{
  "best_metric": 0.6654545454545454,
  "best_model_checkpoint": "BEiT-RD-DA\\checkpoint-2112",
  "epoch": 40.0,
  "eval_steps": 500,
  "global_step": 3840,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.1,
      "learning_rate": 1.3020833333333335e-06,
      "loss": 1.6093,
      "step": 10
    },
    {
      "epoch": 0.21,
      "learning_rate": 2.604166666666667e-06,
      "loss": 1.6082,
      "step": 20
    },
    {
      "epoch": 0.31,
      "learning_rate": 3.90625e-06,
      "loss": 1.6045,
      "step": 30
    },
    {
      "epoch": 0.42,
      "learning_rate": 5.208333333333334e-06,
      "loss": 1.5939,
      "step": 40
    },
    {
      "epoch": 0.52,
      "learning_rate": 6.510416666666667e-06,
      "loss": 1.576,
      "step": 50
    },
    {
      "epoch": 0.62,
      "learning_rate": 7.8125e-06,
      "loss": 1.5501,
      "step": 60
    },
    {
      "epoch": 0.73,
      "learning_rate": 9.114583333333334e-06,
      "loss": 1.5116,
      "step": 70
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.0416666666666668e-05,
      "loss": 1.462,
      "step": 80
    },
    {
      "epoch": 0.94,
      "learning_rate": 1.171875e-05,
      "loss": 1.4123,
      "step": 90
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.49272727272727274,
      "eval_loss": 1.4099408388137817,
      "eval_runtime": 7.0436,
      "eval_samples_per_second": 78.085,
      "eval_steps_per_second": 4.969,
      "step": 96
    },
    {
      "epoch": 1.04,
      "learning_rate": 1.3020833333333334e-05,
      "loss": 1.3572,
      "step": 100
    },
    {
      "epoch": 1.15,
      "learning_rate": 1.4322916666666666e-05,
      "loss": 1.2997,
      "step": 110
    },
    {
      "epoch": 1.25,
      "learning_rate": 1.5625e-05,
      "loss": 1.2359,
      "step": 120
    },
    {
      "epoch": 1.35,
      "learning_rate": 1.6927083333333336e-05,
      "loss": 1.1928,
      "step": 130
    },
    {
      "epoch": 1.46,
      "learning_rate": 1.8229166666666668e-05,
      "loss": 1.1188,
      "step": 140
    },
    {
      "epoch": 1.56,
      "learning_rate": 1.953125e-05,
      "loss": 1.0731,
      "step": 150
    },
    {
      "epoch": 1.67,
      "learning_rate": 2.0833333333333336e-05,
      "loss": 1.0396,
      "step": 160
    },
    {
      "epoch": 1.77,
      "learning_rate": 2.2135416666666668e-05,
      "loss": 0.9933,
      "step": 170
    },
    {
      "epoch": 1.88,
      "learning_rate": 2.34375e-05,
      "loss": 0.9498,
      "step": 180
    },
    {
      "epoch": 1.98,
      "learning_rate": 2.4739583333333336e-05,
      "loss": 0.9503,
      "step": 190
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.49272727272727274,
      "eval_loss": 1.8851675987243652,
      "eval_runtime": 3.6814,
      "eval_samples_per_second": 149.401,
      "eval_steps_per_second": 9.507,
      "step": 192
    },
    {
      "epoch": 2.08,
      "learning_rate": 2.604166666666667e-05,
      "loss": 0.9379,
      "step": 200
    },
    {
      "epoch": 2.19,
      "learning_rate": 2.734375e-05,
      "loss": 0.8963,
      "step": 210
    },
    {
      "epoch": 2.29,
      "learning_rate": 2.8645833333333333e-05,
      "loss": 0.9422,
      "step": 220
    },
    {
      "epoch": 2.4,
      "learning_rate": 2.994791666666667e-05,
      "loss": 0.904,
      "step": 230
    },
    {
      "epoch": 2.5,
      "learning_rate": 3.125e-05,
      "loss": 0.9131,
      "step": 240
    },
    {
      "epoch": 2.6,
      "learning_rate": 3.255208333333333e-05,
      "loss": 0.9078,
      "step": 250
    },
    {
      "epoch": 2.71,
      "learning_rate": 3.385416666666667e-05,
      "loss": 0.851,
      "step": 260
    },
    {
      "epoch": 2.81,
      "learning_rate": 3.5156250000000004e-05,
      "loss": 0.9054,
      "step": 270
    },
    {
      "epoch": 2.92,
      "learning_rate": 3.6458333333333336e-05,
      "loss": 0.8284,
      "step": 280
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.5072727272727273,
      "eval_loss": 2.1701955795288086,
      "eval_runtime": 3.6684,
      "eval_samples_per_second": 149.93,
