 5000
    },
    {
      "epoch": 1.88,
      "grad_norm": 0.8256997466087341,
      "learning_rate": 0.0006875,
      "loss": 2.7088,
      "step": 5500
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.4579552806400202,
      "eval_loss": 2.78731632232666,
      "eval_runtime": 3.4275,
      "eval_samples_per_second": 1310.267,
      "eval_steps_per_second": 10.503,
      "step": 5856
    },
    {
      "epoch": 2.05,
      "grad_norm": 0.7644203901290894,
      "learning_rate": 0.00075,
      "loss": 2.6787,
      "step": 6000
    },
    {
      "epoch": 2.22,
      "grad_norm": 0.8065159320831299,
      "learning_rate": 0.0008125,
      "loss": 2.6291,
      "step": 6500
    },
    {
      "epoch": 2.39,
      "grad_norm": 0.7839713096618652,
      "learning_rate": 0.0008750000000000001,
      "loss": 2.6217,
      "step": 7000
    },
    {
      "epoch": 2.56,
      "grad_norm": 0.7241440415382385,
      "learning_rate": 0.0009375,
      "loss": 2.6003,
      "step": 7500
    },
    {
      "epoch": 2.73,
      "grad_norm": 0.7345842719078064,
      "learning_rate": 0.001,
      "loss": 2.6077,
      "step": 8000
    },
    {
      "epoch": 2.9,
      "grad_norm": 0.670850396156311,
      "learning_rate": 0.0010625,
      "loss": 2.586,
      "step": 8500
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.4688046258746715,
      "eval_loss": 2.695575714111328,
      "eval_runtime": 3.4412,
      "eval_samples_per_second": 1305.068,
      "eval_steps_per_second": 10.461,
      "step": 8784
    },
    {
      "epoch": 3.07,
      "grad_norm": 0.6207906603813171,
      "learning_rate": 0.0011250000000000001,
      "loss": 2.5489,
      "step": 9000
    },
    {
      "epoch": 3.24,
      "grad_norm": 0.5981828570365906,
      "learning_rate": 0.0011875,
      "loss": 2.5196,
      "step": 9500
    },
    {
      "epoch": 3.42,
      "grad_norm": 0.593421995639801,
      "learning_rate": 0.00125,
      "loss": 2.5176,
      "step": 10000
    },
    {
      "epoch": 3.59,
      "grad_norm": 0.5348471999168396,
      "learning_rate": 0.0013125,
      "loss": 2.5265,
      "step": 10500
    },
    {
      "epoch": 3.76,
      "grad_norm": 0.5802651047706604,
      "learning_rate": 0.001375,
      "loss": 2.5181,
      "step": 11000
    },
    {
      "epoch": 3.93,
      "grad_norm": 0.4744044542312622,
      "learning_rate": 0.0014375000000000002,
      "loss": 2.5037,
      "step": 11500
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.4771555359187316,
      "eval_loss": 2.6362032890319824,
      "eval_runtime": 3.3838,
      "eval_samples_per_second": 1327.204,
      "eval_steps_per_second": 10.639,
      "step": 11712
    },
    {
      "epoch": 4.1,
      "grad_norm": 0.5075201392173767,
      "learning_rate": 0.0015,
      "loss": 2.4663,
      "step": 12000
    },
    {
      "epoch": 4.27,
      "grad_norm": 0.47369861602783203,
      "learning_rate": 0.0015625,
      "loss": 2.4459,
      "step": 12500
    },
    {
      "epoch": 4.44,
      "grad_norm": 0.41244271397590637,
      "learning_rate": 0.001625,
      "loss": 2.4541,
      "step": 13000
    },
    {
      "epoch": 4.61,
      "grad_norm": 0.4284743070602417,
      "learning_rate": 0.0016875,
      "loss": 2.4717,
      "step": 13500
    },
    {
      "epoch": 4.78,
      "grad_norm": 0.4499437212944031,
      "learning_rate": 0.0017500000000000003,
      "loss": 2.4669,
      "step": 14000
    },
    {
      "epoch": 4.95,
      "grad_norm": 0.36233699321746826,
      "learning_rate": 0.0018124999999999999,
      "loss": 2.466,
      "step": 14500
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.4786510904573802,
      "eval_loss": 2.612313985824585,
      "eval_runtime": 3.4422,
      "eval_samples_per_second": 1304.691,
      "eval_steps_per_second": 10.458,
      "step": 14640
    },
    {
      "epoch": 5.12,
      "grad_norm": 0.444653183221817,
      "learning_rate": 0.001875,
      "loss": 2.405,
      "step": 15000
    },
    {
      "epoch": 5.29,
      "grad_norm": 0.41395285725593567,
      "learning_rate": 0.0019375000000000002,
      "loss": 2.4002,
      "step": 15500
    },
    {
      "epoch": 5.46,
      "grad_norm": 0.38018137216567993,
      "learning_rate": 0.002,
      "loss": 2.4142,
      "step": 16000
    },
    {
      "epoch": 5.64,
      "grad_norm": 0.3837694823741913,
      "learning_rate": 0.0020625,
      "loss": 2.4187,
      "step": 16500
    },
    {
      "epoch": 5.81,
      "grad_norm": 0.36019667983055115,
      "learning_rate": 0.002125,
      "loss": 2.4364,
      "step": 17000
    },
    {
      "epoch": 5.98,
      "grad_norm": 0.317461222410202,
      "learning_rate": 0.0021874999999999998,
      "loss": 2.4203,
      "step": 17500
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.48276605704848014,
      "eval_loss": 2.587817430496216,
      "eval_runtime": 3.4189,
      "eval_samples_per_second": 1313.564,
      "eval_steps_per_second": 10.53,
      "step": 17568
    },
    {
      "epoch": 6.15,
