{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 1.9905213270142181,
  "eval_steps": 100,
  "global_step": 210,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.01,
      "grad_norm": 177.85299255533252,
      "learning_rate": 5e-09,
      "logits/chosen": 129.0,
      "logits/rejected": 125.5,
      "logps/chosen": -428.0,
      "logps/rejected": -470.0,
      "loss": 0.6914,
      "rewards/accuracies": 0.0,
      "rewards/chosen": 0.0,
      "rewards/margins": 0.0,
      "rewards/rejected": 0.0,
      "step": 1
    },
    {
      "epoch": 0.09,
      "grad_norm": 171.69218216908772,
      "learning_rate": 5e-08,
      "logits/chosen": 125.5,
      "logits/rejected": 133.0,
      "logps/chosen": -414.0,
      "logps/rejected": -448.0,
      "loss": 0.7199,
      "rewards/accuracies": 0.3888888955116272,
      "rewards/chosen": -0.017822265625,
      "rewards/margins": -0.0498046875,
      "rewards/rejected": 0.031982421875,
      "step": 10
    },
    {
      "epoch": 0.19,
      "grad_norm": 180.59129253006714,
      "learning_rate": 1e-07,
      "logits/chosen": 136.0,
      "logits/rejected": 136.0,
      "logps/chosen": -400.0,
      "logps/rejected": -430.0,
      "loss": 0.7187,
      "rewards/accuracies": 0.4437499940395355,
      "rewards/chosen": -0.0242919921875,
      "rewards/margins": -0.02197265625,
      "rewards/rejected": -0.0023956298828125,
      "step": 20
    },
    {
      "epoch": 0.28,
      "grad_norm": 171.658251860157,
      "learning_rate": 1.5e-07,
      "logits/chosen": 130.0,
      "logits/rejected": 130.0,
      "logps/chosen": -374.0,
      "logps/rejected": -388.0,
      "loss": 0.7132,
      "rewards/accuracies": 0.4625000059604645,
      "rewards/chosen": 0.09521484375,
      "rewards/margins": 0.004791259765625,
      "rewards/rejected": 0.09033203125,
      "step": 30
    },
    {
      "epoch": 0.38,
      "grad_norm": 170.37244697891163,
      "learning_rate": 2e-07,
      "logits/chosen": 114.5,
      "logits/rejected": 128.0,
      "logps/chosen": -336.0,
      "logps/rejected": -428.0,
      "loss": 0.659,
      "rewards/accuracies": 0.625,
      "rewards/chosen": 0.30859375,
      "rewards/margins": 0.23828125,
      "rewards/rejected": 0.0712890625,
      "step": 40
    },
    {
      "epoch": 0.47,
      "grad_norm": 156.01063486797565,
      "learning_rate": 2.5e-07,
      "logits/chosen": 124.5,
      "logits/rejected": 124.0,
      "logps/chosen": -372.0,
      "logps/rejected": -388.0,
      "loss": 0.6405,
      "rewards/accuracies": 0.5874999761581421,
      "rewards/chosen": 0.173828125,
      "rewards/margins": 0.318359375,
      "rewards/rejected": -0.14453125,
      "step": 50
    },
    {
      "epoch": 0.57,
      "grad_norm": 191.09116619653446,
      "learning_rate": 3e-07,
      "logits/chosen": 116.0,
      "logits/rejected": 124.0,
      "logps/chosen": -342.0,
      "logps/rejected": -416.0,
      "loss": 0.5833,
      "rewards/accuracies": 0.731249988079071,
      "rewards/chosen": -0.0191650390625,
      "rewards/margins": 0.67578125,
      "rewards/rejected": -0.69140625,
      "step": 60
    },
    {
      "epoch": 0.66,
      "grad_norm": 155.04958273291422,
      "learning_rate": 3.5e-07,
      "logits/chosen": 122.0,
      "logits/rejected": 117.5,
      "logps/chosen": -386.0,
      "logps/rejected": -394.0,
      "loss": 0.5745,
      "rewards/accuracies": 0.7124999761581421,
      "rewards/chosen": -0.09375,
      "rewards/margins": 0.79296875,
      "rewards/rejected": -0.88671875,
      "step": 70
    },
    {
      "epoch": 0.76,
      "grad_norm": 142.1796149100647,
      "learning_rate": 4e-07,
