{
  "best_metric": 0.9453551912568307,
  "best_model_checkpoint": "./neikexue_v2_seq_512__1e-5_model_results/checkpoint-3000",
  "epoch": 26.08695652173913,
  "eval_steps": 500,
  "global_step": 3000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.8695652173913043,
      "grad_norm": 7.171332836151123,
      "learning_rate": 9.713043478260871e-06,
      "loss": 5.7808,
      "step": 100
    },
    {
      "epoch": 1.7391304347826086,
      "grad_norm": 8.207396507263184,
      "learning_rate": 9.423188405797102e-06,
      "loss": 5.498,
      "step": 200
    },
    {
      "epoch": 2.608695652173913,
      "grad_norm": 9.517407417297363,
      "learning_rate": 9.133333333333335e-06,
      "loss": 4.8788,
      "step": 300
    },
    {
      "epoch": 3.4782608695652173,
      "grad_norm": 9.667986869812012,
      "learning_rate": 8.843478260869565e-06,
      "loss": 4.2702,
      "step": 400
    },
    {
      "epoch": 4.3478260869565215,
      "grad_norm": 10.862434387207031,
      "learning_rate": 8.553623188405798e-06,
      "loss": 3.7082,
      "step": 500
    },
    {
      "epoch": 4.3478260869565215,
      "eval_accuracy": 0.5109289617486339,
      "eval_f1": 0.4128540311348767,
      "eval_loss": 3.472090244293213,
      "eval_precision": 0.3813047710709234,
      "eval_recall": 0.5109289617486339,
      "eval_runtime": 2.3816,
      "eval_samples_per_second": 153.679,
      "eval_steps_per_second": 9.657,
      "step": 500
    },
    {
      "epoch": 5.217391304347826,
      "grad_norm": 10.429506301879883,
      "learning_rate": 8.26376811594203e-06,
      "loss": 3.232,
      "step": 600
    },
    {
      "epoch": 6.086956521739131,
      "grad_norm": 9.496939659118652,
      "learning_rate": 7.973913043478261e-06,
      "loss": 2.7575,
      "step": 700
    },
    {
      "epoch": 6.956521739130435,
      "grad_norm": 8.775982856750488,
      "learning_rate": 7.684057971014494e-06,
      "loss": 2.3513,
      "step": 800
    },
    {
      "epoch": 7.826086956521739,
      "grad_norm": 9.64019775390625,
      "learning_rate": 7.394202898550725e-06,
      "loss": 1.9976,
      "step": 900
    },
    {
      "epoch": 8.695652173913043,
      "grad_norm": 8.96278190612793,
      "learning_rate": 7.104347826086957e-06,
      "loss": 1.6696,
      "step": 1000
    },
    {
      "epoch": 8.695652173913043,
      "eval_accuracy": 0.8661202185792349,
      "eval_f1": 0.8266237557221163,
      "eval_loss": 1.9026175737380981,
      "eval_precision": 0.8064055859137825,
      "eval_recall": 0.8661202185792349,
      "eval_runtime": 2.4785,
      "eval_samples_per_second": 147.672,
      "eval_steps_per_second": 9.28,
      "step": 1000
    },
    {
      "epoch": 9.565217391304348,
      "grad_norm": 7.726362228393555,
      "learning_rate": 6.814492753623188e-06,
      "loss": 1.4203,
      "step": 1100
    },
    {
      "epoch": 10.434782608695652,
      "grad_norm": 7.591976642608643,
      "learning_rate": 6.52463768115942e-06,
      "loss": 1.1892,
      "step": 1200
    },
    {
      "epoch": 11.304347826086957,
      "grad_norm": 6.4590020179748535,
      "learning_rate": 6.234782608695653e-06,
      "loss": 1.0007,
      "step": 1300
    },
    {
      "epoch": 12.173913043478262,
      "grad_norm": 5.544084548950195,
      "learning_rate": 5.944927536231885e-06,
      "loss": 0.8247,
      "step": 1400
    },
    {
      "epoch": 13.043478260869565,
      "grad_norm": 5.065005779266357,
      "learning_rate": 5.655072463768117e-06,
      "loss": 0.7049,
      "step": 1500
    },
    {
      "epoch": 13.043478260869565,
      "eval_accuracy": 0.9316939890710383,
      "eval_f1": 0.9106991066007459,
      "eval_loss": 1.0993165969848633,
      "eval_precision": 0.9006830601092896,
      "eval_recall": 0.9316939890710383,
      "eval_runtime": 2.3962,
      "eval_samples_per_second": 152.745,
      "eval_steps_per_second": 9.599,
      "step": 1500
    },
    {
