{
  "best_global_step": 500,
  "best_metric": 0.0026578546967357397,
  "best_model_checkpoint": "outputs/checkpoints/checkpoint-500",
  "epoch": 0.24516564003554903,
  "eval_steps": 100,
  "global_step": 500,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.00490331280071098,
      "grad_norm": 38.97287368774414,
      "learning_rate": 1.8000000000000001e-06,
      "loss": 11.7965,
      "step": 10
    },
    {
      "epoch": 0.00980662560142196,
      "grad_norm": 37.44226837158203,
      "learning_rate": 3.8000000000000005e-06,
      "loss": 11.1348,
      "step": 20
    },
    {
      "epoch": 0.01470993840213294,
      "grad_norm": 39.490840911865234,
      "learning_rate": 5.8e-06,
      "loss": 10.3908,
      "step": 30
    },
    {
      "epoch": 0.01961325120284392,
      "grad_norm": 55.910770416259766,
      "learning_rate": 7.800000000000002e-06,
      "loss": 9.7397,
      "step": 40
    },
    {
      "epoch": 0.0245165640035549,
      "grad_norm": 69.5174789428711,
      "learning_rate": 9.800000000000001e-06,
      "loss": 9.0773,
      "step": 50
    },
    {
      "epoch": 0.02941987680426588,
      "grad_norm": 72.5235366821289,
      "learning_rate": 9.954773869346734e-06,
      "loss": 8.4558,
      "step": 60
    },
    {
      "epoch": 0.034323189604976864,
      "grad_norm": 74.18756103515625,
      "learning_rate": 9.904522613065327e-06,
      "loss": 7.8523,
      "step": 70
    },
    {
      "epoch": 0.03922650240568784,
      "grad_norm": 75.60600280761719,
      "learning_rate": 9.85427135678392e-06,
      "loss": 7.2893,
      "step": 80
    },
    {
      "epoch": 0.044129815206398824,
      "grad_norm": 76.96356201171875,
      "learning_rate": 9.804020100502513e-06,
      "loss": 6.7552,
      "step": 90
    },
    {
      "epoch": 0.0490331280071098,
      "grad_norm": 78.16697692871094,
      "learning_rate": 9.753768844221107e-06,
      "loss": 6.1989,
      "step": 100
    },
    {
      "epoch": 0.0490331280071098,
      "eval_bleu": 0.2568511156961756,
      "eval_exact_match": 0.02403846153846154,
      "eval_loss": 5.520397663116455,
      "eval_overlap_f1": 0.44806093548612735,
      "eval_rouge1": 0.43472462113402405,
      "eval_rouge2": 0.389460301819342,
      "eval_rougeL": 0.41802956544256537,
      "eval_runtime": 51.1798,
      "eval_samples_per_second": 3.908,
      "eval_steps_per_second": 0.254,
      "eval_substring_accuracy": 0.4375,
      "step": 100
    },
    {
      "epoch": 0.053936440807820785,
      "grad_norm": 78.8167953491211,
      "learning_rate": 9.7035175879397e-06,
      "loss": 5.6758,
      "step": 110
    },
    {
      "epoch": 0.05883975360853176,
      "grad_norm": 79.5491714477539,
      "learning_rate": 9.653266331658291e-06,
      "loss": 5.1469,
      "step": 120
    },
    {
      "epoch": 0.06374306640924274,
      "grad_norm": 79.28946685791016,
      "learning_rate": 9.603015075376885e-06,
      "loss": 4.607,
      "step": 130
    },
    {
      "epoch": 0.06864637920995373,
      "grad_norm": 79.43553924560547,
      "learning_rate": 9.552763819095478e-06,
      "loss": 4.0762,
      "step": 140
    },
    {
      "epoch": 0.0735496920106647,
      "grad_norm": 79.19751739501953,
      "learning_rate": 9.502512562814071e-06,
      "loss": 3.5677,
      "step": 150
    },
    {
      "epoch": 0.07845300481137568,
      "grad_norm": 77.9931411743164,
      "learning_rate": 9.452261306532664e-06,
      "loss": 3.0641,
      "step": 160
    },
    {
      "epoch": 0.08335631761208667,
      "grad_norm": 74.37619018554688,
      "learning_rate": 9.402010050251256e-06,
      "loss": 2.5681,
      "step": 170
    },
    {
      "epoch": 0.08825963041279765,
      "grad_norm": 69.27262878417969,
      "learning_rate": 9.35175879396985e-06,
      "loss": 2.1015,
      "step": 180
    },
    {
      "epoch": 0.09316294321350863,
      "grad_norm": 60.969303131103516,
      "learning_rate": 9.301507537688444e-06,
      "loss": 1.6754,
