trainer_state.json

{
  "best_metric": 0.0,
  "best_model_checkpoint": "/workspace/previous_works/RadFM/output/RadFM-Llama3-8B-pretrain-0002-embed_tokens-depth32-lora/checkpoint-2",
  "epoch": 3.0,
  "eval_steps": 2,
  "global_step": 9,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.3333333333333333,
      "grad_norm": 62.661258697509766,
      "learning_rate": 0.0001,
      "loss": 2.7344,
      "step": 1
    },
    {
      "epoch": 0.6666666666666666,
      "grad_norm": 83.97257232666016,
      "learning_rate": 9.619397662556435e-05,
      "loss": 2.7904,
      "step": 2
    },
    {
      "epoch": 0.6666666666666666,
      "eval_accuracy": 0.0,
      "eval_loss": 2.184687852859497,
      "eval_runtime": 6.6014,
      "eval_samples_per_second": 1.515,
      "eval_steps_per_second": 0.454,
      "step": 2
    },
    {
      "epoch": 1.0,
      "grad_norm": 22.563146591186523,
      "learning_rate": 8.535533905932738e-05,
      "loss": 2.1197,
      "step": 3
    },
    {
      "epoch": 1.3333333333333333,
      "grad_norm": 8.40546703338623,
      "learning_rate": 6.91341716182545e-05,
      "loss": 1.7536,
      "step": 4
    },
    {
      "epoch": 1.3333333333333333,
      "eval_accuracy": 0.0,
      "eval_loss": 1.7347915172576904,
      "eval_runtime": 6.3856,
      "eval_samples_per_second": 1.566,
      "eval_steps_per_second": 0.47,
      "step": 4
    },
    {
      "epoch": 1.6666666666666665,
      "grad_norm": 9.354619979858398,
      "learning_rate": 5e-05,
      "loss": 1.4943,
      "step": 5
    },
    {
      "epoch": 2.0,
      "grad_norm": 6.209439754486084,
      "learning_rate": 3.086582838174551e-05,
      "loss": 1.4328,
      "step": 6
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.0,
      "eval_loss": 1.5421178340911865,
      "eval_runtime": 6.4157,
      "eval_samples_per_second": 1.559,
      "eval_steps_per_second": 0.468,
      "step": 6
    },
    {
      "epoch": 2.3333333333333335,
      "grad_norm": 5.153646469116211,
      "learning_rate": 1.4644660940672627e-05,
      "loss": 1.2755,
      "step": 7
    },
    {
      "epoch": 2.6666666666666665,
      "grad_norm": 4.289973735809326,
      "learning_rate": 3.8060233744356633e-06,
      "loss": 1.0907,
      "step": 8
    },
    {
      "epoch": 2.6666666666666665,
      "eval_accuracy": 0.0,
      "eval_loss": 1.4806464910507202,
      "eval_runtime": 6.5249,
      "eval_samples_per_second": 1.533,
      "eval_steps_per_second": 0.46,
      "step": 8
    },
    {
      "epoch": 3.0,
      "grad_norm": 9.46379566192627,
      "learning_rate": 0.0,
      "loss": 1.0722,
      "step": 9
    },
    {
      "epoch": 3.0,
      "step": 9,
      "total_flos": 1717398675652608.0,
      "train_loss": 1.7515179581112332,
      "train_runtime": 313.4275,
      "train_samples_per_second": 0.096,
      "train_steps_per_second": 0.029
    }
  ],
  "logging_steps": 1,
  "max_steps": 9,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 2,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 1717398675652608.0,
  "train_batch_size": 2,
  "trial_name": null,
  "trial_params": null
}