checkpoint-72/trainer_state.json

{
  "best_global_step": 27,
  "best_metric": 0.5779597759246826,
  "best_model_checkpoint": "//outputs/task7_microsoft/Phi-3.5-mini-instruct/checkpoint-27",
  "epoch": 8.0,
  "eval_steps": 500,
  "global_step": 72,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.5882352941176471,
      "grad_norm": 0.2979797124862671,
      "learning_rate": 0.00015,
      "loss": 0.7889,
      "step": 5
    },
    {
      "epoch": 1.0,
      "eval_loss": 0.6737979650497437,
      "eval_runtime": 3.3675,
      "eval_samples_per_second": 4.454,
      "eval_steps_per_second": 0.594,
      "step": 9
    },
    {
      "epoch": 1.1176470588235294,
      "grad_norm": 0.20464661717414856,
      "learning_rate": 0.0003,
      "loss": 0.6771,
      "step": 10
    },
    {
      "epoch": 1.7058823529411766,
      "grad_norm": 0.1992536187171936,
      "learning_rate": 0.00029948767395100045,
      "loss": 0.6255,
      "step": 15
    },
    {
      "epoch": 2.0,
      "eval_loss": 0.6006895303726196,
      "eval_runtime": 3.3642,
      "eval_samples_per_second": 4.459,
      "eval_steps_per_second": 0.594,
      "step": 18
    },
    {
      "epoch": 2.235294117647059,
      "grad_norm": 0.19050121307373047,
      "learning_rate": 0.00029795419551040833,
      "loss": 0.5302,
      "step": 20
    },
    {
      "epoch": 2.8235294117647056,
      "grad_norm": 0.17813433706760406,
      "learning_rate": 0.0002954100398908995,
      "loss": 0.4439,
      "step": 25
    },
    {
      "epoch": 3.0,
      "eval_loss": 0.5779597759246826,
      "eval_runtime": 3.3656,
      "eval_samples_per_second": 4.457,
      "eval_steps_per_second": 0.594,
      "step": 27
    },
    {
      "epoch": 3.3529411764705883,
      "grad_norm": 0.23658457398414612,
      "learning_rate": 0.00029187258625509513,
      "loss": 0.4257,
      "step": 30
    },
    {
      "epoch": 3.9411764705882355,
      "grad_norm": 0.27489331364631653,
      "learning_rate": 0.00028736599899825856,
      "loss": 0.3254,
      "step": 35
    },
    {
      "epoch": 4.0,
      "eval_loss": 0.5997567772865295,
      "eval_runtime": 3.3651,
      "eval_samples_per_second": 4.458,
      "eval_steps_per_second": 0.594,
      "step": 36
    },
    {
      "epoch": 4.470588235294118,
      "grad_norm": 0.3510420620441437,
      "learning_rate": 0.00028192106268097334,
      "loss": 0.2985,
      "step": 40
    },
    {
      "epoch": 5.0,
      "grad_norm": 0.4653756320476532,
      "learning_rate": 0.00027557497173937923,
      "loss": 0.2828,
      "step": 45
    },
    {
      "epoch": 5.0,
      "eval_loss": 0.6584295034408569,
      "eval_runtime": 3.3668,
      "eval_samples_per_second": 4.455,
      "eval_steps_per_second": 0.594,
      "step": 45
    },
    {
      "epoch": 5.588235294117647,
      "grad_norm": 0.3469710052013397,
      "learning_rate": 0.00026837107640945905,
      "loss": 0.2038,
      "step": 50
    },
    {
      "epoch": 6.0,
      "eval_loss": 0.751620888710022,
      "eval_runtime": 3.3651,
      "eval_samples_per_second": 4.458,
      "eval_steps_per_second": 0.594,
      "step": 54
    },
    {
      "epoch": 6.117647058823529,
      "grad_norm": 0.3903103470802307,
      "learning_rate": 0.0002603585866009697,
      "loss": 0.1656,
      "step": 55
    },
    {
      "epoch": 6.705882352941177,
      "grad_norm": 0.38518330454826355,
      "learning_rate": 0.00025159223574386114,
      "loss": 0.1296,
      "step": 60
    },
    {
      "epoch": 7.0,
      "eval_loss": 0.8753386735916138,
      "eval_runtime": 3.3654,
      "eval_samples_per_second": 4.457,
      "eval_steps_per_second": 0.594,
      "step": 63
    },
    {
      "epoch": 7.235294117647059,
      "grad_norm": 0.24312527477741241,
      "learning_rate": 0.00024213190690345018,
      "loss": 0.085,
      "step": 65
    },
    {
      "epoch": 7.823529411764706,
      "grad_norm": 0.3267911970615387,
      "learning_rate": 0.00023204222371836405,
      "loss": 0.0751,
      "step": 70
    },
    {
      "epoch": 8.0,
      "eval_loss": 0.8750011920928955,
      "eval_runtime": 3.3678,
      "eval_samples_per_second": 4.454,
      "eval_steps_per_second": 0.594,
      "step": 72
    }
  ],
  "logging_steps": 5,
  "max_steps": 200,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 25,
  "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": 3.746749366989619e+16,
  "train_batch_size": 2,
  "trial_name": null,
  "trial_params": null
}