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": 7.77264868494275,
      "learning_rate": 8.333333333333333e-08,
      "logits/chosen": -2.807276487350464,
      "logits/rejected": -2.7759768962860107,
      "logps/chosen": -315.42626953125,
      "logps/rejected": -227.5915985107422,
      "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": 7.445776158724625,
      "learning_rate": 4.930057285201027e-07,
      "logits/chosen": -2.755523204803467,
      "logits/rejected": -2.746175765991211,
      "logps/chosen": -271.99896240234375,
      "logps/rejected": -260.7083740234375,
      "loss": 0.6915,
      "rewards/accuracies": 0.5208333134651184,
      "rewards/chosen": 0.004709932953119278,
      "rewards/margins": 0.0031843450851738453,
      "rewards/rejected": 0.0015255878679454327,
      "step": 10
    },
    {
      "epoch": 0.33472803347280333,
      "grad_norm": 6.713329773051,
      "learning_rate": 4.187457503795526e-07,
      "logits/chosen": -2.786980152130127,
      "logits/rejected": -2.768200159072876,
      "logps/chosen": -261.47723388671875,
      "logps/rejected": -250.88986206054688,
      "loss": 0.6745,
      "rewards/accuracies": 0.7124999761581421,
      "rewards/chosen": 0.0320444330573082,
      "rewards/margins": 0.04051815718412399,
      "rewards/rejected": -0.008473716676235199,
      "step": 20
    },
    {
      "epoch": 0.502092050209205,
      "grad_norm": 7.420805855578415,
      "learning_rate": 2.8691164100062034e-07,
      "logits/chosen": -2.7878026962280273,
      "logits/rejected": -2.768773317337036,
      "logps/chosen": -293.83514404296875,
      "logps/rejected": -254.23193359375,
      "loss": 0.6486,
      "rewards/accuracies": 0.6812499761581421,
      "rewards/chosen": -0.02379072643816471,
      "rewards/margins": 0.12694817781448364,
      "rewards/rejected": -0.150738924741745,
      "step": 30
    },
    {
      "epoch": 0.6694560669456067,
      "grad_norm": 8.103120434319091,
      "learning_rate": 1.4248369943086995e-07,
      "logits/chosen": -2.760453701019287,
      "logits/rejected": -2.737915515899658,
      "logps/chosen": -264.1527404785156,
      "logps/rejected": -255.6009521484375,
      "loss": 0.6354,
      "rewards/accuracies": 0.684374988079071,
      "rewards/chosen": -0.11233267933130264,
      "rewards/margins": 0.1605701595544815,
      "rewards/rejected": -0.27290281653404236,
      "step": 40
    },
    {
      "epoch": 0.8368200836820083,
      "grad_norm": 8.477755620407805,
      "learning_rate": 3.473909705816111e-08,
      "logits/chosen": -2.7701656818389893,
      "logits/rejected": -2.7501308917999268,
      "logps/chosen": -279.05072021484375,
      "logps/rejected": -287.25872802734375,
      "loss": 0.62,
      "rewards/accuracies": 0.628125011920929,
      "rewards/chosen": -0.2204204499721527,
      "rewards/margins": 0.13704219460487366,
      "rewards/rejected": -0.35746267437934875,
      "step": 50
    },
    {
      "epoch": 0.9874476987447699,
      "step": 59,
      "total_flos": 0.0,
      "train_loss": 0.649006043450307,
      "train_runtime": 1632.9184,
      "train_samples_per_second": 9.359,
      "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
}