      "eval_steps_per_second": 9.541,
      "step": 288
    },
    {
      "epoch": 3.02,
      "learning_rate": 3.776041666666667e-05,
      "loss": 0.7817,
      "step": 290
    },
    {
      "epoch": 3.12,
      "learning_rate": 3.90625e-05,
      "loss": 0.8264,
      "step": 300
    },
    {
      "epoch": 3.23,
      "learning_rate": 4.036458333333333e-05,
      "loss": 0.8271,
      "step": 310
    },
    {
      "epoch": 3.33,
      "learning_rate": 4.166666666666667e-05,
      "loss": 0.8185,
      "step": 320
    },
    {
      "epoch": 3.44,
      "learning_rate": 4.2968750000000004e-05,
      "loss": 0.7828,
      "step": 330
    },
    {
      "epoch": 3.54,
      "learning_rate": 4.4270833333333337e-05,
      "loss": 0.7351,
      "step": 340
    },
    {
      "epoch": 3.65,
      "learning_rate": 4.557291666666667e-05,
      "loss": 0.7248,
      "step": 350
    },
    {
      "epoch": 3.75,
      "learning_rate": 4.6875e-05,
      "loss": 0.7981,
      "step": 360
    },
    {
      "epoch": 3.85,
      "learning_rate": 4.817708333333333e-05,
      "loss": 0.8329,
      "step": 370
    },
    {
      "epoch": 3.96,
      "learning_rate": 4.947916666666667e-05,
      "loss": 0.7677,
      "step": 380
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.5345454545454545,
      "eval_loss": 2.0407912731170654,
      "eval_runtime": 4.124,
      "eval_samples_per_second": 133.366,
      "eval_steps_per_second": 8.487,
      "step": 384
    },
    {
      "epoch": 4.06,
      "learning_rate": 4.991319444444445e-05,
      "loss": 0.7784,
      "step": 390
    },
    {
      "epoch": 4.17,
      "learning_rate": 4.976851851851852e-05,
      "loss": 0.7742,
      "step": 400
    },
    {
      "epoch": 4.27,
      "learning_rate": 4.96238425925926e-05,
      "loss": 0.7786,
      "step": 410
    },
    {
      "epoch": 4.38,
      "learning_rate": 4.947916666666667e-05,
      "loss": 0.7264,
      "step": 420
    },
    {
      "epoch": 4.48,
      "learning_rate": 4.933449074074074e-05,
      "loss": 0.752,
      "step": 430
    },
    {
      "epoch": 4.58,
      "learning_rate": 4.9189814814814815e-05,
      "loss": 0.7815,
      "step": 440
    },
    {
      "epoch": 4.69,
      "learning_rate": 4.904513888888889e-05,
      "loss": 0.7081,
      "step": 450
    },
    {
      "epoch": 4.79,
      "learning_rate": 4.8900462962962965e-05,
      "loss": 0.7476,
      "step": 460
    },
    {
      "epoch": 4.9,
      "learning_rate": 4.875578703703704e-05,
      "loss": 0.6773,
      "step": 470
    },
    {
      "epoch": 5.0,
      "learning_rate": 4.8611111111111115e-05,
      "loss": 0.788,
      "step": 480
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.5127272727272727,
      "eval_loss": 2.7991154193878174,
      "eval_runtime": 3.8459,
      "eval_samples_per_second": 143.009,
      "eval_steps_per_second": 9.101,
      "step": 480
    },
    {
      "epoch": 5.1,
      "learning_rate": 4.846643518518519e-05,
      "loss": 0.7058,
      "step": 490
    },
    {
      "epoch": 5.21,
      "learning_rate": 4.8321759259259265e-05,
      "loss": 0.7083,
      "step": 500
    },
    {
      "epoch": 5.31,
      "learning_rate": 4.817708333333333e-05,
      "loss": 0.6479,
      "step": 510
    },
    {
      "epoch": 5.42,
      "learning_rate": 4.803240740740741e-05,
      "loss": 0.658,
      "step": 520
    },
    {
      "epoch": 5.52,
      "learning_rate": 4.788773148148148e-05,
      "loss": 0.7756,
      "step": 530
    },
    {
      "epoch": 5.62,
      "learning_rate": 4.774305555555556e-05,
      "loss": 0.6294,
      "step": 540
    },
    {
      "epoch": 5.73,
      "learning_rate": 4.759837962962963e-05,
      "loss": 0.6727,
      "step": 550
    },
    {
      "epoch": 5.83,
      "learning_rate": 4.745370370370371e-05,
      "loss": 0.6476,
      "step": 560
    },
    {
      "epoch": 5.94,
      "learning_rate": 4.730902777777778e-05,
      "loss": 0.5822,
      "step": 570
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.5636363636363636,
      "eval_loss": 2.0950703620910645,
      "eval_runtime": 3.9745,
