gemma-2-9b_pct_reverse/checkpoint-40/trainer_state.json

{
  "best_metric": 3.0729734897613525,
  "best_model_checkpoint": "/home/datta0/models/lora_final/gemma-2-9b_pct_reverse/checkpoint-8",
  "epoch": 0.10301810865191147,
  "eval_steps": 8,
  "global_step": 40,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.002575452716297787,
      "grad_norm": 10.251696586608887,
      "learning_rate": 1.25e-05,
      "loss": 2.1764,
      "step": 1
    },
    {
      "epoch": 0.010301810865191148,
      "grad_norm": 8.478413581848145,
      "learning_rate": 5e-05,
      "loss": 2.2413,
      "step": 4
    },
    {
      "epoch": 0.020603621730382295,
      "grad_norm": 7.021224021911621,
      "learning_rate": 0.0001,
      "loss": 2.3931,
      "step": 8
    },
    {
      "epoch": 0.020603621730382295,
      "eval_loss": 3.0729734897613525,
      "eval_runtime": 708.7312,
      "eval_samples_per_second": 0.346,
      "eval_steps_per_second": 0.346,
      "step": 8
    },
    {
      "epoch": 0.03090543259557344,
      "grad_norm": 8.235655784606934,
      "learning_rate": 9.997266286704631e-05,
      "loss": 6.3624,
      "step": 12
    },
    {
      "epoch": 0.04120724346076459,
      "grad_norm": 6.444003582000732,
      "learning_rate": 9.989068136093873e-05,
      "loss": 10.6943,
      "step": 16
    },
    {
      "epoch": 0.04120724346076459,
      "eval_loss": 10.896873474121094,
      "eval_runtime": 565.8534,
      "eval_samples_per_second": 0.433,
      "eval_steps_per_second": 0.433,
      "step": 16
    },
    {
      "epoch": 0.05150905432595573,
      "grad_norm": 1.8430720567703247,
      "learning_rate": 9.975414512725057e-05,
      "loss": 11.1032,
      "step": 20
    },
    {
      "epoch": 0.06181086519114688,
      "grad_norm": 1.1158581972122192,
      "learning_rate": 9.956320346634876e-05,
      "loss": 11.3663,
      "step": 24
    },
    {
      "epoch": 0.06181086519114688,
      "eval_loss": 11.264862060546875,
      "eval_runtime": 375.8938,
      "eval_samples_per_second": 0.652,
      "eval_steps_per_second": 0.652,
      "step": 24
    },
    {
      "epoch": 0.07211267605633803,
      "grad_norm": 1.7539372444152832,
      "learning_rate": 9.931806517013612e-05,
      "loss": 11.4815,
      "step": 28
    },
    {
      "epoch": 0.08241448692152918,
      "grad_norm": 2.6324522495269775,
      "learning_rate": 9.901899829374047e-05,
      "loss": 11.5978,
      "step": 32
    },
    {
      "epoch": 0.08241448692152918,
      "eval_loss": 11.56078052520752,
      "eval_runtime": 567.8708,
      "eval_samples_per_second": 0.431,
      "eval_steps_per_second": 0.431,
      "step": 32
    },
    {
      "epoch": 0.09271629778672032,
      "grad_norm": 1.0548502206802368,
      "learning_rate": 9.86663298624003e-05,
      "loss": 11.6535,
      "step": 36
    },
    {
      "epoch": 0.10301810865191147,
      "grad_norm": 1.1618255376815796,
      "learning_rate": 9.826044551386744e-05,
      "loss": 11.6505,
      "step": 40
    },
    {
      "epoch": 0.10301810865191147,
      "eval_loss": 11.69133472442627,
      "eval_runtime": 367.9621,
      "eval_samples_per_second": 0.666,
      "eval_steps_per_second": 0.666,
      "step": 40
    }
  ],
  "logging_steps": 4,
  "max_steps": 388,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 8,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 5.709775783098778e+16,
  "train_batch_size": 1,
  "trial_name": null,
  "trial_params": null
}