trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 1.0,
  "eval_steps": 500,
  "global_step": 274,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.18248175182481752,
      "grad_norm": 0.11348239332437515,
      "learning_rate": 9.803954791481238e-07,
      "logits/chosen": 0.7525291442871094,
      "logits/rejected": 0.862596869468689,
      "logps/chosen": -57.87046432495117,
      "logps/rejected": -91.25581359863281,
      "loss": 0.042,
      "rewards/accuracies": 1.0,
      "rewards/chosen": 3.735088586807251,
      "rewards/margins": 5.059007167816162,
      "rewards/rejected": -1.3239188194274902,
      "step": 50
    },
    {
      "epoch": 0.36496350364963503,
      "grad_norm": 2.2142693996429443,
      "learning_rate": 8.03112705483319e-07,
      "logits/chosen": 0.7221750020980835,
      "logits/rejected": 0.8843123912811279,
      "logps/chosen": -54.99198913574219,
      "logps/rejected": -95.34280395507812,
      "loss": 0.0344,
      "rewards/accuracies": 1.0,
      "rewards/chosen": 3.793600559234619,
      "rewards/margins": 5.281071186065674,
      "rewards/rejected": -1.4874707460403442,
      "step": 100
    },
    {
      "epoch": 0.5474452554744526,
      "grad_norm": 1.848400354385376,
      "learning_rate": 5.063851773579869e-07,
      "logits/chosen": 0.7279171943664551,
      "logits/rejected": 0.8646959662437439,
      "logps/chosen": -55.23713684082031,
      "logps/rejected": -100.29174041748047,
      "loss": 0.031,
      "rewards/accuracies": 1.0,
      "rewards/chosen": 3.7684295177459717,
      "rewards/margins": 5.4548797607421875,
      "rewards/rejected": -1.6864502429962158,
      "step": 150
    },
    {
      "epoch": 0.7299270072992701,
      "grad_norm": 0.9244675040245056,
      "learning_rate": 2.071415028359026e-07,
      "logits/chosen": 0.7067604064941406,
      "logits/rejected": 0.8388283252716064,
      "logps/chosen": -50.11094284057617,
      "logps/rejected": -98.2806625366211,
      "loss": 0.0274,
      "rewards/accuracies": 1.0,
      "rewards/chosen": 4.066249847412109,
      "rewards/margins": 5.748508930206299,
      "rewards/rejected": -1.6822593212127686,
      "step": 200
    },
    {
      "epoch": 0.9124087591240876,
      "grad_norm": 0.7083641886711121,
      "learning_rate": 2.3301803972534728e-08,
      "logits/chosen": 0.7070822715759277,
      "logits/rejected": 0.857603132724762,
      "logps/chosen": -51.985496520996094,
      "logps/rejected": -93.11666870117188,
      "loss": 0.0305,
      "rewards/accuracies": 1.0,
      "rewards/chosen": 3.9429595470428467,
      "rewards/margins": 5.623107433319092,
      "rewards/rejected": -1.6801483631134033,
      "step": 250
    }
  ],
  "logging_steps": 50,
  "max_steps": 274,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "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": 0.0,
  "train_batch_size": 2,
  "trial_name": null,
  "trial_params": null
}