      "eval_samples_per_second": 138.384,
      "eval_steps_per_second": 8.806,
      "step": 576
    },
    {
      "epoch": 6.04,
      "learning_rate": 4.716435185185186e-05,
      "loss": 0.5656,
      "step": 580
    },
    {
      "epoch": 6.15,
      "learning_rate": 4.7019675925925926e-05,
      "loss": 0.6338,
      "step": 590
    },
    {
      "epoch": 6.25,
      "learning_rate": 4.6875e-05,
      "loss": 0.62,
      "step": 600
    },
    {
      "epoch": 6.35,
      "learning_rate": 4.6730324074074076e-05,
      "loss": 0.5603,
      "step": 610
    },
    {
      "epoch": 6.46,
      "learning_rate": 4.658564814814815e-05,
      "loss": 0.5726,
      "step": 620
    },
    {
      "epoch": 6.56,
      "learning_rate": 4.6440972222222226e-05,
      "loss": 0.5767,
      "step": 630
    },
    {
      "epoch": 6.67,
      "learning_rate": 4.62962962962963e-05,
      "loss": 0.5546,
      "step": 640
    },
    {
      "epoch": 6.77,
      "learning_rate": 4.6151620370370376e-05,
      "loss": 0.5003,
      "step": 650
    },
    {
      "epoch": 6.88,
      "learning_rate": 4.6006944444444444e-05,
      "loss": 0.564,
      "step": 660
    },
    {
      "epoch": 6.98,
      "learning_rate": 4.586226851851852e-05,
      "loss": 0.5172,
      "step": 670
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.5363636363636364,
      "eval_loss": 2.5976808071136475,
      "eval_runtime": 4.1525,
      "eval_samples_per_second": 132.451,
      "eval_steps_per_second": 8.429,
      "step": 672
    },
    {
      "epoch": 7.08,
      "learning_rate": 4.5717592592592594e-05,
      "loss": 0.4963,
      "step": 680
    },
    {
      "epoch": 7.19,
      "learning_rate": 4.557291666666667e-05,
      "loss": 0.5223,
      "step": 690
    },
    {
      "epoch": 7.29,
      "learning_rate": 4.5428240740740744e-05,
      "loss": 0.5279,
      "step": 700
    },
    {
      "epoch": 7.4,
      "learning_rate": 4.528356481481482e-05,
      "loss": 0.4685,
      "step": 710
    },
    {
      "epoch": 7.5,
      "learning_rate": 4.5138888888888894e-05,
      "loss": 0.4652,
      "step": 720
    },
    {
      "epoch": 7.6,
      "learning_rate": 4.499421296296297e-05,
      "loss": 0.3945,
      "step": 730
    },
    {
      "epoch": 7.71,
      "learning_rate": 4.484953703703704e-05,
      "loss": 0.4617,
      "step": 740
    },
    {
      "epoch": 7.81,
      "learning_rate": 4.470486111111111e-05,
      "loss": 0.471,
      "step": 750
    },
    {
      "epoch": 7.92,
      "learning_rate": 4.456018518518519e-05,
      "loss": 0.4615,
      "step": 760
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.58,
      "eval_loss": 2.096754789352417,
      "eval_runtime": 4.078,
      "eval_samples_per_second": 134.871,
      "eval_steps_per_second": 8.583,
      "step": 768
    },
    {
      "epoch": 8.02,
      "learning_rate": 4.441550925925926e-05,
      "loss": 0.4487,
      "step": 770
    },
    {
      "epoch": 8.12,
      "learning_rate": 4.4270833333333337e-05,
      "loss": 0.4692,
      "step": 780
    },
    {
      "epoch": 8.23,
      "learning_rate": 4.412615740740741e-05,
      "loss": 0.4161,
      "step": 790
    },
    {
      "epoch": 8.33,
      "learning_rate": 4.3981481481481486e-05,
      "loss": 0.4709,
      "step": 800
    },
    {
      "epoch": 8.44,
      "learning_rate": 4.383680555555556e-05,
      "loss": 0.4199,
      "step": 810
    },
    {
      "epoch": 8.54,
      "learning_rate": 4.369212962962963e-05,
      "loss": 0.4416,
      "step": 820
    },
    {
      "epoch": 8.65,
      "learning_rate": 4.3547453703703704e-05,
      "loss": 0.4253,
      "step": 830
    },
    {
      "epoch": 8.75,
      "learning_rate": 4.340277777777778e-05,
      "loss": 0.3953,
      "step": 840
    },
    {
      "epoch": 8.85,
      "learning_rate": 4.3258101851851854e-05,
      "loss": 0.4495,
      "step": 850
    },
    {
      "epoch": 8.96,
      "learning_rate": 4.311342592592593e-05,
      "loss": 0.3672,
      "step": 860
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.5436363636363636,
      "eval_loss": 2.8535282611846924,