      "grad_norm": 0.32134848833084106,
      "learning_rate": 0.0022500000000000003,
      "loss": 2.3565,
      "step": 18000
    },
    {
      "epoch": 6.32,
      "grad_norm": 0.36488619446754456,
      "learning_rate": 0.0023125000000000003,
      "loss": 2.3698,
      "step": 18500
    },
    {
      "epoch": 6.49,
      "grad_norm": 0.348846435546875,
      "learning_rate": 0.002375,
      "loss": 2.3797,
      "step": 19000
    },
    {
      "epoch": 6.66,
      "grad_norm": 0.400886207818985,
      "learning_rate": 0.0024375,
      "loss": 2.3919,
      "step": 19500
    },
    {
      "epoch": 6.83,
      "grad_norm": 0.32355600595474243,
      "learning_rate": 0.0025,
      "loss": 2.3871,
      "step": 20000
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.48550644596279174,
      "eval_loss": 2.569087505340576,
      "eval_runtime": 3.425,
      "eval_samples_per_second": 1311.234,
      "eval_steps_per_second": 10.511,
      "step": 20496
    },
    {
      "epoch": 7.0,
      "grad_norm": 0.3506123423576355,
      "learning_rate": 0.0025625,
      "loss": 2.3935,
      "step": 20500
    },
    {
      "epoch": 7.17,
      "grad_norm": 0.3607357144355774,
      "learning_rate": 0.002625,
      "loss": 2.3145,
      "step": 21000
    },
    {
      "epoch": 7.34,
      "grad_norm": 0.35818299651145935,
      "learning_rate": 0.0026875000000000002,
      "loss": 2.3333,
      "step": 21500
    },
    {
      "epoch": 7.51,
      "grad_norm": 0.33337295055389404,
      "learning_rate": 0.00275,
      "loss": 2.3555,
      "step": 22000
    },
    {
      "epoch": 7.68,
      "grad_norm": 0.36606958508491516,
      "learning_rate": 0.0028125,
      "loss": 2.3567,
      "step": 22500
    },
    {
      "epoch": 7.86,
      "grad_norm": 0.3098103404045105,
      "learning_rate": 0.0028750000000000004,
      "loss": 2.367,
      "step": 23000
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.4879943614262646,
      "eval_loss": 2.556701183319092,
      "eval_runtime": 3.419,
      "eval_samples_per_second": 1313.536,
      "eval_steps_per_second": 10.529,
      "step": 23424
    },
    {
      "epoch": 8.03,
      "grad_norm": 0.36782288551330566,
      "learning_rate": 0.0029375,
      "loss": 2.3522,
      "step": 23500
    },
    {
      "epoch": 8.2,
      "grad_norm": 0.36634373664855957,
      "learning_rate": 0.003,
      "loss": 2.3004,
      "step": 24000
    },
    {
      "epoch": 8.37,
      "grad_norm": 0.32430076599121094,
      "learning_rate": 0.002715909090909091,
      "loss": 2.304,
      "step": 24500
    },
    {
      "epoch": 8.54,
      "grad_norm": 0.31956982612609863,
      "learning_rate": 0.0024318181818181817,
      "loss": 2.2997,
      "step": 25000
    },
    {
      "epoch": 8.71,
      "grad_norm": 0.3134121596813202,
      "learning_rate": 0.002147727272727273,
      "loss": 2.3008,
      "step": 25500
    },
    {
      "epoch": 8.88,
      "grad_norm": 0.3530263900756836,
      "learning_rate": 0.0018636363636363638,
      "loss": 2.2871,
      "step": 26000
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.49414840179045755,
      "eval_loss": 2.502568483352661,
      "eval_runtime": 3.3962,
      "eval_samples_per_second": 1322.366,
      "eval_steps_per_second": 10.6,
      "step": 26352
    },
    {
      "epoch": 9.05,
      "grad_norm": 0.33767953515052795,
      "learning_rate": 0.0015795454545454546,
      "loss": 2.2403,
      "step": 26500
    },
    {
      "epoch": 9.22,
      "grad_norm": 0.28383320569992065,
      "learning_rate": 0.0012954545454545456,
      "loss": 2.1669,
      "step": 27000
    },
    {
      "epoch": 9.39,
      "grad_norm": 0.32172703742980957,
      "learning_rate": 0.0010113636363636364,
      "loss": 2.1702,
      "step": 27500
    },
    {
      "epoch": 9.56,
      "grad_norm": 0.3210740387439728,
      "learning_rate": 0.0007272727272727273,
      "loss": 2.1621,
      "step": 28000
    },
    {
      "epoch": 9.73,
      "grad_norm": 0.33859506249427795,
      "learning_rate": 0.0004431818181818182,
      "loss": 2.1613,
      "step": 28500
    },
    {
      "epoch": 9.9,
      "grad_norm": 0.2973349690437317,
      "learning_rate": 0.0001590909090909091,
      "loss": 2.1368,
      "step": 29000
    },
    {
      "epoch": 10.0,
      "eval_accuracy": 0.499527488923604,
      "eval_loss": 2.4752440452575684,
      "eval_runtime": 3.3565,
      "eval_samples_per_second": 1338.013,
      "eval_steps_per_second": 10.726,
      "step": 29280
    },
    {
      "epoch": 10.0,
      "step": 29280,
      "total_flos": 2273237316403200.0,
      "train_loss": 2.5689867467828136,
      "train_runtime": 745.5496,
      "train_samples_per_second": 628.328,
      "train_steps_per_second": 39.273
    }
  ],
  "logging_steps": 500,
  "max_steps": 29280,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 10,
  "save_steps": 2000,
  "total_flos": 2273237316403200.0,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": null
}