{
  "best_metric": 0.8454458293384468,
  "best_model_checkpoint": "test4-pairwise/checkpoint-516",
  "epoch": 2.0,
  "eval_steps": 500,
  "global_step": 516,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.08,
      "learning_rate": 3.0769230769230774e-05,
      "loss": 0.6865,
      "step": 20
    },
    {
      "epoch": 0.16,
      "learning_rate": 5.538461538461539e-05,
      "loss": 0.3337,
      "step": 40
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.998533144893137e-05,
      "loss": 0.1844,
      "step": 60
    },
    {
      "epoch": 0.31,
      "learning_rate": 7.947306090166105e-05,
      "loss": 0.1622,
      "step": 80
    },
    {
      "epoch": 0.39,
      "learning_rate": 7.823808570686448e-05,
      "loss": 0.1337,
      "step": 100
    },
    {
      "epoch": 0.47,
      "learning_rate": 7.63030167868383e-05,
      "loss": 0.1452,
      "step": 120
    },
    {
      "epoch": 0.54,
      "learning_rate": 7.370328294466597e-05,
      "loss": 0.0922,
      "step": 140
    },
    {
      "epoch": 0.62,
      "learning_rate": 7.048648220510546e-05,
      "loss": 0.0946,
      "step": 160
    },
    {
      "epoch": 0.7,
      "learning_rate": 6.671151035154783e-05,
      "loss": 0.0903,
      "step": 180
    },
    {
      "epoch": 0.78,
      "learning_rate": 6.24474826144953e-05,
      "loss": 0.075,
      "step": 200
    },
    {
      "epoch": 0.85,
      "learning_rate": 5.7772468254156155e-05,
      "loss": 0.0801,
      "step": 220
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.27720612054392e-05,
      "loss": 0.0864,
      "step": 240
    },
    {
      "epoch": 1.0,
      "eval_loss": 2.4745593070983887,
      "eval_runtime": 8.6606,
      "eval_samples_per_second": 10129.044,
      "eval_steps_per_second": 79.209,
      "eval_top-1-accuracy": 0.7670182166826462,
      "eval_top-10-accuracy": 0.9311601150527326,
      "eval_top-3-accuracy": 0.8186001917545541,
      "eval_top-5-accuracy": 0.8705656759348035,
      "step": 258
    },
    {
      "epoch": 1.01,
      "learning_rate": 4.753781295513274e-05,
      "loss": 0.0629,
      "step": 260
    },
    {
      "epoch": 1.09,
      "learning_rate": 4.216555634341671e-05,
      "loss": 0.0654,
      "step": 280
    },
    {
      "epoch": 1.16,
      "learning_rate": 3.675365097890276e-05,
      "loss": 0.0585,
      "step": 300
    },
    {
      "epoch": 1.24,
      "learning_rate": 3.140118239155904e-05,
      "loss": 0.0525,
      "step": 320
    },
    {
      "epoch": 1.32,
      "learning_rate": 2.6206147894881193e-05,
      "loss": 0.0377,
      "step": 340
    },
    {
      "epoch": 1.4,
      "learning_rate": 2.12636623720084e-05,
      "loss": 0.0415,
      "step": 360
    },
    {
      "epoch": 1.47,
      "learning_rate": 1.666421683570024e-05,
      "loss": 0.0453,
      "step": 380
    },
    {
      "epoch": 1.55,
      "learning_rate": 1.2492021645863064e-05,
      "loss": 0.0457,
      "step": 400
    },
    {
      "epoch": 1.63,
      "learning_rate": 8.823464718336336e-06,
      "loss": 0.0369,
      "step": 420
    },
    {
      "epoch": 1.71,
      "learning_rate": 5.725712953296439e-06,
      "loss": 0.0272,
      "step": 440
    },
    {
      "epoch": 1.78,
      "learning_rate": 3.2554824894551306e-06,
      "loss": 0.0316,
      "step": 460
    },
    {
      "epoch": 1.86,
      "learning_rate": 1.4580002992310837e-06,
      "loss": 0.0363,
      "step": 480
    },
    {
      "epoch": 1.94,
      "learning_rate": 3.66176136847729e-07,
      "loss": 0.0411,
      "step": 500
    },
    {
      "epoch": 2.0,
      "eval_loss": 3.0463552474975586,
      "eval_runtime": 8.6788,
      "eval_samples_per_second": 10107.877,
      "eval_steps_per_second": 79.043,
      "eval_top-1-accuracy": 0.7946308724832215,
      "eval_top-10-accuracy": 0.9497603068072866,
      "eval_top-3-accuracy": 0.8454458293384468,
      "eval_top-5-accuracy": 0.8928092042186002,
      "step": 516
    }
  ],
  "logging_steps": 20,
  "max_steps": 516,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 2,
  "save_steps": 500,
  "total_flos": 2415688357511168.0,
  "train_batch_size": 64,
  "trial_name": null,
  "trial_params": null
}