{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 3.0,
  "eval_steps": 500,
  "global_step": 1500,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.2,
      "grad_norm": 1440.0,
      "learning_rate": 2.9818992729986362e-05,
      "logits/chosen": -0.7303402423858643,
      "logits/rejected": -0.45186057686805725,
      "logps/chosen": -394.23968505859375,
      "logps/rejected": -349.1568298339844,
      "loss": 3.1086,
      "rewards/accuracies": 0.48500001430511475,
      "rewards/chosen": -4.4438581466674805,
      "rewards/margins": -0.3193548321723938,
      "rewards/rejected": -4.124503135681152,
      "step": 100
    },
    {
      "epoch": 0.4,
      "grad_norm": 556.0,
      "learning_rate": 2.9273111183047722e-05,
      "logits/chosen": -0.5820356011390686,
      "logits/rejected": -0.45071941614151,
      "logps/chosen": -419.5753479003906,
      "logps/rejected": -373.3916320800781,
      "loss": 3.3459,
      "rewards/accuracies": 0.5299999713897705,
      "rewards/chosen": -6.157695770263672,
      "rewards/margins": 0.3280045986175537,
      "rewards/rejected": -6.4857001304626465,
      "step": 200
    },
    {
      "epoch": 0.6,
      "grad_norm": 1688.0,
      "learning_rate": 2.837577826904226e-05,
      "logits/chosen": -0.8348171710968018,
      "logits/rejected": -0.6238606572151184,
      "logps/chosen": -443.1069030761719,
      "logps/rejected": -385.998779296875,
      "loss": 3.7071,
      "rewards/accuracies": 0.5849999785423279,
      "rewards/chosen": -7.805955410003662,
      "rewards/margins": 0.8645437359809875,
      "rewards/rejected": -8.67049789428711,
      "step": 300
    },
    {
      "epoch": 0.8,
      "grad_norm": 7.75,
      "learning_rate": 2.714908930238998e-05,
      "logits/chosen": -0.915019690990448,
      "logits/rejected": -0.7366578578948975,
      "logps/chosen": -466.5214538574219,
      "logps/rejected": -348.214111328125,
      "loss": 4.0855,
      "rewards/accuracies": 0.4699999988079071,
      "rewards/chosen": -7.924253463745117,
      "rewards/margins": -0.6686192154884338,
      "rewards/rejected": -7.255634784698486,
      "step": 400
    },
    {
      "epoch": 1.0,
      "grad_norm": 1392.0,
      "learning_rate": 2.5623249436597142e-05,
      "logits/chosen": -1.1756317615509033,
      "logits/rejected": -1.0033434629440308,
      "logps/chosen": -446.835693359375,
      "logps/rejected": -426.72015380859375,
      "loss": 2.9222,
      "rewards/accuracies": 0.5,
      "rewards/chosen": -9.632830619812012,
      "rewards/margins": 0.8283342123031616,
      "rewards/rejected": -10.461165428161621,
      "step": 500
    },
    {
      "epoch": 1.2,
      "grad_norm": 12.0,
      "learning_rate": 2.3835829913131745e-05,
      "logits/chosen": -1.8936198949813843,
      "logits/rejected": -1.6783454418182373,
      "logps/chosen": -393.1761474609375,
      "logps/rejected": -409.0048522949219,
      "loss": 0.2294,
      "rewards/accuracies": 0.9599999785423279,
      "rewards/chosen": 5.744272232055664,
      "rewards/margins": 40.307220458984375,
      "rewards/rejected": -34.562950134277344,
      "step": 600
    },
    {
      "epoch": 1.4,
      "grad_norm": 0.0,
      "learning_rate": 2.1830842932766803e-05,
      "logits/chosen": -2.0007760524749756,
      "logits/rejected": -1.7464280128479004,
      "logps/chosen": -385.68499755859375,
      "logps/rejected": -453.5122985839844,
      "loss": 0.3262,
      "rewards/accuracies": 0.9549999833106995,
      "rewards/chosen": 8.030089378356934,
      "rewards/margins": 44.05226516723633,
      "rewards/rejected": -36.022178649902344,
      "step": 700
    },
    {
      "epoch": 1.6,
      "grad_norm": 9.12696123123169e-08,
      "learning_rate": 1.9657657929129185e-05,
      "logits/chosen": -1.8635623455047607,
      "logits/rejected": -1.603875994682312,
      "logps/chosen": -411.04840087890625,
      "logps/rejected": -464.8753967285156,
      "loss": 0.1132,