      "eval_runtime": 3.6989,
      "eval_samples_per_second": 148.694,
      "eval_steps_per_second": 9.462,
      "step": 864
    },
    {
      "epoch": 9.06,
      "learning_rate": 4.2968750000000004e-05,
      "loss": 0.3812,
      "step": 870
    },
    {
      "epoch": 9.17,
      "learning_rate": 4.282407407407408e-05,
      "loss": 0.3683,
      "step": 880
    },
    {
      "epoch": 9.27,
      "learning_rate": 4.267939814814815e-05,
      "loss": 0.4462,
      "step": 890
    },
    {
      "epoch": 9.38,
      "learning_rate": 4.253472222222222e-05,
      "loss": 0.31,
      "step": 900
    },
    {
      "epoch": 9.48,
      "learning_rate": 4.23900462962963e-05,
      "loss": 0.3963,
      "step": 910
    },
    {
      "epoch": 9.58,
      "learning_rate": 4.224537037037037e-05,
      "loss": 0.3393,
      "step": 920
    },
    {
      "epoch": 9.69,
      "learning_rate": 4.210069444444445e-05,
      "loss": 0.3853,
      "step": 930
    },
    {
      "epoch": 9.79,
      "learning_rate": 4.195601851851852e-05,
      "loss": 0.4014,
      "step": 940
    },
    {
      "epoch": 9.9,
      "learning_rate": 4.18113425925926e-05,
      "loss": 0.3931,
      "step": 950
    },
    {
      "epoch": 10.0,
      "learning_rate": 4.166666666666667e-05,
      "loss": 0.379,
      "step": 960
    },
    {
      "epoch": 10.0,
      "eval_accuracy": 0.5381818181818182,
      "eval_loss": 2.9514999389648438,
      "eval_runtime": 3.6769,
      "eval_samples_per_second": 149.583,
      "eval_steps_per_second": 9.519,
      "step": 960
    },
    {
      "epoch": 10.1,
      "learning_rate": 4.152199074074074e-05,
      "loss": 0.3544,
      "step": 970
    },
    {
      "epoch": 10.21,
      "learning_rate": 4.1377314814814815e-05,
      "loss": 0.3334,
      "step": 980
    },
    {
      "epoch": 10.31,
      "learning_rate": 4.123263888888889e-05,
      "loss": 0.2906,
      "step": 990
    },
    {
      "epoch": 10.42,
      "learning_rate": 4.1087962962962965e-05,
      "loss": 0.3518,
      "step": 1000
    },
    {
      "epoch": 10.52,
      "learning_rate": 4.094328703703704e-05,
      "loss": 0.3341,
      "step": 1010
    },
    {
      "epoch": 10.62,
      "learning_rate": 4.0798611111111115e-05,
      "loss": 0.3725,
      "step": 1020
    },
    {
      "epoch": 10.73,
      "learning_rate": 4.065393518518519e-05,
      "loss": 0.3435,
      "step": 1030
    },
    {
      "epoch": 10.83,
      "learning_rate": 4.0509259259259265e-05,
      "loss": 0.3219,
      "step": 1040
    },
    {
      "epoch": 10.94,
      "learning_rate": 4.036458333333333e-05,
      "loss": 0.3301,
      "step": 1050
    },
    {
      "epoch": 11.0,
      "eval_accuracy": 0.5581818181818182,
      "eval_loss": 2.7200424671173096,
      "eval_runtime": 3.7269,
      "eval_samples_per_second": 147.576,
      "eval_steps_per_second": 9.391,
      "step": 1056
    },
    {
      "epoch": 11.04,
      "learning_rate": 4.021990740740741e-05,
      "loss": 0.3121,
      "step": 1060
    },
    {
      "epoch": 11.15,
      "learning_rate": 4.007523148148148e-05,
      "loss": 0.3416,
      "step": 1070
    },
    {
      "epoch": 11.25,
      "learning_rate": 3.993055555555556e-05,
      "loss": 0.2896,
      "step": 1080
    },
    {
      "epoch": 11.35,
      "learning_rate": 3.978587962962963e-05,
      "loss": 0.3542,
      "step": 1090
    },
    {
      "epoch": 11.46,
      "learning_rate": 3.964120370370371e-05,
      "loss": 0.3516,
      "step": 1100
    },
    {
      "epoch": 11.56,
      "learning_rate": 3.949652777777778e-05,
      "loss": 0.295,
      "step": 1110
    },
    {
      "epoch": 11.67,
      "learning_rate": 3.935185185185186e-05,
      "loss": 0.3252,
      "step": 1120
    },
    {
      "epoch": 11.77,
      "learning_rate": 3.9207175925925926e-05,
      "loss": 0.2624,
      "step": 1130
    },
    {
      "epoch": 11.88,
      "learning_rate": 3.90625e-05,
      "loss": 0.329,
      "step": 1140
    },
    {
      "epoch": 11.98,
      "learning_rate": 3.8917824074074076e-05,
      "loss": 0.2786,
      "step": 1150
    },
