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": 7.774161688327495,
      "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": 7.451268268455854,
      "learning_rate": 4.930057285201027e-07,
      "logits/chosen": -2.7555460929870605,
      "logits/rejected": -2.7461297512054443,
      "logps/chosen": -271.95458984375,
      "logps/rejected": -260.7107849121094,
      "loss": 0.6914,
      "rewards/accuracies": 0.5347222089767456,
      "rewards/chosen": 0.005153654608875513,
      "rewards/margins": 0.0036522927694022655,
      "rewards/rejected": 0.001501361490227282,
      "step": 10
    },
    {
      "epoch": 0.33472803347280333,
      "grad_norm": 6.7050372954284905,
      "learning_rate": 4.187457503795526e-07,
      "logits/chosen": -2.7869606018066406,
      "logits/rejected": -2.76808762550354,
      "logps/chosen": -261.46710205078125,
      "logps/rejected": -250.8916015625,
      "loss": 0.6746,
      "rewards/accuracies": 0.690625011920929,
      "rewards/chosen": 0.032145868986845016,
      "rewards/margins": 0.04063734412193298,
      "rewards/rejected": -0.008491471409797668,
      "step": 20
    },
    {
      "epoch": 0.502092050209205,
      "grad_norm": 7.445996241869226,
      "learning_rate": 2.8691164100062034e-07,
      "logits/chosen": -2.7877469062805176,
      "logits/rejected": -2.768691062927246,
      "logps/chosen": -293.81890869140625,
      "logps/rejected": -254.21170043945312,
      "loss": 0.6486,
      "rewards/accuracies": 0.675000011920929,
      "rewards/chosen": -0.023628313094377518,
      "rewards/margins": 0.12690803408622742,
      "rewards/rejected": -0.15053634345531464,
      "step": 30
    },
    {
      "epoch": 0.6694560669456067,
      "grad_norm": 8.132911479301741,
      "learning_rate": 1.4248369943086995e-07,
      "logits/chosen": -2.7604308128356934,
      "logits/rejected": -2.7378690242767334,
      "logps/chosen": -264.1637878417969,
      "logps/rejected": -255.6041259765625,
      "loss": 0.6353,
      "rewards/accuracies": 0.6875,
      "rewards/chosen": -0.11244319379329681,
      "rewards/margins": 0.16049158573150635,
      "rewards/rejected": -0.27293476462364197,
      "step": 40
    },
    {
      "epoch": 0.8368200836820083,
      "grad_norm": 8.515463629957392,
      "learning_rate": 3.473909705816111e-08,
      "logits/chosen": -2.7702841758728027,
      "logits/rejected": -2.750366449356079,
      "logps/chosen": -279.08843994140625,
      "logps/rejected": -287.2726745605469,
      "loss": 0.6201,
      "rewards/accuracies": 0.628125011920929,
      "rewards/chosen": -0.22079789638519287,
      "rewards/margins": 0.13680399954319,
      "rewards/rejected": -0.35760191082954407,
      "step": 50
    },
    {
      "epoch": 0.9874476987447699,
      "step": 59,
      "total_flos": 0.0,
      "train_loss": 0.649011991791806,
      "train_runtime": 1630.9121,
      "train_samples_per_second": 9.371,
      "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
}