trainer_state.json

{
  "best_metric": 0.9519919986664445,
  "best_model_checkpoint": "/scratch/camembertv2/runs/results/flue-CLS/camembertv2-base-bf16-p2-17000/max_seq_length-1024-gradient_accumulation_steps-4-precision-fp32-learning_rate-3e-05-epochs-6-lr_scheduler-cosine-warmup_steps-0/SEED-1/checkpoint-1122",
  "epoch": 5.984,
  "eval_steps": 500,
  "global_step": 1122,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.5333333333333333,
      "grad_norm": 5.213508605957031,
      "learning_rate": 2.9415834288858954e-05,
      "loss": 0.3587,
      "step": 100
    },
    {
      "epoch": 0.9973333333333333,
      "eval_accuracy": 0.9234872478746458,
      "eval_loss": 0.21121856570243835,
      "eval_runtime": 52.6425,
      "eval_samples_per_second": 113.957,
      "eval_steps_per_second": 14.247,
      "step": 187
    },
    {
      "epoch": 1.0666666666666667,
      "grad_norm": 7.258457183837891,
      "learning_rate": 2.7708837099178863e-05,
      "loss": 0.2208,
      "step": 200
    },
    {
      "epoch": 1.6,
      "grad_norm": 19.751707077026367,
      "learning_rate": 2.5011964327923105e-05,
      "loss": 0.164,
      "step": 300
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.9384897482913819,
      "eval_loss": 0.1945669800043106,
      "eval_runtime": 52.3758,
      "eval_samples_per_second": 114.538,
      "eval_steps_per_second": 14.32,
      "step": 375
    },
    {
      "epoch": 2.1333333333333333,
      "grad_norm": 8.5791015625,
      "learning_rate": 2.1535272055128672e-05,
      "loss": 0.1581,
      "step": 400
    },
    {
      "epoch": 2.6666666666666665,
      "grad_norm": 5.534102916717529,
      "learning_rate": 1.754955553598965e-05,
      "loss": 0.0974,
      "step": 500
    },
    {
      "epoch": 2.997333333333333,
      "eval_accuracy": 0.9489914985830972,
      "eval_loss": 0.20593389868736267,
      "eval_runtime": 52.314,
      "eval_samples_per_second": 114.673,
      "eval_steps_per_second": 14.337,
      "step": 562
    },
    {
      "epoch": 3.2,
      "grad_norm": 0.21238231658935547,
      "learning_rate": 1.3365257293813959e-05,
      "loss": 0.0864,
      "step": 600
    },
    {
      "epoch": 3.7333333333333334,
      "grad_norm": 0.7444894313812256,
      "learning_rate": 9.308287136370512e-06,
      "loss": 0.0728,
      "step": 700
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.9511585264210701,
      "eval_loss": 0.19625167548656464,
      "eval_runtime": 52.4884,
      "eval_samples_per_second": 114.292,
      "eval_steps_per_second": 14.289,
      "step": 750
    },
    {
      "epoch": 4.266666666666667,
      "grad_norm": 11.124679565429688,
      "learning_rate": 5.694637444606099e-06,
      "loss": 0.0612,
      "step": 800
    },
    {
      "epoch": 4.8,
      "grad_norm": 6.557811260223389,
      "learning_rate": 2.805770917454614e-06,
      "loss": 0.0353,
      "step": 900
    },
    {
      "epoch": 4.997333333333334,
      "eval_accuracy": 0.9516586097682947,
      "eval_loss": 0.2168048769235611,
      "eval_runtime": 52.2955,
      "eval_samples_per_second": 114.713,
      "eval_steps_per_second": 14.342,
      "step": 937
    },
    {
      "epoch": 5.333333333333333,
      "grad_norm": 2.557724952697754,
      "learning_rate": 8.66697790812731e-07,
      "loss": 0.0445,
      "step": 1000
    },
    {
      "epoch": 5.866666666666667,
      "grad_norm": 0.06581971794366837,
      "learning_rate": 2.8450068944338436e-08,
      "loss": 0.035,
      "step": 1100
    },
    {
      "epoch": 5.984,
      "eval_accuracy": 0.9519919986664445,
      "eval_loss": 0.2167392075061798,
      "eval_runtime": 52.341,
      "eval_samples_per_second": 114.614,
      "eval_steps_per_second": 14.329,
      "step": 1122
    },
    {
      "epoch": 5.984,
      "step": 1122,
      "total_flos": 6620464611065820.0,
      "train_loss": 0.11981426341435913,
      "train_runtime": 1269.2954,
      "train_samples_per_second": 28.348,
      "train_steps_per_second": 0.884
    }
  ],
  "logging_steps": 100,
  "max_steps": 1122,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 6,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 6620464611065820.0,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}