    {
      "epoch": 12.0,
      "eval_accuracy": 0.6272727272727273,
      "eval_loss": 1.8999561071395874,
      "eval_runtime": 3.6939,
      "eval_samples_per_second": 148.893,
      "eval_steps_per_second": 9.475,
      "step": 1152
    },
    {
      "epoch": 12.08,
      "learning_rate": 3.877314814814815e-05,
      "loss": 0.2833,
      "step": 1160
    },
    {
      "epoch": 12.19,
      "learning_rate": 3.8628472222222226e-05,
      "loss": 0.2536,
      "step": 1170
    },
    {
      "epoch": 12.29,
      "learning_rate": 3.84837962962963e-05,
      "loss": 0.3003,
      "step": 1180
    },
    {
      "epoch": 12.4,
      "learning_rate": 3.8339120370370376e-05,
      "loss": 0.2713,
      "step": 1190
    },
    {
      "epoch": 12.5,
      "learning_rate": 3.8194444444444444e-05,
      "loss": 0.2854,
      "step": 1200
    },
    {
      "epoch": 12.6,
      "learning_rate": 3.804976851851852e-05,
      "loss": 0.2695,
      "step": 1210
    },
    {
      "epoch": 12.71,
      "learning_rate": 3.7905092592592594e-05,
      "loss": 0.2702,
      "step": 1220
    },
    {
      "epoch": 12.81,
      "learning_rate": 3.776041666666667e-05,
      "loss": 0.3312,
      "step": 1230
    },
    {
      "epoch": 12.92,
      "learning_rate": 3.7615740740740744e-05,
      "loss": 0.2746,
      "step": 1240
    },
    {
      "epoch": 13.0,
      "eval_accuracy": 0.5363636363636364,
      "eval_loss": 3.1768321990966797,
      "eval_runtime": 3.7274,
      "eval_samples_per_second": 147.555,
      "eval_steps_per_second": 9.39,
      "step": 1248
    },
    {
      "epoch": 13.02,
      "learning_rate": 3.747106481481482e-05,
      "loss": 0.2632,
      "step": 1250
    },
    {
      "epoch": 13.12,
      "learning_rate": 3.7326388888888893e-05,
      "loss": 0.2495,
      "step": 1260
    },
    {
      "epoch": 13.23,
      "learning_rate": 3.718171296296297e-05,
      "loss": 0.2761,
      "step": 1270
    },
    {
      "epoch": 13.33,
      "learning_rate": 3.7037037037037037e-05,
      "loss": 0.2879,
      "step": 1280
    },
    {
      "epoch": 13.44,
      "learning_rate": 3.689236111111111e-05,
      "loss": 0.2658,
      "step": 1290
    },
    {
      "epoch": 13.54,
      "learning_rate": 3.6747685185185186e-05,
      "loss": 0.2652,
      "step": 1300
    },
    {
      "epoch": 13.65,
      "learning_rate": 3.660300925925926e-05,
      "loss": 0.2604,
      "step": 1310
    },
    {
      "epoch": 13.75,
      "learning_rate": 3.6458333333333336e-05,
      "loss": 0.2468,
      "step": 1320
    },
    {
      "epoch": 13.85,
      "learning_rate": 3.631365740740741e-05,
      "loss": 0.2552,
      "step": 1330
    },
    {
      "epoch": 13.96,
      "learning_rate": 3.6168981481481486e-05,
      "loss": 0.2298,
      "step": 1340
    },
    {
      "epoch": 14.0,
      "eval_accuracy": 0.5527272727272727,
      "eval_loss": 3.100266218185425,
      "eval_runtime": 3.7819,
      "eval_samples_per_second": 145.429,
      "eval_steps_per_second": 9.255,
      "step": 1344
    },
    {
      "epoch": 14.06,
      "learning_rate": 3.602430555555556e-05,
      "loss": 0.2423,
      "step": 1350
    },
    {
      "epoch": 14.17,
      "learning_rate": 3.587962962962963e-05,
      "loss": 0.2102,
      "step": 1360
    },
    {
      "epoch": 14.27,
      "learning_rate": 3.5734953703703704e-05,
      "loss": 0.207,
      "step": 1370
    },
    {
      "epoch": 14.38,
      "learning_rate": 3.559027777777778e-05,
      "loss": 0.2548,
      "step": 1380
    },
    {
      "epoch": 14.48,
      "learning_rate": 3.5445601851851854e-05,
      "loss": 0.2428,
      "step": 1390
    },
    {
      "epoch": 14.58,
      "learning_rate": 3.530092592592593e-05,
      "loss": 0.2344,
      "step": 1400
    },
    {
      "epoch": 14.69,
      "learning_rate": 3.5156250000000004e-05,
      "loss": 0.2847,
      "step": 1410
    },
    {
      "epoch": 14.79,
      "learning_rate": 3.501157407407408e-05,
      "loss": 0.2296,
      "step": 1420
    },
    {
      "epoch": 14.9,
      "learning_rate": 3.486689814814815e-05,
      "loss": 0.2894,
