trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 1.0,
  "eval_steps": 500,
  "global_step": 48,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.02,
      "learning_rate": 1e-07,
      "logits/chosen": -2.901834726333618,
      "logits/rejected": -2.844017505645752,
      "logps/chosen": -210.10006713867188,
      "logps/pi_response": -112.89700317382812,
      "logps/ref_response": -112.89700317382812,
      "logps/rejected": -243.11831665039062,
      "loss": 0.6931,
      "rewards/accuracies": 0.0,
      "rewards/chosen": 0.0,
      "rewards/margins": 0.0,
      "rewards/rejected": 0.0,
      "step": 1
    },
    {
      "epoch": 0.21,
      "learning_rate": 4.83504027183137e-07,
      "logits/chosen": -2.7638163566589355,
      "logits/rejected": -2.704831838607788,
      "logps/chosen": -225.97950744628906,
      "logps/pi_response": -118.01399993896484,
      "logps/ref_response": -116.89596557617188,
      "logps/rejected": -233.20187377929688,
      "loss": 0.6869,
      "rewards/accuracies": 0.5416666865348816,
      "rewards/chosen": -0.027089955285191536,
      "rewards/margins": 0.013482754118740559,
      "rewards/rejected": -0.04057271406054497,
      "step": 10
    },
    {
      "epoch": 0.42,
      "learning_rate": 3.643105808261596e-07,
      "logits/chosen": -2.8008337020874023,
      "logits/rejected": -2.774320125579834,
      "logps/chosen": -251.0797882080078,
      "logps/pi_response": -113.0524673461914,
      "logps/ref_response": -113.1565170288086,
      "logps/rejected": -278.3044128417969,
      "loss": 0.639,
      "rewards/accuracies": 0.659375011920929,
      "rewards/chosen": -0.14624103903770447,
      "rewards/margins": 0.18217135965824127,
      "rewards/rejected": -0.32841241359710693,
      "step": 20
    },
    {
      "epoch": 0.62,
      "learning_rate": 1.8676665440207977e-07,
      "logits/chosen": -2.711427688598633,
      "logits/rejected": -2.6692211627960205,
      "logps/chosen": -271.66339111328125,
      "logps/pi_response": -147.83273315429688,
      "logps/ref_response": -125.85347747802734,
      "logps/rejected": -317.5698547363281,
      "loss": 0.5981,
      "rewards/accuracies": 0.671875,
      "rewards/chosen": -0.38658708333969116,
      "rewards/margins": 0.37116459012031555,
      "rewards/rejected": -0.7577516436576843,
      "step": 30
    },
    {
      "epoch": 0.83,
      "learning_rate": 4.1500545527530544e-08,
      "logits/chosen": -2.7132654190063477,
      "logits/rejected": -2.6600053310394287,
      "logps/chosen": -278.91387939453125,
      "logps/pi_response": -148.5419158935547,
      "logps/ref_response": -117.31595611572266,
      "logps/rejected": -345.1435852050781,
      "loss": 0.5877,
      "rewards/accuracies": 0.71875,
      "rewards/chosen": -0.4909982681274414,
      "rewards/margins": 0.4453675150871277,
      "rewards/rejected": -0.9363657236099243,
      "step": 40
    },
    {
      "epoch": 1.0,
      "step": 48,
      "total_flos": 0.0,
      "train_loss": 0.6218686103820801,
      "train_runtime": 2894.9247,
      "train_samples_per_second": 4.224,
      "train_steps_per_second": 0.017
    }
  ],
  "logging_steps": 10,
  "max_steps": 48,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 100,
  "total_flos": 0.0,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}