{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 1.971563981042654,
  "eval_steps": 128,
  "global_step": 104,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.018957345971563982,
      "grad_norm": 157.75602672892865,
      "learning_rate": 4.545454545454545e-08,
      "logits/chosen": 117.53560638427734,
      "logits/rejected": 126.8960952758789,
      "logps/chosen": -335.40118408203125,
      "logps/rejected": -439.16552734375,
      "loss": 0.8464,
      "rewards/accuracies": 0.0,
      "rewards/chosen": 0.0,
      "rewards/margins": 0.0,
      "rewards/rejected": 0.0,
      "step": 1
    },
    {
      "epoch": 0.1895734597156398,
      "grad_norm": 159.84694184422614,
      "learning_rate": 4.545454545454545e-07,
      "logits/chosen": 135.0513153076172,
      "logits/rejected": 138.41073608398438,
      "logps/chosen": -395.8233642578125,
      "logps/rejected": -439.4460144042969,
      "loss": 0.7992,
      "rewards/accuracies": 0.4375,
      "rewards/chosen": 0.008658057078719139,
      "rewards/margins": 0.014583365991711617,
      "rewards/rejected": -0.005925314500927925,
      "step": 10
    },
    {
      "epoch": 0.3791469194312796,
      "grad_norm": 152.55623162543827,
      "learning_rate": 4.885348141000122e-07,
      "logits/chosen": 121.6577377319336,
      "logits/rejected": 125.38299560546875,
      "logps/chosen": -372.36334228515625,
      "logps/rejected": -426.1581115722656,
      "loss": 0.6773,
      "rewards/accuracies": 0.59375,
      "rewards/chosen": 0.0024319454096257687,
      "rewards/margins": 0.3107669949531555,
      "rewards/rejected": -0.30833500623703003,
      "step": 20
    },
    {
      "epoch": 0.5687203791469194,
      "grad_norm": 119.76688287434446,
      "learning_rate": 4.5025027361734613e-07,
      "logits/chosen": 142.59500122070312,
      "logits/rejected": 136.1927947998047,
      "logps/chosen": -427.85626220703125,
      "logps/rejected": -474.54498291015625,
      "loss": 0.5982,
      "rewards/accuracies": 0.71875,
      "rewards/chosen": -1.7695486545562744,
      "rewards/margins": 0.9575872421264648,
      "rewards/rejected": -2.7271358966827393,
      "step": 30
    },
    {
      "epoch": 0.7582938388625592,
      "grad_norm": 101.19518327182783,
      "learning_rate": 3.893311157806091e-07,
      "logits/chosen": 125.58316802978516,
      "logits/rejected": 114.24357604980469,
      "logps/chosen": -413.34722900390625,
      "logps/rejected": -439.4224548339844,
      "loss": 0.557,
      "rewards/accuracies": 0.699999988079071,
      "rewards/chosen": -2.9576516151428223,
      "rewards/margins": 1.1197086572647095,
      "rewards/rejected": -4.077361106872559,
      "step": 40
    },
    {
      "epoch": 0.9478672985781991,
      "grad_norm": 124.4654073535278,
      "learning_rate": 3.126631330646801e-07,
      "logits/chosen": 140.7157440185547,
      "logits/rejected": 144.85781860351562,
      "logps/chosen": -476.4051208496094,
      "logps/rejected": -560.07568359375,
      "loss": 0.5143,
      "rewards/accuracies": 0.762499988079071,
      "rewards/chosen": -2.863062620162964,
      "rewards/margins": 1.3244019746780396,
      "rewards/rejected": -4.187464237213135,
      "step": 50
    },
    {
      "epoch": 1.1374407582938388,
      "grad_norm": 60.3803202767405,
      "learning_rate": 2.2891223348923882e-07,
      "logits/chosen": 133.24928283691406,
      "logits/rejected": 136.88035583496094,
      "logps/chosen": -462.84130859375,
      "logps/rejected": -546.5650024414062,
      "loss": 0.3397,
      "rewards/accuracies": 0.875,
      "rewards/chosen": -3.2417819499969482,
      "rewards/margins": 2.2681899070739746,
      "rewards/rejected": -5.509971618652344,
      "step": 60
    },
    {
      "epoch": 1.3270142180094786,
      "grad_norm": 48.59955026130948,
      "learning_rate": 1.4754491880085317e-07,
      "logits/chosen": 126.54573822021484,
      "logits/rejected": 128.51889038085938,
      "logps/chosen": -433.62042236328125,
      "logps/rejected": -533.0380859375,
      "loss": 0.2374,
      "rewards/accuracies": 0.9125000238418579,
      "rewards/chosen": -3.3238282203674316,
      "rewards/margins": 2.6523845195770264,
      "rewards/rejected": -5.976212978363037,
      "step": 70
    },
    {
      "epoch": 1.5165876777251186,
      "grad_norm": 54.174748869518474,
      "learning_rate": 7.775827023107834e-08,
      "logits/chosen": 112.75434875488281,
      "logits/rejected": 129.4608154296875,
      "logps/chosen": -436.5595703125,
      "logps/rejected": -553.3507080078125,
      "loss": 0.2095,
      "rewards/accuracies": 0.9125000238418579,
      "rewards/chosen": -4.145426273345947,
      "rewards/margins": 2.8602230548858643,
      "rewards/rejected": -7.005648612976074,
      "step": 80
    },
    {
      "epoch": 1.7061611374407581,
      "grad_norm": 54.737453014104844,
      "learning_rate": 2.7440387297912122e-08,
      "logits/chosen": 112.3515396118164,
      "logits/rejected": 125.18294525146484,
      "logps/chosen": -466.541259765625,
      "logps/rejected": -578.0580444335938,
      "loss": 0.2032,
      "rewards/accuracies": 0.918749988079071,
      "rewards/chosen": -4.005508899688721,
      "rewards/margins": 3.0133039951324463,
      "rewards/rejected": -7.018812656402588,
      "step": 90
    },
    {
      "epoch": 1.8957345971563981,
      "grad_norm": 52.477271491484125,
      "learning_rate": 2.27878296044029e-09,
      "logits/chosen": 118.53382873535156,
      "logits/rejected": 118.45182800292969,
      "logps/chosen": -454.86248779296875,
      "logps/rejected": -543.6682739257812,
      "loss": 0.1998,
      "rewards/accuracies": 0.9312499761581421,
      "rewards/chosen": -3.6443698406219482,
      "rewards/margins": 2.700843095779419,
      "rewards/rejected": -6.345212936401367,
      "step": 100
    },
    {
      "epoch": 1.971563981042654,
      "step": 104,
      "total_flos": 0.0,
      "train_loss": 0.4319001900462004,
      "train_runtime": 2177.8844,
      "train_samples_per_second": 6.199,
      "train_steps_per_second": 0.048
    }
  ],
  "logging_steps": 10,
  "max_steps": 104,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 2,
  "save_steps": 500,
  "total_flos": 0.0,
  "train_batch_size": 2,
  "trial_name": null,
  "trial_params": null
}