      "step": 1430
    },
    {
      "epoch": 15.0,
      "learning_rate": 3.472222222222222e-05,
      "loss": 0.2013,
      "step": 1440
    },
    {
      "epoch": 15.0,
      "eval_accuracy": 0.6181818181818182,
      "eval_loss": 2.344068765640259,
      "eval_runtime": 3.8014,
      "eval_samples_per_second": 144.683,
      "eval_steps_per_second": 9.207,
      "step": 1440
    },
    {
      "epoch": 15.1,
      "learning_rate": 3.45775462962963e-05,
      "loss": 0.2093,
      "step": 1450
    },
    {
      "epoch": 15.21,
      "learning_rate": 3.443287037037037e-05,
      "loss": 0.2092,
      "step": 1460
    },
    {
      "epoch": 15.31,
      "learning_rate": 3.428819444444444e-05,
      "loss": 0.1995,
      "step": 1470
    },
    {
      "epoch": 15.42,
      "learning_rate": 3.414351851851852e-05,
      "loss": 0.2162,
      "step": 1480
    },
    {
      "epoch": 15.52,
      "learning_rate": 3.39988425925926e-05,
      "loss": 0.2444,
      "step": 1490
    },
    {
      "epoch": 15.62,
      "learning_rate": 3.385416666666667e-05,
      "loss": 0.1702,
      "step": 1500
    },
    {
      "epoch": 15.73,
      "learning_rate": 3.370949074074074e-05,
      "loss": 0.28,
      "step": 1510
    },
    {
      "epoch": 15.83,
      "learning_rate": 3.3564814814814815e-05,
      "loss": 0.2303,
      "step": 1520
    },
    {
      "epoch": 15.94,
      "learning_rate": 3.342013888888889e-05,
      "loss": 0.2225,
      "step": 1530
    },
    {
      "epoch": 16.0,
      "eval_accuracy": 0.5709090909090909,
      "eval_loss": 3.0213537216186523,
      "eval_runtime": 3.7244,
      "eval_samples_per_second": 147.674,
      "eval_steps_per_second": 9.397,
      "step": 1536
    },
    {
      "epoch": 16.04,
      "learning_rate": 3.3275462962962965e-05,
      "loss": 0.2012,
      "step": 1540
    },
    {
      "epoch": 16.15,
      "learning_rate": 3.313078703703704e-05,
      "loss": 0.1761,
      "step": 1550
    },
    {
      "epoch": 16.25,
      "learning_rate": 3.2986111111111115e-05,
      "loss": 0.229,
      "step": 1560
    },
    {
      "epoch": 16.35,
      "learning_rate": 3.284143518518519e-05,
      "loss": 0.2114,
      "step": 1570
    },
    {
      "epoch": 16.46,
      "learning_rate": 3.2696759259259265e-05,
      "loss": 0.2284,
      "step": 1580
    },
    {
      "epoch": 16.56,
      "learning_rate": 3.255208333333333e-05,
      "loss": 0.2137,
      "step": 1590
    },
    {
      "epoch": 16.67,
      "learning_rate": 3.240740740740741e-05,
      "loss": 0.2902,
      "step": 1600
    },
    {
      "epoch": 16.77,
      "learning_rate": 3.226273148148148e-05,
      "loss": 0.2273,
      "step": 1610
    },
    {
      "epoch": 16.88,
      "learning_rate": 3.211805555555556e-05,
      "loss": 0.182,
      "step": 1620
    },
    {
      "epoch": 16.98,
      "learning_rate": 3.197337962962963e-05,
      "loss": 0.2229,
      "step": 1630
    },
    {
      "epoch": 17.0,
      "eval_accuracy": 0.6163636363636363,
      "eval_loss": 2.067636728286743,
      "eval_runtime": 3.8554,
      "eval_samples_per_second": 142.656,
      "eval_steps_per_second": 9.078,
      "step": 1632
    },
    {
      "epoch": 17.08,
      "learning_rate": 3.182870370370371e-05,
      "loss": 0.1704,
      "step": 1640
    },
    {
      "epoch": 17.19,
      "learning_rate": 3.168402777777778e-05,
      "loss": 0.2184,
      "step": 1650
    },
    {
      "epoch": 17.29,
      "learning_rate": 3.153935185185186e-05,
      "loss": 0.1662,
      "step": 1660
    },
    {
      "epoch": 17.4,
      "learning_rate": 3.1394675925925926e-05,
      "loss": 0.1965,
      "step": 1670
    },
    {
      "epoch": 17.5,
      "learning_rate": 3.125e-05,
      "loss": 0.1384,
      "step": 1680
    },
    {
      "epoch": 17.6,
      "learning_rate": 3.1105324074074076e-05,
      "loss": 0.213,
      "step": 1690
    },
    {
      "epoch": 17.71,
      "learning_rate": 3.0960648148148144e-05,
      "loss": 0.2053,
      "step": 1700
    },
    {
      "epoch": 17.81,
      "learning_rate": 3.0815972222222225e-05,
      "loss": 0.1938,
