checkpoint-75/trainer_state.json

{
  "best_metric": 1.2520735263824463,
  "best_model_checkpoint": "/mnt/data/computer_design/lora_checkpoints/DeepSeek-R1-Distill-Llama-8B__emotion_chinese_2k__ral_8_16_0.0003_6/checkpoint-75",
  "epoch": 5.333333333333333,
  "eval_steps": 25,
  "global_step": 75,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.35555555555555557,
      "grad_norm": 9.459488868713379,
      "learning_rate": 0.00015,
      "loss": 3.4762,
      "step": 5
    },
    {
      "epoch": 0.7111111111111111,
      "grad_norm": 5.444821834564209,
      "learning_rate": 0.0003,
      "loss": 2.138,
      "step": 10
    },
    {
      "epoch": 1.0666666666666667,
      "grad_norm": 1.5013136863708496,
      "learning_rate": 0.0002797297297297297,
      "loss": 1.4366,
      "step": 15
    },
    {
      "epoch": 1.4222222222222223,
      "grad_norm": 1.1052565574645996,
      "learning_rate": 0.00025945945945945944,
      "loss": 1.3689,
      "step": 20
    },
    {
      "epoch": 1.7777777777777777,
      "grad_norm": 0.9474283456802368,
      "learning_rate": 0.00023918918918918917,
      "loss": 1.2958,
      "step": 25
    },
    {
      "epoch": 1.7777777777777777,
      "eval_loss": 1.3025931119918823,
      "eval_runtime": 8.0107,
      "eval_samples_per_second": 24.967,
      "eval_steps_per_second": 3.121,
      "step": 25
    },
    {
      "epoch": 2.1333333333333333,
      "grad_norm": 0.84736168384552,
      "learning_rate": 0.0002189189189189189,
      "loss": 1.2529,
      "step": 30
    },
    {
      "epoch": 2.488888888888889,
      "grad_norm": 0.8068817853927612,
      "learning_rate": 0.00019864864864864863,
      "loss": 1.2063,
      "step": 35
    },
    {
      "epoch": 2.8444444444444446,
      "grad_norm": 0.8153584003448486,
      "learning_rate": 0.00017837837837837839,
      "loss": 1.2033,
      "step": 40
    },
    {
      "epoch": 3.2,
      "grad_norm": 0.787850022315979,
      "learning_rate": 0.0001581081081081081,
      "loss": 1.1714,
      "step": 45
    },
    {
      "epoch": 3.5555555555555554,
      "grad_norm": 0.7746137380599976,
      "learning_rate": 0.00013783783783783782,
      "loss": 1.1831,
      "step": 50
    },
    {
      "epoch": 3.5555555555555554,
      "eval_loss": 1.2562888860702515,
      "eval_runtime": 7.999,
      "eval_samples_per_second": 25.003,
      "eval_steps_per_second": 3.125,
      "step": 50
    },
    {
      "epoch": 3.911111111111111,
      "grad_norm": 0.7717641592025757,
      "learning_rate": 0.00011756756756756755,
      "loss": 1.1497,
      "step": 55
    },
    {
      "epoch": 4.266666666666667,
      "grad_norm": 0.7889072299003601,
      "learning_rate": 9.72972972972973e-05,
      "loss": 1.1336,
      "step": 60
    },
    {
      "epoch": 4.622222222222222,
      "grad_norm": 0.8381452560424805,
      "learning_rate": 7.702702702702701e-05,
      "loss": 1.1241,
      "step": 65
    },
    {
      "epoch": 4.977777777777778,
      "grad_norm": 0.8982987999916077,
      "learning_rate": 5.6756756756756757e-05,
      "loss": 1.1431,
      "step": 70
    },
    {
      "epoch": 5.333333333333333,
      "grad_norm": 0.8278347253799438,
      "learning_rate": 3.648648648648649e-05,
      "loss": 1.1178,
      "step": 75
    },
    {
      "epoch": 5.333333333333333,
      "eval_loss": 1.2520735263824463,
      "eval_runtime": 7.9722,
      "eval_samples_per_second": 25.087,
      "eval_steps_per_second": 3.136,
      "step": 75
    }
  ],
  "logging_steps": 5,
  "max_steps": 84,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 6,
  "save_steps": 25,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 6.635612022413722e+16,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}