      "step": 190
    },
    {
      "epoch": 0.0980662560142196,
      "grad_norm": 50.06096649169922,
      "learning_rate": 9.251256281407036e-06,
      "loss": 1.3013,
      "step": 200
    },
    {
      "epoch": 0.0980662560142196,
      "eval_bleu": 0.14716894569508807,
      "eval_exact_match": 0.02403846153846154,
      "eval_loss": 0.9516403675079346,
      "eval_overlap_f1": 0.3883582231248927,
      "eval_rouge1": 0.36852942137992606,
      "eval_rouge2": 0.33485300994926664,
      "eval_rougeL": 0.35705470249538007,
      "eval_runtime": 62.2584,
      "eval_samples_per_second": 3.212,
      "eval_steps_per_second": 0.209,
      "eval_substring_accuracy": 0.2644230769230769,
      "step": 200
    },
    {
      "epoch": 0.10296956881493059,
      "grad_norm": 38.215309143066406,
      "learning_rate": 9.201005025125629e-06,
      "loss": 0.9937,
      "step": 210
    },
    {
      "epoch": 0.10787288161564157,
      "grad_norm": 26.703927993774414,
      "learning_rate": 9.150753768844222e-06,
      "loss": 0.7609,
      "step": 220
    },
    {
      "epoch": 0.11277619441635255,
      "grad_norm": 17.413555145263672,
      "learning_rate": 9.100502512562814e-06,
      "loss": 0.5952,
      "step": 230
    },
    {
      "epoch": 0.11767950721706352,
      "grad_norm": 10.940084457397461,
      "learning_rate": 9.050251256281409e-06,
      "loss": 0.4761,
      "step": 240
    },
    {
      "epoch": 0.12258282001777451,
      "grad_norm": 6.992828845977783,
      "learning_rate": 9e-06,
      "loss": 0.4064,
      "step": 250
    },
    {
      "epoch": 0.12748613281848548,
      "grad_norm": 4.597628593444824,
      "learning_rate": 8.949748743718593e-06,
      "loss": 0.3508,
      "step": 260
    },
    {
      "epoch": 0.13238944561919647,
      "grad_norm": 3.1910245418548584,
      "learning_rate": 8.899497487437187e-06,
      "loss": 0.2929,
      "step": 270
    },
    {
      "epoch": 0.13729275841990746,
      "grad_norm": 2.1990866661071777,
      "learning_rate": 8.84924623115578e-06,
      "loss": 0.241,
      "step": 280
    },
    {
      "epoch": 0.14219607122061842,
      "grad_norm": 1.5336138010025024,
      "learning_rate": 8.798994974874373e-06,
      "loss": 0.441,
      "step": 290
    },
    {
      "epoch": 0.1470993840213294,
      "grad_norm": 1.0439887046813965,
      "learning_rate": 8.748743718592966e-06,
      "loss": 0.0839,
      "step": 300
    },
    {
      "epoch": 0.1470993840213294,
      "eval_bleu": 0.8777994634083494,
      "eval_exact_match": 0.2548076923076923,
      "eval_loss": 0.06626231968402863,
      "eval_overlap_f1": 0.9088659585838987,
      "eval_rouge1": 0.9063033282561512,
      "eval_rouge2": 0.8942021365610976,
      "eval_rougeL": 0.9015459692112401,
      "eval_runtime": 22.2958,
      "eval_samples_per_second": 8.97,
      "eval_steps_per_second": 0.583,
      "eval_substring_accuracy": 0.33653846153846156,
      "step": 300
    },
    {
      "epoch": 0.1520026968220404,
      "grad_norm": 0.8918174505233765,
      "learning_rate": 8.698492462311558e-06,
      "loss": 0.0603,
      "step": 310
    },
    {
      "epoch": 0.15690600962275136,
      "grad_norm": 0.8070284724235535,
      "learning_rate": 8.648241206030151e-06,
      "loss": 0.0471,
      "step": 320
    },
    {
      "epoch": 0.16180932242346235,
      "grad_norm": 2.4222471714019775,
      "learning_rate": 8.597989949748744e-06,
      "loss": 0.0295,
      "step": 330
    },
    {
      "epoch": 0.16671263522417334,
      "grad_norm": 0.49611347913742065,
      "learning_rate": 8.547738693467338e-06,
      "loss": 0.0255,
      "step": 340
    },
    {
      "epoch": 0.1716159480248843,
      "grad_norm": 0.2609233558177948,
      "learning_rate": 8.497487437185931e-06,
      "loss": 0.0105,
      "step": 350
    },
    {
      "epoch": 0.1765192608255953,
      "grad_norm": 0.2548035979270935,
      "learning_rate": 8.447236180904524e-06,
      "loss": 0.0128,
