{
  "best_metric": 0.6806846415754118,
  "best_model_checkpoint": "output/pretraining/vihealthbert-w_unsup-SynPD/lr3e-5_wr0.1_wd0.0/checkpoint-42000",
  "epoch": 7.237635705669481,
  "eval_steps": 2000,
  "global_step": 42000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.00017232465965879716,
      "grad_norm": 35.608882904052734,
      "learning_rate": 5.169739789763915e-09,
      "loss": 10.7382,
      "step": 1
    },
    {
      "epoch": 0.34464931931759435,
      "grad_norm": 4.884454250335693,
      "learning_rate": 1.033947957952783e-05,
      "loss": 5.8543,
      "step": 2000
    },
    {
      "epoch": 0.34464931931759435,
      "eval_accuracy": 0.3950195480446223,
      "eval_loss": 3.896709680557251,
      "eval_runtime": 24.6709,
      "eval_samples_per_second": 396.175,
      "eval_steps_per_second": 24.766,
      "step": 2000
    },
    {
      "epoch": 0.6892986386351887,
      "grad_norm": 5.278345108032227,
      "learning_rate": 2.067895915905566e-05,
      "loss": 3.4544,
      "step": 4000
    },
    {
      "epoch": 0.6892986386351887,
      "eval_accuracy": 0.530626322082222,
      "eval_loss": 2.811899423599243,
      "eval_runtime": 24.6443,
      "eval_samples_per_second": 396.602,
      "eval_steps_per_second": 24.793,
      "step": 4000
    },
    {
      "epoch": 1.033947957952783,
      "grad_norm": 4.731122016906738,
      "learning_rate": 2.988684014015739e-05,
      "loss": 2.8312,
      "step": 6000
    },
    {
      "epoch": 1.033947957952783,
      "eval_accuracy": 0.5771132321637026,
      "eval_loss": 2.4039628505706787,
      "eval_runtime": 24.6266,
      "eval_samples_per_second": 396.889,
      "eval_steps_per_second": 24.811,
      "step": 6000
    },
    {
      "epoch": 1.3785972772703774,
      "grad_norm": 12.213573455810547,
      "learning_rate": 2.873800907576541e-05,
      "loss": 2.5914,
      "step": 8000
    },
    {
      "epoch": 1.3785972772703774,
      "eval_accuracy": 0.5349752225383133,
      "eval_loss": 2.6481692790985107,
      "eval_runtime": 24.6033,
      "eval_samples_per_second": 397.264,
      "eval_steps_per_second": 24.834,
      "step": 8000
    },
    {
      "epoch": 1.7232465965879717,
      "grad_norm": 5.172228813171387,
      "learning_rate": 2.758917801137343e-05,
      "loss": 2.5649,
      "step": 10000
    },
    {
      "epoch": 1.7232465965879717,
      "eval_accuracy": 0.6087076297937304,
      "eval_loss": 2.13348388671875,
      "eval_runtime": 24.6211,
      "eval_samples_per_second": 396.976,
      "eval_steps_per_second": 24.816,
      "step": 10000
    },
    {
      "epoch": 2.067895915905566,
      "grad_norm": 4.871993064880371,
      "learning_rate": 2.6440346946981447e-05,
      "loss": 2.2749,
      "step": 12000
    },
    {
      "epoch": 2.067895915905566,
      "eval_accuracy": 0.6281529792701336,
      "eval_loss": 1.9895449876785278,
      "eval_runtime": 24.453,
      "eval_samples_per_second": 399.705,
      "eval_steps_per_second": 24.987,
      "step": 12000
    },
    {
      "epoch": 2.4125452352231607,
      "grad_norm": 8.037242889404297,
      "learning_rate": 2.5291515882589467e-05,
      "loss": 2.1572,
      "step": 14000
    },
    {
      "epoch": 2.4125452352231607,
      "eval_accuracy": 0.6353159252095423,
      "eval_loss": 1.9312896728515625,
      "eval_runtime": 24.4658,
      "eval_samples_per_second": 399.496,
      "eval_steps_per_second": 24.974,
      "step": 14000
    },
    {
      "epoch": 2.757194554540755,
      "grad_norm": 4.320974349975586,
      "learning_rate": 2.4142684818197487e-05,
      "loss": 2.1009,
      "step": 16000
    },
    {
      "epoch": 2.757194554540755,
      "eval_accuracy": 0.6428579553856829,
      "eval_loss": 1.864585280418396,
      "eval_runtime": 24.4545,
      "eval_samples_per_second": 399.68,
      "eval_steps_per_second": 24.985,
      "step": 16000
    },
    {
      "epoch": 3.101843873858349,
      "grad_norm": 4.6416168212890625,
      "learning_rate": 2.2993853753805504e-05,
      "loss": 2.0609,
      "step": 18000
    },
    {
      "epoch": 3.101843873858349,
      "eval_accuracy": 0.6450457483325106,
      "eval_loss": 1.8571828603744507,
      "eval_runtime": 24.4634,
      "eval_samples_per_second": 399.535,
      "eval_steps_per_second": 24.976,
      "step": 18000
    },
    {
      "epoch": 3.4464931931759435,
      "grad_norm": 4.341440677642822,
      "learning_rate": 2.184502268941352e-05,
      "loss": 2.0885,
      "step": 20000
    },
    {
      "epoch": 3.4464931931759435,
      "eval_accuracy": 0.6284792764252838,
      "eval_loss": 1.9489394426345825,
      "eval_runtime": 24.6433,
      "eval_samples_per_second": 396.618,
      "eval_steps_per_second": 24.794,
      "step": 20000
    },
    {
      "epoch": 3.791142512493538,
      "grad_norm": 4.660053253173828,
