trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 1.0,
  "eval_steps": 500,
  "global_step": 24,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.04,
      "grad_norm": 5.126630481226491,
      "learning_rate": 1.6666666666666665e-07,
      "logits/chosen": -2.8051414489746094,
      "logits/rejected": -2.824976921081543,
      "logps/chosen": -140.9189453125,
      "logps/rejected": -111.96463775634766,
      "loss": 0.6931,
      "rewards/accuracies": 0.0,
      "rewards/chosen": 0.0,
      "rewards/margins": 0.0,
      "rewards/rejected": 0.0,
      "step": 1
    },
    {
      "epoch": 0.21,
      "grad_norm": 4.3434825310394585,
      "learning_rate": 4.888932014465352e-07,
      "logits/chosen": -2.7855749130249023,
      "logits/rejected": -2.789283275604248,
      "logps/chosen": -91.09937286376953,
      "logps/rejected": -97.13131713867188,
      "loss": 0.6931,
      "rewards/accuracies": 0.3828125,
      "rewards/chosen": 0.0011320528574287891,
      "rewards/margins": 0.0004516332410275936,
      "rewards/rejected": 0.0006804197328165174,
      "step": 5
    },
    {
      "epoch": 0.42,
      "grad_norm": 4.884271965568999,
      "learning_rate": 3.75e-07,
      "logits/chosen": -2.8252010345458984,
      "logits/rejected": -2.8388328552246094,
      "logps/chosen": -114.865234375,
      "logps/rejected": -117.52699279785156,
      "loss": 0.6917,
      "rewards/accuracies": 0.53125,
      "rewards/chosen": 0.021610241383314133,
      "rewards/margins": 0.003150947391986847,
      "rewards/rejected": 0.018459293991327286,
      "step": 10
    },
    {
      "epoch": 0.62,
      "grad_norm": 4.267689170274191,
      "learning_rate": 1.9436976651092142e-07,
      "logits/chosen": -2.7744898796081543,
      "logits/rejected": -2.7769482135772705,
      "logps/chosen": -99.48928833007812,
      "logps/rejected": -101.80488586425781,
      "loss": 0.69,
      "rewards/accuracies": 0.518750011920929,
      "rewards/chosen": 0.04334767535328865,
      "rewards/margins": 0.005900852382183075,
      "rewards/rejected": 0.03744681924581528,
      "step": 15
    },
    {
      "epoch": 0.83,
      "grad_norm": 4.8670057291088185,
      "learning_rate": 4.3440306421001324e-08,
      "logits/chosen": -2.803718328475952,
      "logits/rejected": -2.8046982288360596,
      "logps/chosen": -84.83966827392578,
      "logps/rejected": -93.15818786621094,
      "loss": 0.6868,
      "rewards/accuracies": 0.5,
      "rewards/chosen": 0.05009683966636658,
      "rewards/margins": 0.012487905099987984,
      "rewards/rejected": 0.037608928978443146,
      "step": 20
    },
    {
      "epoch": 1.0,
      "step": 24,
      "total_flos": 0.0,
      "train_loss": 0.6898899773756663,
      "train_runtime": 644.9682,
      "train_samples_per_second": 4.738,
      "train_steps_per_second": 0.037
    }
  ],
  "logging_steps": 5,
  "max_steps": 24,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 500,
  "total_flos": 0.0,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}