      "step": 1710
    },
    {
      "epoch": 17.92,
      "learning_rate": 3.06712962962963e-05,
      "loss": 0.2024,
      "step": 1720
    },
    {
      "epoch": 18.0,
      "eval_accuracy": 0.5672727272727273,
      "eval_loss": 2.6478159427642822,
      "eval_runtime": 3.8359,
      "eval_samples_per_second": 143.382,
      "eval_steps_per_second": 9.124,
      "step": 1728
    },
    {
      "epoch": 18.02,
      "learning_rate": 3.0526620370370375e-05,
      "loss": 0.1643,
      "step": 1730
    },
    {
      "epoch": 18.12,
      "learning_rate": 3.0381944444444444e-05,
      "loss": 0.2376,
      "step": 1740
    },
    {
      "epoch": 18.23,
      "learning_rate": 3.023726851851852e-05,
      "loss": 0.2088,
      "step": 1750
    },
    {
      "epoch": 18.33,
      "learning_rate": 3.0092592592592593e-05,
      "loss": 0.181,
      "step": 1760
    },
    {
      "epoch": 18.44,
      "learning_rate": 2.994791666666667e-05,
      "loss": 0.19,
      "step": 1770
    },
    {
      "epoch": 18.54,
      "learning_rate": 2.980324074074074e-05,
      "loss": 0.2008,
      "step": 1780
    },
    {
      "epoch": 18.65,
      "learning_rate": 2.9658564814814815e-05,
      "loss": 0.2109,
      "step": 1790
    },
    {
      "epoch": 18.75,
      "learning_rate": 2.951388888888889e-05,
      "loss": 0.1904,
      "step": 1800
    },
    {
      "epoch": 18.85,
      "learning_rate": 2.9369212962962965e-05,
      "loss": 0.2136,
      "step": 1810
    },
    {
      "epoch": 18.96,
      "learning_rate": 2.9224537037037036e-05,
      "loss": 0.1401,
      "step": 1820
    },
    {
      "epoch": 19.0,
      "eval_accuracy": 0.5636363636363636,
      "eval_loss": 2.8951961994171143,
      "eval_runtime": 3.8754,
      "eval_samples_per_second": 141.92,
      "eval_steps_per_second": 9.031,
      "step": 1824
    },
    {
      "epoch": 19.06,
      "learning_rate": 2.907986111111111e-05,
      "loss": 0.1974,
      "step": 1830
    },
    {
      "epoch": 19.17,
      "learning_rate": 2.8935185185185186e-05,
      "loss": 0.1752,
      "step": 1840
    },
    {
      "epoch": 19.27,
      "learning_rate": 2.879050925925926e-05,
      "loss": 0.1814,
      "step": 1850
    },
    {
      "epoch": 19.38,
      "learning_rate": 2.8645833333333333e-05,
      "loss": 0.1853,
      "step": 1860
    },
    {
      "epoch": 19.48,
      "learning_rate": 2.8501157407407408e-05,
      "loss": 0.1992,
      "step": 1870
    },
    {
      "epoch": 19.58,
      "learning_rate": 2.8356481481481483e-05,
      "loss": 0.1774,
      "step": 1880
    },
    {
      "epoch": 19.69,
      "learning_rate": 2.821180555555556e-05,
      "loss": 0.1836,
      "step": 1890
    },
    {
      "epoch": 19.79,
      "learning_rate": 2.806712962962963e-05,
      "loss": 0.1556,
      "step": 1900
    },
    {
      "epoch": 19.9,
      "learning_rate": 2.7922453703703704e-05,
      "loss": 0.175,
      "step": 1910
    },
    {
      "epoch": 20.0,
      "learning_rate": 2.777777777777778e-05,
      "loss": 0.1984,
      "step": 1920
    },
    {
      "epoch": 20.0,
      "eval_accuracy": 0.6145454545454545,
      "eval_loss": 2.308309555053711,
      "eval_runtime": 3.9255,
      "eval_samples_per_second": 140.111,
      "eval_steps_per_second": 8.916,
      "step": 1920
    },
    {
      "epoch": 20.1,
      "learning_rate": 2.7633101851851857e-05,
      "loss": 0.1598,
      "step": 1930
    },
    {
      "epoch": 20.21,
      "learning_rate": 2.7488425925925926e-05,
      "loss": 0.1761,
      "step": 1940
    },
    {
      "epoch": 20.31,
      "learning_rate": 2.734375e-05,
      "loss": 0.1647,
      "step": 1950
    },
    {
      "epoch": 20.42,
      "learning_rate": 2.7199074074074076e-05,
      "loss": 0.1553,
      "step": 1960
    },
    {
      "epoch": 20.52,
      "learning_rate": 2.7054398148148147e-05,
      "loss": 0.1591,
      "step": 1970
    },
    {
      "epoch": 20.62,
      "learning_rate": 2.6909722222222222e-05,
      "loss": 0.1736,
      "step": 1980
    },
    {
      "epoch": 20.73,
      "learning_rate": 2.6765046296296297e-05,