      "step": 360
    },
    {
      "epoch": 0.18142257362630626,
      "grad_norm": 0.18216729164123535,
      "learning_rate": 8.396984924623116e-06,
      "loss": 0.0084,
      "step": 370
    },
    {
      "epoch": 0.18632588642701725,
      "grad_norm": 0.15114252269268036,
      "learning_rate": 8.346733668341709e-06,
      "loss": 0.0064,
      "step": 380
    },
    {
      "epoch": 0.19122919922772824,
      "grad_norm": 0.1348075270652771,
      "learning_rate": 8.296482412060302e-06,
      "loss": 0.0062,
      "step": 390
    },
    {
      "epoch": 0.1961325120284392,
      "grad_norm": 0.25008365511894226,
      "learning_rate": 8.246231155778895e-06,
      "loss": 0.0062,
      "step": 400
    },
    {
      "epoch": 0.1961325120284392,
      "eval_bleu": 0.8264721397464441,
      "eval_exact_match": 0.6009615384615384,
      "eval_loss": 0.0034012680407613516,
      "eval_overlap_f1": 0.9396967958900664,
      "eval_rouge1": 0.9287252310234548,
      "eval_rouge2": 0.924869521658068,
      "eval_rougeL": 0.92772120031963,
      "eval_runtime": 18.5847,
      "eval_samples_per_second": 10.762,
      "eval_steps_per_second": 0.7,
      "eval_substring_accuracy": 0.8942307692307693,
      "step": 400
    },
    {
      "epoch": 0.2010358248291502,
      "grad_norm": 0.11969061940908432,
      "learning_rate": 8.195979899497489e-06,
      "loss": 0.0054,
      "step": 410
    },
    {
      "epoch": 0.20593913762986119,
      "grad_norm": 0.10634017735719681,
      "learning_rate": 8.14572864321608e-06,
      "loss": 0.0053,
      "step": 420
    },
    {
      "epoch": 0.21084245043057215,
      "grad_norm": 0.10716363042593002,
      "learning_rate": 8.095477386934673e-06,
      "loss": 0.0052,
      "step": 430
    },
    {
      "epoch": 0.21574576323128314,
      "grad_norm": 0.09278230369091034,
      "learning_rate": 8.045226130653267e-06,
      "loss": 0.0051,
      "step": 440
    },
    {
      "epoch": 0.22064907603199413,
      "grad_norm": 0.08859017491340637,
      "learning_rate": 7.99497487437186e-06,
      "loss": 0.0049,
      "step": 450
    },
    {
      "epoch": 0.2255523888327051,
      "grad_norm": 0.08707090467214584,
      "learning_rate": 7.944723618090453e-06,
      "loss": 0.0053,
      "step": 460
    },
    {
      "epoch": 0.23045570163341608,
      "grad_norm": 0.13823187351226807,
      "learning_rate": 7.894472361809046e-06,
      "loss": 0.0047,
      "step": 470
    },
    {
      "epoch": 0.23535901443412705,
      "grad_norm": 0.07414552569389343,
      "learning_rate": 7.844221105527638e-06,
      "loss": 0.0042,
      "step": 480
    },
    {
      "epoch": 0.24026232723483804,
      "grad_norm": 0.08094857633113861,
      "learning_rate": 7.793969849246233e-06,
      "loss": 0.0046,
      "step": 490
    },
    {
      "epoch": 0.24516564003554903,
      "grad_norm": 0.07222504168748856,
      "learning_rate": 7.743718592964824e-06,
      "loss": 0.0041,
      "step": 500
    },
    {
      "epoch": 0.24516564003554903,
      "eval_bleu": 0.7604122865829173,
      "eval_exact_match": 0.4855769230769231,
      "eval_loss": 0.0026578546967357397,
      "eval_overlap_f1": 0.9140666246940682,
      "eval_rouge1": 0.8993297480179168,
      "eval_rouge2": 0.8943847041495613,
      "eval_rougeL": 0.897489490164824,
      "eval_runtime": 17.607,
      "eval_samples_per_second": 11.359,
      "eval_steps_per_second": 0.738,
      "eval_substring_accuracy": 0.9038461538461539,
      "step": 500
    }
  ],
  "logging_steps": 10,
  "max_steps": 2040,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 100,
  "stateful_callbacks": {
    "EarlyStoppingCallback": {
      "args": {
        "early_stopping_patience": 3,
        "early_stopping_threshold": 0.001
      },
      "attributes": {
        "early_stopping_patience_counter": 1
      }
    },
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 2.77392610295808e+17,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}