max_seq_length_128_experiments/distilbert/distilroberta_base_twitter/checkpoint-200/trainer_state.json

{
  "best_metric": 0.48123905062675476,
  "best_model_checkpoint": "../../experiments_checkpoints/MAdAiLab/distilbert/distilroberta_base_twitter/checkpoint-200",
  "epoch": 1.4705882352941178,
  "eval_steps": 50,
  "global_step": 200,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.07,
      "grad_norm": 1.314621925354004,
      "learning_rate": 1.950980392156863e-05,
      "loss": 0.6097,
      "step": 10
    },
    {
      "epoch": 0.15,
      "grad_norm": 2.451284408569336,
      "learning_rate": 1.9019607843137255e-05,
      "loss": 0.6246,
      "step": 20
    },
    {
      "epoch": 0.22,
      "grad_norm": 8.775372505187988,
      "learning_rate": 1.8529411764705884e-05,
      "loss": 0.5457,
      "step": 30
    },
    {
      "epoch": 0.29,
      "grad_norm": 3.2106316089630127,
      "learning_rate": 1.8039215686274513e-05,
      "loss": 0.5083,
      "step": 40
    },
    {
      "epoch": 0.37,
      "grad_norm": 2.1298840045928955,
      "learning_rate": 1.7549019607843138e-05,
      "loss": 0.4863,
      "step": 50
    },
    {
      "epoch": 0.37,
      "eval_accuracy": 0.7582720588235294,
      "eval_f1_macro": 0.7215736986471983,
      "eval_f1_micro": 0.7582720588235294,
      "eval_loss": 0.4847523272037506,
      "eval_runtime": 0.5343,
      "eval_samples_per_second": 2036.144,
      "eval_steps_per_second": 31.815,
      "step": 50
    },
    {
      "epoch": 0.44,
      "grad_norm": 2.0992162227630615,
      "learning_rate": 1.7058823529411767e-05,
      "loss": 0.479,
      "step": 60
    },
    {
      "epoch": 0.51,
      "grad_norm": 4.017960548400879,
      "learning_rate": 1.6568627450980395e-05,
      "loss": 0.5089,
      "step": 70
    },
    {
      "epoch": 0.59,
      "grad_norm": 2.547389268875122,
      "learning_rate": 1.607843137254902e-05,
      "loss": 0.4534,
      "step": 80
    },
    {
      "epoch": 0.66,
      "grad_norm": 5.25526762008667,
      "learning_rate": 1.558823529411765e-05,
      "loss": 0.5069,
      "step": 90
    },
    {
      "epoch": 0.74,
      "grad_norm": 2.0313820838928223,
      "learning_rate": 1.5098039215686276e-05,
      "loss": 0.4736,
      "step": 100
    },
    {
      "epoch": 0.74,
      "eval_accuracy": 0.7628676470588235,
      "eval_f1_macro": 0.7078921078921079,
      "eval_f1_micro": 0.7628676470588235,
      "eval_loss": 0.486197292804718,
      "eval_runtime": 0.5322,
      "eval_samples_per_second": 2044.173,
      "eval_steps_per_second": 31.94,
      "step": 100
    },
    {
      "epoch": 0.81,
      "grad_norm": 3.0662012100219727,
      "learning_rate": 1.4607843137254903e-05,
      "loss": 0.4831,
      "step": 110
    },
    {
      "epoch": 0.88,
      "grad_norm": 2.1004221439361572,
      "learning_rate": 1.4117647058823532e-05,
      "loss": 0.4413,
      "step": 120
    },
    {
      "epoch": 0.96,
      "grad_norm": 5.478739261627197,
      "learning_rate": 1.3627450980392158e-05,
      "loss": 0.4581,
      "step": 130
    },
    {
      "epoch": 1.03,
      "grad_norm": 3.0216825008392334,
      "learning_rate": 1.3137254901960785e-05,
      "loss": 0.4658,
      "step": 140
    },
    {
      "epoch": 1.1,
      "grad_norm": 2.5667436122894287,
      "learning_rate": 1.2647058823529412e-05,
      "loss": 0.463,
      "step": 150
    },
    {
      "epoch": 1.1,
      "eval_accuracy": 0.7738970588235294,
      "eval_f1_macro": 0.7356166061472933,
      "eval_f1_micro": 0.7738970588235294,
      "eval_loss": 0.49347999691963196,
      "eval_runtime": 0.5316,
      "eval_samples_per_second": 2046.653,
      "eval_steps_per_second": 31.979,
      "step": 150
    },
    {
      "epoch": 1.18,
      "grad_norm": 3.2256863117218018,
      "learning_rate": 1.215686274509804e-05,
      "loss": 0.4481,
      "step": 160
    },
    {
      "epoch": 1.25,
      "grad_norm": 2.5312764644622803,
      "learning_rate": 1.1666666666666668e-05,
      "loss": 0.4336,
      "step": 170
    },
    {
      "epoch": 1.32,
      "grad_norm": 3.472776174545288,
      "learning_rate": 1.1176470588235295e-05,
      "loss": 0.4159,
      "step": 180
    },
    {
      "epoch": 1.4,
      "grad_norm": 3.2552549839019775,
      "learning_rate": 1.0686274509803922e-05,
      "loss": 0.4319,
      "step": 190
    },
    {
      "epoch": 1.47,
      "grad_norm": 2.458969831466675,
      "learning_rate": 1.0196078431372549e-05,
      "loss": 0.4703,
      "step": 200
    },
    {
      "epoch": 1.47,
      "eval_accuracy": 0.7720588235294118,
      "eval_f1_macro": 0.7372512970724334,
      "eval_f1_micro": 0.7720588235294118,
      "eval_loss": 0.48123905062675476,
      "eval_runtime": 0.5286,
      "eval_samples_per_second": 2058.189,
      "eval_steps_per_second": 32.159,
      "step": 200
    }
  ],
  "logging_steps": 10,
  "max_steps": 408,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 50,
  "total_flos": 423895656038400.0,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": null
}