{
  "best_metric": 0.47333332896232605,
  "best_model_checkpoint": "./zephyr/08-04-24-Weni-WeniGPT-Agents-Zephyr-1.0.17-KTO_Hyperparameter search, altering lora params for KTO task.-2_max_steps-145_batch_16_2024-04-08_ppid_9/checkpoint-100",
  "epoch": 0.684931506849315,
  "eval_steps": 50,
  "global_step": 100,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.14,
      "grad_norm": 0.0,
      "learning_rate": 0.00018,
      "loss": 0.4638,
      "step": 20,
      "train/kl": 7.083856105804443,
      "train/logps/chosen": -426.40190360915494,
      "train/logps/rejected": -550.5845768960675,
      "train/rewards/chosen": -14.309082890900088,
      "train/rewards/margins": 11.978627234626599,
      "train/rewards/rejected": -26.287710125526687
    },
    {
      "epoch": 0.27,
      "grad_norm": 0.0,
      "learning_rate": 0.00015142857142857143,
      "loss": 0.4844,
      "step": 40,
      "train/kl": 0.0,
      "train/logps/chosen": -3204.072177419355,
      "train/logps/rejected": -3086.9876893939395,
      "train/rewards/chosen": -292.8815524193548,
      "train/rewards/margins": -13.51173707844572,
      "train/rewards/rejected": -279.3698153409091
    },
    {
      "epoch": 0.34,
      "eval/kl": 0.0,
      "eval/logps/chosen": -2537.9452024647885,
      "eval/logps/rejected": -2313.8192246835442,
      "eval/rewards/chosen": -225.4286971830986,
      "eval/rewards/margins": -20.483273429142884,
      "eval/rewards/rejected": -204.9454237539557,
      "eval_loss": 0.47333332896232605,
      "eval_runtime": 143.3034,
      "eval_samples_per_second": 2.093,
      "eval_steps_per_second": 0.523,
      "step": 50
    },
    {
      "epoch": 0.41,
      "grad_norm": 0.0,
      "learning_rate": 0.00012285714285714287,
      "loss": 0.5,
      "step": 60,
      "train/kl": 0.0,
      "train/logps/chosen": -2750.593359375,
      "train/logps/rejected": -2645.1216796875,
      "train/rewards/chosen": -247.13798828125,
      "train/rewards/margins": -11.170458984374989,
      "train/rewards/rejected": -235.967529296875
    },
    {
      "epoch": 0.55,
      "grad_norm": 0.0,
      "learning_rate": 9.428571428571429e-05,
      "loss": 0.425,
      "step": 80,
      "train/kl": 0.0,
      "train/logps/chosen": -2778.7603400735293,
      "train/logps/rejected": -2450.7654551630435,
      "train/rewards/chosen": -248.5206801470588,
      "train/rewards/margins": -31.70898687260228,
      "train/rewards/rejected": -216.81169327445653
    },
    {
      "epoch": 0.68,
      "grad_norm": 0.0,
      "learning_rate": 6.571428571428571e-05,
      "loss": 0.5031,
      "step": 100,
      "train/kl": 0.0,
      "train/logps/chosen": -2629.9400232919256,
      "train/logps/rejected": -2546.1786556603774,
      "train/rewards/chosen": -234.98452057453417,
      "train/rewards/margins": -9.083110393716566,
      "train/rewards/rejected": -225.9014101808176
    },
    {
      "epoch": 0.68,
      "eval/kl": 0.0,
      "eval/logps/chosen": -2499.3208626760565,
      "eval/logps/rejected": -2280.931566455696,
      "eval/rewards/chosen": -221.56628246038733,
      "eval/rewards/margins": -19.909649252317706,
      "eval/rewards/rejected": -201.65663320806962,
      "eval_loss": 0.47333332896232605,
      "eval_runtime": 143.2853,
      "eval_samples_per_second": 2.094,
      "eval_steps_per_second": 0.523,
      "step": 100
    }
  ],
  "logging_steps": 20,
  "max_steps": 145,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 100,
  "total_flos": 0.0,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}