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": 28.310565581646102,
      "learning_rate": 8.333333333333333e-08,
      "logits/chosen": -2.282123565673828,
      "logits/rejected": -2.180680274963379,
      "logps/chosen": -366.17987060546875,
      "logps/rejected": -389.84954833984375,
      "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": 26.63593514238783,
      "learning_rate": 4.930057285201027e-07,
      "logits/chosen": -2.1524295806884766,
      "logits/rejected": -2.1412665843963623,
      "logps/chosen": -333.9299011230469,
      "logps/rejected": -436.9288330078125,
      "loss": 0.6639,
      "rewards/accuracies": 0.6493055820465088,
      "rewards/chosen": -0.07371152937412262,
      "rewards/margins": 0.06567458808422089,
      "rewards/rejected": -0.1393861174583435,
      "step": 10
    },
    {
      "epoch": 0.33472803347280333,
      "grad_norm": 45.04577240979691,
      "learning_rate": 4.187457503795526e-07,
      "logits/chosen": -1.623558759689331,
      "logits/rejected": -1.549423336982727,
      "logps/chosen": -467.49139404296875,
      "logps/rejected": -620.3740844726562,
      "loss": 0.5604,
      "rewards/accuracies": 0.731249988079071,
      "rewards/chosen": -1.3230326175689697,
      "rewards/margins": 0.725892961025238,
      "rewards/rejected": -2.0489256381988525,
      "step": 20
    },
    {
      "epoch": 0.502092050209205,
      "grad_norm": 25.13032401537719,
      "learning_rate": 2.8691164100062034e-07,
      "logits/chosen": -1.4312714338302612,
      "logits/rejected": -1.241246223449707,
      "logps/chosen": -468.46240234375,
      "logps/rejected": -618.2733764648438,
      "loss": 0.5045,
      "rewards/accuracies": 0.737500011920929,
      "rewards/chosen": -1.284248948097229,
      "rewards/margins": 0.9667768478393555,
      "rewards/rejected": -2.251025676727295,
      "step": 30
    },
    {
      "epoch": 0.6694560669456067,
      "grad_norm": 28.72537330625965,
      "learning_rate": 1.4248369943086995e-07,
      "logits/chosen": -1.2682111263275146,
      "logits/rejected": -1.0946087837219238,
      "logps/chosen": -389.8838806152344,
      "logps/rejected": -542.1454467773438,
      "loss": 0.4978,
      "rewards/accuracies": 0.7875000238418579,
      "rewards/chosen": -0.8395187258720398,
      "rewards/margins": 0.7310727834701538,
      "rewards/rejected": -1.570591688156128,
      "step": 40
    },
    {
      "epoch": 0.8368200836820083,
      "grad_norm": 32.05290862112005,
      "learning_rate": 3.473909705816111e-08,
      "logits/chosen": -1.0692652463912964,
      "logits/rejected": -0.9052573442459106,
      "logps/chosen": -400.3882141113281,
      "logps/rejected": -565.8882446289062,
      "loss": 0.4863,
      "rewards/accuracies": 0.746874988079071,
      "rewards/chosen": -0.8596526980400085,
      "rewards/margins": 0.7219886183738708,
      "rewards/rejected": -1.581641435623169,
      "step": 50
    },
    {
      "epoch": 0.9874476987447699,
      "step": 59,
      "total_flos": 0.0,
      "train_loss": 0.5407808513964637,
      "train_runtime": 1626.1325,
      "train_samples_per_second": 9.398,
      "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
}