trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.9874476987447699,
  "eval_steps": 500,
  "global_step": 59,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.016736401673640166,
      "grad_norm": 7.7724597597287675,
      "learning_rate": 8.333333333333333e-08,
      "logits/chosen": -2.807276487350464,
      "logits/rejected": -2.7759768962860107,
      "logps/chosen": -315.42626953125,
      "logps/rejected": -227.5915985107422,
      "loss": 0.6931,
      "rewards/accuracies": 0.0,
      "rewards/chosen": 0.0,
      "rewards/margins": 0.0,
      "rewards/rejected": 0.0,
      "step": 1
    },
    {
      "epoch": 0.16736401673640167,
      "grad_norm": 7.43868551766701,
      "learning_rate": 4.930057285201027e-07,
      "logits/chosen": -2.75545072555542,
      "logits/rejected": -2.7460691928863525,
      "logps/chosen": -271.9334716796875,
      "logps/rejected": -260.69281005859375,
      "loss": 0.6914,
      "rewards/accuracies": 0.5208333134651184,
      "rewards/chosen": 0.005364930257201195,
      "rewards/margins": 0.0036840070970356464,
      "rewards/rejected": 0.001680923392996192,
      "step": 10
    },
    {
      "epoch": 0.33472803347280333,
      "grad_norm": 6.731470928419319,
      "learning_rate": 4.187457503795526e-07,
      "logits/chosen": -2.787079334259033,
      "logits/rejected": -2.768202543258667,
      "logps/chosen": -261.4521789550781,
      "logps/rejected": -250.88571166992188,
      "loss": 0.6745,
      "rewards/accuracies": 0.7093750238418579,
      "rewards/chosen": 0.03229519724845886,
      "rewards/margins": 0.04072761535644531,
      "rewards/rejected": -0.008432422764599323,
      "step": 20
    },
    {
      "epoch": 0.502092050209205,
      "grad_norm": 7.353543442886458,
      "learning_rate": 2.8691164100062034e-07,
      "logits/chosen": -2.7879936695098877,
      "logits/rejected": -2.7689452171325684,
      "logps/chosen": -293.8031921386719,
      "logps/rejected": -254.1697235107422,
      "loss": 0.6488,
      "rewards/accuracies": 0.6781250238418579,
      "rewards/chosen": -0.02347097545862198,
      "rewards/margins": 0.12664587795734406,
      "rewards/rejected": -0.15011683106422424,
      "step": 30
    },
    {
      "epoch": 0.6694560669456067,
      "grad_norm": 8.125000777751103,
      "learning_rate": 1.4248369943086995e-07,
      "logits/chosen": -2.7606043815612793,
      "logits/rejected": -2.7380428314208984,
      "logps/chosen": -264.1278381347656,
      "logps/rejected": -255.55029296875,
      "loss": 0.6353,
      "rewards/accuracies": 0.6812499761581421,
      "rewards/chosen": -0.11208363622426987,
      "rewards/margins": 0.1603127121925354,
      "rewards/rejected": -0.27239635586738586,
      "step": 40
    },
    {
      "epoch": 0.8368200836820083,
      "grad_norm": 8.574824255298934,
      "learning_rate": 3.473909705816111e-08,
      "logits/chosen": -2.770317554473877,
      "logits/rejected": -2.7504336833953857,
      "logps/chosen": -279.09698486328125,
      "logps/rejected": -287.2370300292969,
      "loss": 0.6202,
      "rewards/accuracies": 0.628125011920929,
      "rewards/chosen": -0.22088325023651123,
      "rewards/margins": 0.13636226952075958,
      "rewards/rejected": -0.3572455048561096,
      "step": 50
    },
    {
      "epoch": 0.9874476987447699,
      "step": 59,
      "total_flos": 0.0,
      "train_loss": 0.6490422426644018,
      "train_runtime": 1622.868,
      "train_samples_per_second": 9.417,
      "train_steps_per_second": 0.036
    }
  ],
  "logging_steps": 10,
  "max_steps": 59,
  "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
}