{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 1000.0,
  "eval_steps": 500,
  "global_step": 8000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 62.5,
      "grad_norm": 0.2072206288576126,
      "learning_rate": 0.0009375,
      "loss": 0.3349,
      "step": 500
    },
    {
      "epoch": 62.5,
      "eval_loss": 0.21073544025421143,
      "eval_runtime": 1.7576,
      "eval_samples_per_second": 4470.361,
      "eval_steps_per_second": 35.276,
      "step": 500
    },
    {
      "epoch": 125.0,
      "grad_norm": 0.20736312866210938,
      "learning_rate": 0.000875,
      "loss": 0.1147,
      "step": 1000
    },
    {
      "epoch": 125.0,
      "eval_loss": 0.2409573197364807,
      "eval_runtime": 1.7582,
      "eval_samples_per_second": 4468.854,
      "eval_steps_per_second": 35.264,
      "step": 1000
    },
    {
      "epoch": 187.5,
      "grad_norm": 0.17418691515922546,
      "learning_rate": 0.0008125000000000001,
      "loss": 0.071,
      "step": 1500
    },
    {
      "epoch": 187.5,
      "eval_loss": 0.2686581015586853,
      "eval_runtime": 1.7698,
      "eval_samples_per_second": 4439.56,
      "eval_steps_per_second": 35.033,
      "step": 1500
    },
    {
      "epoch": 250.0,
      "grad_norm": 0.1891753375530243,
      "learning_rate": 0.00075,
      "loss": 0.0502,
      "step": 2000
    },
    {
      "epoch": 250.0,
      "eval_loss": 0.29014259576797485,
      "eval_runtime": 1.7819,
      "eval_samples_per_second": 4409.274,
      "eval_steps_per_second": 34.794,
      "step": 2000
    },
    {
      "epoch": 312.5,
      "grad_norm": 0.16434858739376068,
      "learning_rate": 0.0006875,
      "loss": 0.0373,
      "step": 2500
    },
    {
      "epoch": 312.5,
      "eval_loss": 0.3033287525177002,
      "eval_runtime": 1.7585,
      "eval_samples_per_second": 4468.05,
      "eval_steps_per_second": 35.258,
      "step": 2500
    },
    {
      "epoch": 375.0,
      "grad_norm": 0.11065812408924103,
      "learning_rate": 0.000625,
      "loss": 0.0301,
      "step": 3000
    },
    {
      "epoch": 375.0,
      "eval_loss": 0.31410983204841614,
      "eval_runtime": 1.7765,
      "eval_samples_per_second": 4422.629,
      "eval_steps_per_second": 34.899,
      "step": 3000
    },
    {
      "epoch": 437.5,
      "grad_norm": 0.12541820108890533,
      "learning_rate": 0.0005625000000000001,
      "loss": 0.025,
      "step": 3500
    },
    {
      "epoch": 437.5,
      "eval_loss": 0.3234836161136627,
      "eval_runtime": 1.8714,
      "eval_samples_per_second": 4198.44,
      "eval_steps_per_second": 33.13,
      "step": 3500
    },
    {
      "epoch": 500.0,
      "grad_norm": 0.12353639304637909,
      "learning_rate": 0.0005,
      "loss": 0.0212,
      "step": 4000
    },
    {
      "epoch": 500.0,
      "eval_loss": 0.33116382360458374,
      "eval_runtime": 2.0737,
      "eval_samples_per_second": 3788.95,
      "eval_steps_per_second": 29.899,
      "step": 4000
    },
    {
      "epoch": 562.5,
      "grad_norm": 0.14147008955478668,
      "learning_rate": 0.0004375,
      "loss": 0.0187,
      "step": 4500
    },
    {
      "epoch": 562.5,
      "eval_loss": 0.3403891623020172,
      "eval_runtime": 1.7439,
      "eval_samples_per_second": 4505.522,
      "eval_steps_per_second": 35.553,
      "step": 4500
    },
    {
      "epoch": 625.0,
      "grad_norm": 0.17466695606708527,
      "learning_rate": 0.000375,
      "loss": 0.017,
      "step": 5000
    },
    {
      "epoch": 625.0,
      "eval_loss": 0.3371485471725464,
      "eval_runtime": 1.7727,
      "eval_samples_per_second": 4432.252,
      "eval_steps_per_second": 34.975,
      "step": 5000
    },
    {
      "epoch": 687.5,
      "grad_norm": 0.1278233528137207,
      "learning_rate": 0.0003125,
      "loss": 0.0148,
      "step": 5500
    },
    {
      "epoch": 687.5,
      "eval_loss": 0.3466494679450989,
      "eval_runtime": 1.7776,
      "eval_samples_per_second": 4420.019,
      "eval_steps_per_second": 34.879,
      "step": 5500
    },
    {
      "epoch": 750.0,
      "grad_norm": 0.08440423011779785,
      "learning_rate": 0.00025,
      "loss": 0.0139,
      "step": 6000
    },
    {
      "epoch": 750.0,
      "eval_loss": 0.34804055094718933,
      "eval_runtime": 1.7744,
      "eval_samples_per_second": 4427.887,
      "eval_steps_per_second": 34.941,
      "step": 6000
    },
    {
      "epoch": 812.5,
      "grad_norm": 0.1064932644367218,
      "learning_rate": 0.0001875,
      "loss": 0.0124,
      "step": 6500
    },
    {
      "epoch": 812.5,
      "eval_loss": 0.3551761209964752,
      "eval_runtime": 1.7547,
      "eval_samples_per_second": 4477.816,
      "eval_steps_per_second": 35.335,
      "step": 6500
    },
    {
      "epoch": 875.0,
      "grad_norm": 0.07584909349679947,
      "learning_rate": 0.000125,
      "loss": 0.0118,
      "step": 7000
    },
    {
      "epoch": 875.0,
      "eval_loss": 0.3594026267528534,
      "eval_runtime": 1.7559,
      "eval_samples_per_second": 4474.601,
      "eval_steps_per_second": 35.309,
      "step": 7000
    },
    {
      "epoch": 937.5,
      "grad_norm": 0.04748733341693878,
      "learning_rate": 6.25e-05,
      "loss": 0.0112,
      "step": 7500
    },
    {
      "epoch": 937.5,
      "eval_loss": 0.35810717940330505,
      "eval_runtime": 1.7577,
      "eval_samples_per_second": 4469.966,
      "eval_steps_per_second": 35.273,
      "step": 7500
    },
    {
      "epoch": 1000.0,
      "grad_norm": 0.0842234194278717,
      "learning_rate": 0.0,
      "loss": 0.0106,
      "step": 8000
    },
    {
      "epoch": 1000.0,
      "eval_loss": 0.35844165086746216,
      "eval_runtime": 1.7498,
      "eval_samples_per_second": 4490.355,
      "eval_steps_per_second": 35.434,
      "step": 8000
    },
    {
      "epoch": 1000.0,
      "step": 8000,
      "total_flos": 3.308878023570227e+16,
      "train_loss": 0.04967061561346054,
      "train_runtime": 721.4078,
      "train_samples_per_second": 1386.178,
      "train_steps_per_second": 11.089
    }
  ],
  "logging_steps": 500,
  "max_steps": 8000,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1000,
  "save_steps": 500,
  "total_flos": 3.308878023570227e+16,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": null
}