last-checkpoint/trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.005925925925925926,
  "eval_steps": 500,
  "global_step": 10,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.0005925925925925926,
      "grad_norm": 10.858694076538086,
      "learning_rate": 5e-07,
      "logits/chosen": -4.444676399230957,
      "logits/rejected": -4.0909342765808105,
      "logps/chosen": -186.1875,
      "logps/rejected": -228.68560791015625,
      "loss": 0.6931,
      "rewards/accuracies": 0.0,
      "rewards/chosen": 0.0,
      "rewards/margins": 0.0,
      "rewards/rejected": 0.0,
      "step": 1
    },
    {
      "epoch": 0.0011851851851851852,
      "grad_norm": 7.565941333770752,
      "learning_rate": 4.849231551964771e-07,
      "logits/chosen": -4.530362606048584,
      "logits/rejected": -3.981240749359131,
      "logps/chosen": -192.2100830078125,
      "logps/rejected": -213.73086547851562,
      "loss": 0.6931,
      "rewards/accuracies": 0.0,
      "rewards/chosen": 0.0,
      "rewards/margins": 0.0,
      "rewards/rejected": 0.0,
      "step": 2
    },
    {
      "epoch": 0.0017777777777777779,
      "grad_norm": 13.622398376464844,
      "learning_rate": 4.415111107797445e-07,
      "logits/chosen": -4.5415449142456055,
      "logits/rejected": -4.322567462921143,
      "logps/chosen": -200.85960388183594,
      "logps/rejected": -224.73626708984375,
      "loss": 0.6619,
      "rewards/accuracies": 1.0,
      "rewards/chosen": 0.03391075134277344,
      "rewards/margins": 0.06416893005371094,
      "rewards/rejected": -0.03025818057358265,
      "step": 3
    },
    {
      "epoch": 0.0023703703703703703,
      "grad_norm": 10.704012870788574,
      "learning_rate": 3.75e-07,
      "logits/chosen": -3.9766385555267334,
      "logits/rejected": -4.474390029907227,
      "logps/chosen": -239.95339965820312,
      "logps/rejected": -202.6293182373047,
      "loss": 0.7176,
      "rewards/accuracies": 0.25,
      "rewards/chosen": -0.037302400916814804,
      "rewards/margins": -0.04801139980554581,
      "rewards/rejected": 0.010708999820053577,
      "step": 4
    },
    {
      "epoch": 0.002962962962962963,
      "grad_norm": 6.752383232116699,
      "learning_rate": 2.934120444167326e-07,
      "logits/chosen": -4.133105754852295,
      "logits/rejected": -4.693289279937744,
      "logps/chosen": -168.95433044433594,
      "logps/rejected": -149.30447387695312,
      "loss": 0.6772,
      "rewards/accuracies": 0.75,
      "rewards/chosen": 0.03451080620288849,
      "rewards/margins": 0.03242149576544762,
      "rewards/rejected": 0.002089309971779585,
      "step": 5
    },
    {
      "epoch": 0.0035555555555555557,
      "grad_norm": 12.878663063049316,
      "learning_rate": 2.065879555832674e-07,
      "logits/chosen": -5.0531182289123535,
      "logits/rejected": -5.144861221313477,
      "logps/chosen": -193.86325073242188,
      "logps/rejected": -202.404052734375,
      "loss": 0.6904,
      "rewards/accuracies": 0.5,
      "rewards/chosen": -0.009803391061723232,
      "rewards/margins": 0.0061431871727108955,
      "rewards/rejected": -0.015946578234434128,
      "step": 6
    },
    {
      "epoch": 0.004148148148148148,
      "grad_norm": 6.771181106567383,
      "learning_rate": 1.2500000000000005e-07,
      "logits/chosen": -4.2494401931762695,
      "logits/rejected": -3.684239387512207,
      "logps/chosen": -176.15640258789062,
      "logps/rejected": -223.1680450439453,
      "loss": 0.6937,
      "rewards/accuracies": 0.25,
      "rewards/chosen": 0.02459411695599556,
      "rewards/margins": -0.0004608146846294403,
      "rewards/rejected": 0.025054931640625,
      "step": 7
    },
    {
      "epoch": 0.004740740740740741,
      "grad_norm": 11.16915225982666,
      "learning_rate": 5.848888922025552e-08,
      "logits/chosen": -4.407987117767334,
      "logits/rejected": -4.415105819702148,
      "logps/chosen": -182.2259063720703,
      "logps/rejected": -196.61068725585938,
      "loss": 0.6884,
      "rewards/accuracies": 0.5,
      "rewards/chosen": -0.05178241804242134,
      "rewards/margins": 0.010274123400449753,
      "rewards/rejected": -0.062056541442871094,
      "step": 8
    },
    {
      "epoch": 0.005333333333333333,
      "grad_norm": 9.21722412109375,
      "learning_rate": 1.507684480352292e-08,
      "logits/chosen": -4.9174699783325195,
      "logits/rejected": -4.600230693817139,
      "logps/chosen": -154.47506713867188,
      "logps/rejected": -178.89224243164062,
      "loss": 0.7014,
      "rewards/accuracies": 0.25,
      "rewards/chosen": 0.024329185485839844,
      "rewards/margins": -0.01587962917983532,
      "rewards/rejected": 0.04020881652832031,
      "step": 9
    },
    {
      "epoch": 0.005925925925925926,
      "grad_norm": 16.037979125976562,
      "learning_rate": 0.0,
      "logits/chosen": -3.811877965927124,
      "logits/rejected": -3.696570873260498,
      "logps/chosen": -232.4647216796875,
      "logps/rejected": -253.81515502929688,
      "loss": 0.7093,
      "rewards/accuracies": 0.25,
      "rewards/chosen": -0.00266952533274889,
      "rewards/margins": -0.03174018859863281,
      "rewards/rejected": 0.029070664197206497,
      "step": 10
    }
  ],
  "logging_steps": 1,
  "max_steps": 10,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 0.0,
  "train_batch_size": 1,
  "trial_name": null,
  "trial_params": null
}