trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.9905956112852664,
  "eval_steps": 500,
  "global_step": 79,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.01,
      "learning_rate": 6.25e-08,
      "logits/chosen": -2.4037704467773438,
      "logits/rejected": -2.4510624408721924,
      "logps/chosen": -452.3387145996094,
      "logps/rejected": -474.736572265625,
      "loss": 0.6931,
      "rewards/accuracies": 0.0,
      "rewards/chosen": 0.0,
      "rewards/margins": 0.0,
      "rewards/rejected": 0.0,
      "step": 1
    },
    {
      "epoch": 0.13,
      "learning_rate": 4.990217055187362e-07,
      "logits/chosen": -2.3159937858581543,
      "logits/rejected": -2.326166868209839,
      "logps/chosen": -363.0790100097656,
      "logps/rejected": -444.7694091796875,
      "loss": 0.6767,
      "rewards/accuracies": 0.625,
      "rewards/chosen": -0.03609583526849747,
      "rewards/margins": 0.03676144406199455,
      "rewards/rejected": -0.07285728305578232,
      "step": 10
    },
    {
      "epoch": 0.25,
      "learning_rate": 4.655786431300069e-07,
      "logits/chosen": -2.0593743324279785,
      "logits/rejected": -1.9507596492767334,
      "logps/chosen": -487.2562561035156,
      "logps/rejected": -611.5640869140625,
      "loss": 0.5575,
      "rewards/accuracies": 0.746874988079071,
      "rewards/chosen": -1.0315128564834595,
      "rewards/margins": 0.6375976800918579,
      "rewards/rejected": -1.6691105365753174,
      "step": 20
    },
    {
      "epoch": 0.38,
      "learning_rate": 3.9061232191019517e-07,
      "logits/chosen": -1.8185665607452393,
      "logits/rejected": -1.7463910579681396,
      "logps/chosen": -481.860107421875,
      "logps/rejected": -633.5632934570312,
      "loss": 0.5326,
      "rewards/accuracies": 0.746874988079071,
      "rewards/chosen": -1.2174549102783203,
      "rewards/margins": 0.8040310144424438,
      "rewards/rejected": -2.0214858055114746,
      "step": 30
    },
    {
      "epoch": 0.5,
      "learning_rate": 2.8856223324132555e-07,
      "logits/chosen": -1.8887481689453125,
      "logits/rejected": -1.7590510845184326,
      "logps/chosen": -461.4515075683594,
      "logps/rejected": -576.8236083984375,
      "loss": 0.4994,
      "rewards/accuracies": 0.7593749761581421,
      "rewards/chosen": -0.7517538070678711,
      "rewards/margins": 0.758992075920105,
      "rewards/rejected": -1.510745882987976,
      "step": 40
    },
    {
      "epoch": 0.63,
      "learning_rate": 1.7908455541642582e-07,
      "logits/chosen": -1.7753263711929321,
      "logits/rejected": -1.5831645727157593,
      "logps/chosen": -477.20263671875,
      "logps/rejected": -597.608642578125,
      "loss": 0.4955,
      "rewards/accuracies": 0.796875,
      "rewards/chosen": -0.8421181440353394,
      "rewards/margins": 0.8058697581291199,
      "rewards/rejected": -1.647987961769104,
      "step": 50
    },
    {
      "epoch": 0.75,
      "learning_rate": 8.32661172908373e-08,
      "logits/chosen": -1.562878966331482,
      "logits/rejected": -1.4060611724853516,
      "logps/chosen": -438.7434997558594,
      "logps/rejected": -601.6849365234375,
      "loss": 0.4876,
      "rewards/accuracies": 0.800000011920929,
      "rewards/chosen": -0.8998396992683411,
      "rewards/margins": 0.9105955362319946,
      "rewards/rejected": -1.8104350566864014,
      "step": 60
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.956279997278043e-08,
      "logits/chosen": -1.5130259990692139,
      "logits/rejected": -1.2868883609771729,
      "logps/chosen": -486.8975524902344,
      "logps/rejected": -630.5179443359375,
      "loss": 0.4762,
      "rewards/accuracies": 0.7718750238418579,
      "rewards/chosen": -1.0723719596862793,
      "rewards/margins": 0.8739804029464722,
      "rewards/rejected": -1.9463523626327515,
      "step": 70
    },
    {
      "epoch": 0.99,
      "step": 79,
      "total_flos": 0.0,
      "train_loss": 0.5280408738534662,
      "train_runtime": 2579.8853,
      "train_samples_per_second": 7.899,
      "train_steps_per_second": 0.031
    }
  ],
  "logging_steps": 10,
  "max_steps": 79,
  "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
}