      "learning_rate": 2.069619162502154e-05,
      "loss": 1.9891,
      "step": 22000
    },
    {
      "epoch": 3.791142512493538,
      "eval_accuracy": 0.6582960564701575,
      "eval_loss": 1.7700080871582031,
      "eval_runtime": 24.616,
      "eval_samples_per_second": 397.059,
      "eval_steps_per_second": 24.821,
      "step": 22000
    },
    {
      "epoch": 4.135791831811132,
      "grad_norm": 3.9445292949676514,
      "learning_rate": 1.954736056062956e-05,
      "loss": 1.9368,
      "step": 24000
    },
    {
      "epoch": 4.135791831811132,
      "eval_accuracy": 0.6608816779512006,
      "eval_loss": 1.7397598028182983,
      "eval_runtime": 24.6251,
      "eval_samples_per_second": 396.912,
      "eval_steps_per_second": 24.812,
      "step": 24000
    },
    {
      "epoch": 4.480441151128726,
      "grad_norm": 4.561331748962402,
      "learning_rate": 1.839852949623758e-05,
      "loss": 1.9003,
      "step": 26000
    },
    {
      "epoch": 4.480441151128726,
      "eval_accuracy": 0.666368433126864,
      "eval_loss": 1.7164973020553589,
      "eval_runtime": 24.4665,
      "eval_samples_per_second": 399.485,
      "eval_steps_per_second": 24.973,
      "step": 26000
    },
    {
      "epoch": 4.825090470446321,
      "grad_norm": 4.358778476715088,
      "learning_rate": 1.72496984318456e-05,
      "loss": 1.9058,
      "step": 28000
    },
    {
      "epoch": 4.825090470446321,
      "eval_accuracy": 0.666961039556835,
      "eval_loss": 1.703224539756775,
      "eval_runtime": 24.4455,
      "eval_samples_per_second": 399.829,
      "eval_steps_per_second": 24.994,
      "step": 28000
    },
    {
      "epoch": 5.1697397897639155,
      "grad_norm": 4.522342205047607,
      "learning_rate": 1.6100867367453616e-05,
      "loss": 1.859,
      "step": 30000
    },
    {
      "epoch": 5.1697397897639155,
      "eval_accuracy": 0.6717793867023528,
      "eval_loss": 1.677147626876831,
      "eval_runtime": 24.4697,
      "eval_samples_per_second": 399.432,
      "eval_steps_per_second": 24.97,
      "step": 30000
    },
    {
      "epoch": 5.51438910908151,
      "grad_norm": 4.162671089172363,
      "learning_rate": 1.4952036303061636e-05,
      "loss": 1.8401,
      "step": 32000
    },
    {
      "epoch": 5.51438910908151,
      "eval_accuracy": 0.6709595439247201,
      "eval_loss": 1.665189504623413,
      "eval_runtime": 24.6261,
      "eval_samples_per_second": 396.897,
      "eval_steps_per_second": 24.811,
      "step": 32000
    },
    {
      "epoch": 5.859038428399104,
      "grad_norm": 4.618478775024414,
      "learning_rate": 1.3803205238669653e-05,
      "loss": 1.8181,
      "step": 34000
    },
    {
      "epoch": 5.859038428399104,
      "eval_accuracy": 0.675394469972155,
      "eval_loss": 1.6417487859725952,
      "eval_runtime": 24.5944,
      "eval_samples_per_second": 397.408,
      "eval_steps_per_second": 24.843,
      "step": 34000
    },
    {
      "epoch": 6.203687747716698,
      "grad_norm": 4.046966075897217,
      "learning_rate": 1.2654374174277672e-05,
      "loss": 1.8133,
      "step": 36000
    },
    {
      "epoch": 6.203687747716698,
      "eval_accuracy": 0.6748118789132058,
      "eval_loss": 1.6430697441101074,
      "eval_runtime": 24.6651,
      "eval_samples_per_second": 396.268,
      "eval_steps_per_second": 24.772,
      "step": 36000
    },
    {
      "epoch": 6.548337067034293,
      "grad_norm": 4.522408962249756,
      "learning_rate": 1.1505543109885692e-05,
      "loss": 1.7888,
      "step": 38000
    },
    {
      "epoch": 6.548337067034293,
      "eval_accuracy": 0.6755271447952356,
      "eval_loss": 1.6363412141799927,
      "eval_runtime": 24.6146,
      "eval_samples_per_second": 397.081,
      "eval_steps_per_second": 24.823,
      "step": 38000
    },
    {
      "epoch": 6.892986386351887,
      "grad_norm": 4.334969520568848,
      "learning_rate": 1.035671204549371e-05,
      "loss": 1.7811,
      "step": 40000
    },
    {
      "epoch": 6.892986386351887,
      "eval_accuracy": 0.679264661758257,
      "eval_loss": 1.6205224990844727,
      "eval_runtime": 24.6327,
      "eval_samples_per_second": 396.79,
      "eval_steps_per_second": 24.804,
      "step": 40000
    },
    {
      "epoch": 7.237635705669481,
      "grad_norm": 3.82973313331604,
      "learning_rate": 9.20788098110173e-06,
      "loss": 1.7481,
      "step": 42000
    },
    {
      "epoch": 7.237635705669481,
      "eval_accuracy": 0.6806846415754118,
      "eval_loss": 1.61899995803833,
      "eval_runtime": 24.6165,
      "eval_samples_per_second": 397.051,
      "eval_steps_per_second": 24.821,
      "step": 42000
    }
  ],
  "logging_steps": 2000,
  "max_steps": 58030,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 10,
  "save_steps": 2000,
  "total_flos": 0.0,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": null
}