      "logits/chosen": 116.5,
      "logits/rejected": 116.5,
      "logps/chosen": -338.0,
      "logps/rejected": -406.0,
      "loss": 0.537,
      "rewards/accuracies": 0.7562500238418579,
      "rewards/chosen": -0.33203125,
      "rewards/margins": 0.96484375,
      "rewards/rejected": -1.296875,
      "step": 80
    },
    {
      "epoch": 0.85,
      "grad_norm": 199.30217047660693,
      "learning_rate": 4.5e-07,
      "logits/chosen": 130.0,
      "logits/rejected": 132.0,
      "logps/chosen": -412.0,
      "logps/rejected": -444.0,
      "loss": 0.4894,
      "rewards/accuracies": 0.7437499761581421,
      "rewards/chosen": -0.8515625,
      "rewards/margins": 1.0078125,
      "rewards/rejected": -1.859375,
      "step": 90
    },
    {
      "epoch": 0.95,
      "grad_norm": 130.24358680507126,
      "learning_rate": 5e-07,
      "logits/chosen": 117.5,
      "logits/rejected": 126.5,
      "logps/chosen": -392.0,
      "logps/rejected": -490.0,
      "loss": 0.5324,
      "rewards/accuracies": 0.7250000238418579,
      "rewards/chosen": -0.98828125,
      "rewards/margins": 1.2578125,
      "rewards/rejected": -2.25,
      "step": 100
    },
    {
      "epoch": 0.95,
      "eval_logits/chosen": 92.0,
      "eval_logits/rejected": 94.0,
      "eval_logps/chosen": -376.0,
      "eval_logps/rejected": -456.0,
      "eval_loss": 0.4851927161216736,
      "eval_rewards/accuracies": 0.7340425252914429,
      "eval_rewards/chosen": -1.1640625,
      "eval_rewards/margins": 1.3046875,
      "eval_rewards/rejected": -2.46875,
      "eval_runtime": 130.6909,
      "eval_samples_per_second": 5.739,
      "eval_steps_per_second": 0.36,
      "step": 100
    },
    {
      "epoch": 1.04,
      "grad_norm": 73.74185148689295,
      "learning_rate": 4.898732434036243e-07,
      "logits/chosen": 119.5,
      "logits/rejected": 114.5,
      "logps/chosen": -406.0,
      "logps/rejected": -458.0,
      "loss": 0.3895,
      "rewards/accuracies": 0.8187500238418579,
      "rewards/chosen": -1.109375,
      "rewards/margins": 1.6953125,
      "rewards/rejected": -2.8125,
      "step": 110
    },
    {
      "epoch": 1.14,
      "grad_norm": 77.9102123166481,
      "learning_rate": 4.603133832077953e-07,
      "logits/chosen": 113.5,
      "logits/rejected": 122.0,
      "logps/chosen": -352.0,
      "logps/rejected": -470.0,
      "loss": 0.2308,
      "rewards/accuracies": 0.918749988079071,
      "rewards/chosen": -0.04150390625,
      "rewards/margins": 2.859375,
      "rewards/rejected": -2.90625,
      "step": 120
    },
    {
      "epoch": 1.23,
      "grad_norm": 72.1155658512914,
      "learning_rate": 4.137151834863213e-07,
      "logits/chosen": 107.0,
      "logits/rejected": 108.5,
      "logps/chosen": -392.0,
      "logps/rejected": -462.0,
      "loss": 0.2175,
      "rewards/accuracies": 0.949999988079071,
      "rewards/chosen": -0.68359375,
      "rewards/margins": 2.796875,
      "rewards/rejected": -3.46875,
      "step": 130
    },
    {
      "epoch": 1.33,
      "grad_norm": 83.89189869364462,
      "learning_rate": 3.5385375325047163e-07,
      "logits/chosen": 116.0,
      "logits/rejected": 123.0,
      "logps/chosen": -396.0,
      "logps/rejected": -510.0,
      "loss": 0.2066,
      "rewards/accuracies": 0.9312499761581421,
      "rewards/chosen": -1.6875,
      "rewards/margins": 2.796875,
      "rewards/rejected": -4.46875,
      "step": 140
    },
    {
      "epoch": 1.42,
      "grad_norm": 67.05775374408921,
      "learning_rate": 2.8557870956832133e-07,
      "logits/chosen": 104.5,
      "logits/rejected": 107.5,
      "logps/chosen": -390.0,
