trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 1.0,
  "eval_steps": 500,
  "global_step": 80,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.01,
      "learning_rate": 6.25e-08,
      "logits/chosen": -2.6214611530303955,
      "logits/rejected": -2.5660409927368164,
      "logps/chosen": -188.08230590820312,
      "logps/pi_response": -124.86982727050781,
      "logps/ref_response": -124.86982727050781,
      "logps/rejected": -255.6346435546875,
      "loss": 0.6931,
      "rewards/accuracies": 0.0,
      "rewards/chosen": 0.0,
      "rewards/margins": 0.0,
      "rewards/rejected": 0.0,
      "step": 1
    },
    {
      "epoch": 0.12,
      "learning_rate": 4.990486745229364e-07,
      "logits/chosen": -2.51465106010437,
      "logits/rejected": -2.499199867248535,
      "logps/chosen": -234.35617065429688,
      "logps/pi_response": -144.1409912109375,
      "logps/ref_response": -142.20367431640625,
      "logps/rejected": -258.1360778808594,
      "loss": 0.6869,
      "rewards/accuracies": 0.4791666567325592,
      "rewards/chosen": -0.029688755050301552,
      "rewards/margins": 0.012110848911106586,
      "rewards/rejected": -0.04179960489273071,
      "step": 10
    },
    {
      "epoch": 0.25,
      "learning_rate": 4.6650635094610966e-07,
      "logits/chosen": -2.4308013916015625,
      "logits/rejected": -2.438734769821167,
      "logps/chosen": -228.0776824951172,
      "logps/pi_response": -156.75161743164062,
      "logps/ref_response": -143.56219482421875,
      "logps/rejected": -305.380615234375,
      "loss": 0.6582,
      "rewards/accuracies": 0.59375,
      "rewards/chosen": -0.21737465262413025,
      "rewards/margins": 0.20173434913158417,
      "rewards/rejected": -0.4191090166568756,
      "step": 20
    },
    {
      "epoch": 0.38,
      "learning_rate": 3.933941090877615e-07,
      "logits/chosen": -2.440542459487915,
      "logits/rejected": -2.3943440914154053,
      "logps/chosen": -224.8809051513672,
      "logps/pi_response": -155.57861328125,
      "logps/ref_response": -143.30210876464844,
      "logps/rejected": -315.95989990234375,
      "loss": 0.6526,
      "rewards/accuracies": 0.581250011920929,
      "rewards/chosen": -0.2337447851896286,
      "rewards/margins": 0.252723753452301,
      "rewards/rejected": -0.48646849393844604,
      "step": 30
    },
    {
      "epoch": 0.5,
      "learning_rate": 2.934120444167326e-07,
      "logits/chosen": -2.3429083824157715,
      "logits/rejected": -2.3265538215637207,
      "logps/chosen": -307.50848388671875,
      "logps/pi_response": -218.9744110107422,
      "logps/ref_response": -160.5868377685547,
      "logps/rejected": -345.1627502441406,
      "loss": 0.6284,
      "rewards/accuracies": 0.6187499761581421,
      "rewards/chosen": -0.6049060821533203,
      "rewards/margins": 0.14748108386993408,
      "rewards/rejected": -0.7523871660232544,
      "step": 40
    },
    {
      "epoch": 0.62,
      "learning_rate": 1.8529523872436977e-07,
      "logits/chosen": -2.3797526359558105,
      "logits/rejected": -2.336573839187622,
      "logps/chosen": -271.7038269042969,
      "logps/pi_response": -180.560302734375,
      "logps/ref_response": -147.65762329101562,
      "logps/rejected": -290.31854248046875,
      "loss": 0.6262,
      "rewards/accuracies": 0.65625,
      "rewards/chosen": -0.319719135761261,
      "rewards/margins": 0.23344525694847107,
      "rewards/rejected": -0.5531644225120544,
      "step": 50
    },
    {
      "epoch": 0.75,
      "learning_rate": 8.930309757836516e-08,
      "logits/chosen": -2.3760523796081543,
      "logits/rejected": -2.32749605178833,
      "logps/chosen": -274.39361572265625,
      "logps/pi_response": -180.96006774902344,
      "logps/ref_response": -138.89596557617188,
      "logps/rejected": -319.18914794921875,
      "loss": 0.6005,
      "rewards/accuracies": 0.6625000238418579,
      "rewards/chosen": -0.49327319860458374,
      "rewards/margins": 0.29331183433532715,
      "rewards/rejected": -0.7865849733352661,
      "step": 60
    },
    {
      "epoch": 0.88,
      "learning_rate": 2.3423053240837514e-08,
      "logits/chosen": -2.4027113914489746,
      "logits/rejected": -2.3286519050598145,
      "logps/chosen": -291.0926818847656,
      "logps/pi_response": -238.1649627685547,
      "logps/ref_response": -181.65982055664062,
      "logps/rejected": -385.4344177246094,
      "loss": 0.6027,
      "rewards/accuracies": 0.737500011920929,
      "rewards/chosen": -0.44906187057495117,
      "rewards/margins": 0.48221874237060547,
      "rewards/rejected": -0.9312806129455566,
      "step": 70
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0,
      "logits/chosen": -2.3703107833862305,
      "logits/rejected": -2.331961154937744,
      "logps/chosen": -283.42547607421875,
      "logps/pi_response": -191.7020263671875,
      "logps/ref_response": -152.6740264892578,
      "logps/rejected": -312.0522766113281,
      "loss": 0.6139,
      "rewards/accuracies": 0.637499988079071,
      "rewards/chosen": -0.4424005448818207,
      "rewards/margins": 0.24074189364910126,
      "rewards/rejected": -0.6831424832344055,
      "step": 80
    },
    {
      "epoch": 1.0,
      "step": 80,
      "total_flos": 0.0,
      "train_loss": 0.633746999502182,
      "train_runtime": 2276.5265,
      "train_samples_per_second": 4.476,
      "train_steps_per_second": 0.035
    }
  ],
  "logging_steps": 10,
  "max_steps": 80,
  "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
}