      "epoch": 13.91304347826087,
      "grad_norm": 4.597533702850342,
      "learning_rate": 5.365217391304348e-06,
      "loss": 0.5942,
      "step": 1600
    },
    {
      "epoch": 14.782608695652174,
      "grad_norm": 3.6905860900878906,
      "learning_rate": 5.07536231884058e-06,
      "loss": 0.4948,
      "step": 1700
    },
    {
      "epoch": 15.652173913043478,
      "grad_norm": 3.5639991760253906,
      "learning_rate": 4.785507246376812e-06,
      "loss": 0.4243,
      "step": 1800
    },
    {
      "epoch": 16.52173913043478,
      "grad_norm": 2.8523664474487305,
      "learning_rate": 4.495652173913044e-06,
      "loss": 0.3603,
      "step": 1900
    },
    {
      "epoch": 17.391304347826086,
      "grad_norm": 3.5141196250915527,
      "learning_rate": 4.205797101449276e-06,
      "loss": 0.3128,
      "step": 2000
    },
    {
      "epoch": 17.391304347826086,
      "eval_accuracy": 0.9371584699453552,
      "eval_f1": 0.9183190215977102,
      "eval_loss": 0.7386174201965332,
      "eval_precision": 0.910792349726776,
      "eval_recall": 0.9371584699453552,
      "eval_runtime": 2.5874,
      "eval_samples_per_second": 141.456,
      "eval_steps_per_second": 8.889,
      "step": 2000
    },
    {
      "epoch": 18.26086956521739,
      "grad_norm": 2.3429932594299316,
      "learning_rate": 3.915942028985508e-06,
      "loss": 0.2712,
      "step": 2100
    },
    {
      "epoch": 19.130434782608695,
      "grad_norm": 1.9925650358200073,
      "learning_rate": 3.6260869565217393e-06,
      "loss": 0.2384,
      "step": 2200
    },
    {
      "epoch": 20.0,
      "grad_norm": 2.1215438842773438,
      "learning_rate": 3.336231884057971e-06,
      "loss": 0.2121,
      "step": 2300
    },
    {
      "epoch": 20.869565217391305,
      "grad_norm": 1.7758429050445557,
      "learning_rate": 3.0463768115942034e-06,
      "loss": 0.1882,
      "step": 2400
    },
    {
      "epoch": 21.73913043478261,
      "grad_norm": 1.4673223495483398,
      "learning_rate": 2.756521739130435e-06,
      "loss": 0.1714,
      "step": 2500
    },
    {
      "epoch": 21.73913043478261,
      "eval_accuracy": 0.9371584699453552,
      "eval_f1": 0.9192601266371758,
      "eval_loss": 0.5610440969467163,
      "eval_precision": 0.9120673952641166,
      "eval_recall": 0.9371584699453552,
      "eval_runtime": 2.4135,
      "eval_samples_per_second": 151.65,
      "eval_steps_per_second": 9.53,
      "step": 2500
    },
    {
      "epoch": 22.608695652173914,
      "grad_norm": 1.3567324876785278,
      "learning_rate": 2.466666666666667e-06,
      "loss": 0.1535,
      "step": 2600
    },
    {
      "epoch": 23.47826086956522,
      "grad_norm": 1.2968732118606567,
      "learning_rate": 2.1768115942028987e-06,
      "loss": 0.1424,
      "step": 2700
    },
    {
      "epoch": 24.347826086956523,
      "grad_norm": 1.2187700271606445,
      "learning_rate": 1.8869565217391305e-06,
      "loss": 0.1318,
      "step": 2800
    },
    {
      "epoch": 25.217391304347824,
      "grad_norm": 1.1091492176055908,
      "learning_rate": 1.5971014492753626e-06,
      "loss": 0.123,
      "step": 2900
    },
    {
      "epoch": 26.08695652173913,
      "grad_norm": 1.013292670249939,
      "learning_rate": 1.3072463768115942e-06,
      "loss": 0.1163,
      "step": 3000
    },
    {
      "epoch": 26.08695652173913,
      "eval_accuracy": 0.9453551912568307,
      "eval_f1": 0.9292783415734236,
      "eval_loss": 0.4996342957019806,
      "eval_precision": 0.9220856102003644,
      "eval_recall": 0.9453551912568307,
      "eval_runtime": 2.3949,
      "eval_samples_per_second": 152.826,
      "eval_steps_per_second": 9.604,
      "step": 3000
    }
  ],
  "logging_steps": 100,
  "max_steps": 3450,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 30,
  "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": 4.453760493084672e+16,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": null
}