      "rewards/accuracies": 0.9700000286102295,
      "rewards/chosen": 11.565813064575195,
      "rewards/margins": 47.49372100830078,
      "rewards/rejected": -35.92790985107422,
      "step": 800
    },
    {
      "epoch": 1.8,
      "grad_norm": 0.0,
      "learning_rate": 1.736978592939598e-05,
      "logits/chosen": -1.9231587648391724,
      "logits/rejected": -1.6588664054870605,
      "logps/chosen": -414.5479736328125,
      "logps/rejected": -444.71441650390625,
      "loss": 0.0873,
      "rewards/accuracies": 0.9750000238418579,
      "rewards/chosen": 10.788975715637207,
      "rewards/margins": 44.054561614990234,
      "rewards/rejected": -33.26558303833008,
      "step": 900
    },
    {
      "epoch": 2.0,
      "grad_norm": 0.091796875,
      "learning_rate": 1.5023561935212463e-05,
      "logits/chosen": -1.952134132385254,
      "logits/rejected": -1.622839093208313,
      "logps/chosen": -362.3935852050781,
      "logps/rejected": -455.61383056640625,
      "loss": 0.3025,
      "rewards/accuracies": 0.9750000238418579,
      "rewards/chosen": 9.277441024780273,
      "rewards/margins": 44.19200134277344,
      "rewards/rejected": -34.91455078125,
      "step": 1000
    },
    {
      "epoch": 2.2,
      "grad_norm": 6.288290023803711e-06,
      "learning_rate": 1.2676757767986436e-05,
      "logits/chosen": -1.8726714849472046,
      "logits/rejected": -1.6393769979476929,
      "logps/chosen": -369.1127624511719,
      "logps/rejected": -453.4897766113281,
      "loss": 0.0,
      "rewards/accuracies": 1.0,
      "rewards/chosen": 9.544073104858398,
      "rewards/margins": 45.42412185668945,
      "rewards/rejected": -35.88005065917969,
      "step": 1100
    },
    {
      "epoch": 2.4,
      "grad_norm": 4.26173210144043e-06,
      "learning_rate": 1.038715953491149e-05,
      "logits/chosen": -1.909058690071106,
      "logits/rejected": -1.5853359699249268,
      "logps/chosen": -406.30902099609375,
      "logps/rejected": -417.96429443359375,
      "loss": 0.0,
      "rewards/accuracies": 1.0,
      "rewards/chosen": 12.841102600097656,
      "rewards/margins": 48.4782829284668,
      "rewards/rejected": -35.63717269897461,
      "step": 1200
    },
    {
      "epoch": 2.6,
      "grad_norm": 0.09912109375,
      "learning_rate": 8.211144743226819e-06,
      "logits/chosen": -1.847670555114746,
      "logits/rejected": -1.6100419759750366,
      "logps/chosen": -385.2706298828125,
      "logps/rejected": -452.8170166015625,
      "loss": 0.0,
      "rewards/accuracies": 1.0,
      "rewards/chosen": 12.797687530517578,
      "rewards/margins": 51.022464752197266,
      "rewards/rejected": -38.22477722167969,
      "step": 1300
    },
    {
      "epoch": 2.8,
      "grad_norm": 6.693881005048752e-10,
      "learning_rate": 6.202294098882747e-06,
      "logits/chosen": -1.887666940689087,
      "logits/rejected": -1.6139150857925415,
      "logps/chosen": -366.9139709472656,
      "logps/rejected": -446.2010803222656,
      "loss": 0.0002,
      "rewards/accuracies": 1.0,
      "rewards/chosen": 10.917724609375,
      "rewards/margins": 48.82371139526367,
      "rewards/rejected": -37.90598678588867,
      "step": 1400
    },
    {
      "epoch": 3.0,
      "grad_norm": 1.0695657692849636e-09,
      "learning_rate": 4.410072171736081e-06,
      "logits/chosen": -1.8766179084777832,
      "logits/rejected": -1.5880229473114014,
      "logps/chosen": -401.2027587890625,
      "logps/rejected": -485.7976989746094,
      "loss": 0.0001,
      "rewards/accuracies": 1.0,
      "rewards/chosen": 14.522769927978516,
      "rewards/margins": 52.987579345703125,
      "rewards/rejected": -38.46479797363281,
      "step": 1500
    }
  ],
  "logging_steps": 100,
  "max_steps": 2000,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 4,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 0.0,
  "train_batch_size": 1,
  "trial_name": null,
  "trial_params": null
}