      "logps/rejected": -490.0,
      "loss": 0.1569,
      "rewards/accuracies": 0.9312499761581421,
      "rewards/chosen": -1.484375,
      "rewards/margins": 3.25,
      "rewards/rejected": -4.71875,
      "step": 150
    },
    {
      "epoch": 1.52,
      "grad_norm": 91.02831609820949,
      "learning_rate": 2.1442129043167873e-07,
      "logits/chosen": 101.5,
      "logits/rejected": 121.0,
      "logps/chosen": -394.0,
      "logps/rejected": -506.0,
      "loss": 0.1956,
      "rewards/accuracies": 0.918749988079071,
      "rewards/chosen": -1.28125,
      "rewards/margins": 3.671875,
      "rewards/rejected": -4.9375,
      "step": 160
    },
    {
      "epoch": 1.61,
      "grad_norm": 78.82758233414538,
      "learning_rate": 1.461462467495284e-07,
      "logits/chosen": 109.5,
      "logits/rejected": 110.5,
      "logps/chosen": -406.0,
      "logps/rejected": -476.0,
      "loss": 0.1575,
      "rewards/accuracies": 0.956250011920929,
      "rewards/chosen": -0.9375,
      "rewards/margins": 3.484375,
      "rewards/rejected": -4.4375,
      "step": 170
    },
    {
      "epoch": 1.71,
      "grad_norm": 95.52959889189601,
      "learning_rate": 8.628481651367875e-08,
      "logits/chosen": 105.0,
      "logits/rejected": 117.0,
      "logps/chosen": -406.0,
      "logps/rejected": -536.0,
      "loss": 0.1574,
      "rewards/accuracies": 0.9624999761581421,
      "rewards/chosen": -0.83203125,
      "rewards/margins": 3.3125,
      "rewards/rejected": -4.15625,
      "step": 180
    },
    {
      "epoch": 1.8,
      "grad_norm": 67.11817234106277,
      "learning_rate": 3.968661679220467e-08,
      "logits/chosen": 114.0,
      "logits/rejected": 115.0,
      "logps/chosen": -414.0,
      "logps/rejected": -512.0,
      "loss": 0.1866,
      "rewards/accuracies": 0.918749988079071,
      "rewards/chosen": -0.8125,
      "rewards/margins": 3.5625,
      "rewards/rejected": -4.375,
      "step": 190
    },
    {
      "epoch": 1.9,
      "grad_norm": 83.00810869226768,
      "learning_rate": 1.0126756596375685e-08,
      "logits/chosen": 96.5,
      "logits/rejected": 104.0,
      "logps/chosen": -384.0,
      "logps/rejected": -464.0,
      "loss": 0.1822,
      "rewards/accuracies": 0.887499988079071,
      "rewards/chosen": -1.265625,
      "rewards/margins": 3.125,
      "rewards/rejected": -4.375,
      "step": 200
    },
    {
      "epoch": 1.9,
      "eval_logits/chosen": 82.0,
      "eval_logits/rejected": 84.0,
      "eval_logps/chosen": -388.0,
      "eval_logps/rejected": -480.0,
      "eval_loss": 0.4446233808994293,
      "eval_rewards/accuracies": 0.7446808218955994,
      "eval_rewards/chosen": -1.796875,
      "eval_rewards/margins": 1.8359375,
      "eval_rewards/rejected": -3.640625,
      "eval_runtime": 133.5481,
      "eval_samples_per_second": 5.616,
      "eval_steps_per_second": 0.352,
      "step": 200
    },
    {
      "epoch": 1.99,
      "grad_norm": 47.68700617963266,
      "learning_rate": 0.0,
      "logits/chosen": 96.5,
      "logits/rejected": 98.5,
      "logps/chosen": -390.0,
      "logps/rejected": -446.0,
      "loss": 0.2176,
      "rewards/accuracies": 0.918749988079071,
      "rewards/chosen": -1.140625,
      "rewards/margins": 3.125,
      "rewards/rejected": -4.28125,
      "step": 210
    },
    {
      "epoch": 1.99,
      "step": 210,
      "total_flos": 0.0,
      "train_loss": 0.4030106709116981,
      "train_runtime": 4928.2992,
      "train_samples_per_second": 2.739,
      "train_steps_per_second": 0.043
    }
  ],
  "logging_steps": 10,
  "max_steps": 210,
  "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
}