      "loss": 0.1892,
      "step": 1990
    },
    {
      "epoch": 20.83,
      "learning_rate": 2.6620370370370372e-05,
      "loss": 0.1596,
      "step": 2000
    },
    {
      "epoch": 20.94,
      "learning_rate": 2.6475694444444443e-05,
      "loss": 0.1788,
      "step": 2010
    },
    {
      "epoch": 21.0,
      "eval_accuracy": 0.52,
      "eval_loss": 3.7701735496520996,
      "eval_runtime": 3.9464,
      "eval_samples_per_second": 139.366,
      "eval_steps_per_second": 8.869,
      "step": 2016
    },
    {
      "epoch": 21.04,
      "learning_rate": 2.633101851851852e-05,
      "loss": 0.1783,
      "step": 2020
    },
    {
      "epoch": 21.15,
      "learning_rate": 2.6186342592592593e-05,
      "loss": 0.1855,
      "step": 2030
    },
    {
      "epoch": 21.25,
      "learning_rate": 2.604166666666667e-05,
      "loss": 0.2003,
      "step": 2040
    },
    {
      "epoch": 21.35,
      "learning_rate": 2.589699074074074e-05,
      "loss": 0.1726,
      "step": 2050
    },
    {
      "epoch": 21.46,
      "learning_rate": 2.5752314814814815e-05,
      "loss": 0.153,
      "step": 2060
    },
    {
      "epoch": 21.56,
      "learning_rate": 2.560763888888889e-05,
      "loss": 0.1474,
      "step": 2070
    },
    {
      "epoch": 21.67,
      "learning_rate": 2.5462962962962965e-05,
      "loss": 0.1354,
      "step": 2080
    },
    {
      "epoch": 21.77,
      "learning_rate": 2.5318287037037036e-05,
      "loss": 0.1553,
      "step": 2090
    },
    {
      "epoch": 21.88,
      "learning_rate": 2.517361111111111e-05,
      "loss": 0.1832,
      "step": 2100
    },
    {
      "epoch": 21.98,
      "learning_rate": 2.5028935185185186e-05,
      "loss": 0.1907,
      "step": 2110
    },
    {
      "epoch": 22.0,
      "eval_accuracy": 0.6654545454545454,
      "eval_loss": 1.9616814851760864,
      "eval_runtime": 3.8632,
      "eval_samples_per_second": 142.368,
      "eval_steps_per_second": 9.06,
      "step": 2112
    },
    {
      "epoch": 22.08,
      "learning_rate": 2.488425925925926e-05,
      "loss": 0.159,
      "step": 2120
    },
    {
      "epoch": 22.19,
      "learning_rate": 2.4739583333333336e-05,
      "loss": 0.1397,
      "step": 2130
    },
    {
      "epoch": 22.29,
      "learning_rate": 2.4594907407407408e-05,
      "loss": 0.1511,
      "step": 2140
    },
    {
      "epoch": 22.4,
      "learning_rate": 2.4450231481481483e-05,
      "loss": 0.1704,
      "step": 2150
    },
    {
      "epoch": 22.5,
      "learning_rate": 2.4305555555555558e-05,
      "loss": 0.1475,
      "step": 2160
    },
    {
      "epoch": 22.6,
      "learning_rate": 2.4160879629629633e-05,
      "loss": 0.1568,
      "step": 2170
    },
    {
      "epoch": 22.71,
      "learning_rate": 2.4016203703703704e-05,
      "loss": 0.1394,
      "step": 2180
    },
    {
      "epoch": 22.81,
      "learning_rate": 2.387152777777778e-05,
      "loss": 0.1274,
      "step": 2190
    },
    {
      "epoch": 22.92,
      "learning_rate": 2.3726851851851854e-05,
      "loss": 0.1113,
      "step": 2200
    },
    {
      "epoch": 23.0,
      "eval_accuracy": 0.5963636363636363,
      "eval_loss": 2.654604434967041,
      "eval_runtime": 3.8399,
      "eval_samples_per_second": 143.233,
      "eval_steps_per_second": 9.115,
      "step": 2208
    },
    {
      "epoch": 23.02,
      "learning_rate": 2.358217592592593e-05,
      "loss": 0.1994,
      "step": 2210
    },
    {
      "epoch": 23.12,
      "learning_rate": 2.34375e-05,
      "loss": 0.1513,
      "step": 2220
    },
    {
      "epoch": 23.23,
      "learning_rate": 2.3292824074074075e-05,
      "loss": 0.1596,
      "step": 2230
    },
    {
      "epoch": 23.33,
      "learning_rate": 2.314814814814815e-05,
      "loss": 0.1276,
      "step": 2240
    },
    {
      "epoch": 23.44,
      "learning_rate": 2.3003472222222222e-05,
      "loss": 0.1552,
      "step": 2250
    },
    {
      "epoch": 23.54,
      "learning_rate": 2.2858796296296297e-05,
      "loss": 0.1419,
      "step": 2260
    },
    {
      "epoch": 23.65,
      "learning_rate": 2.2714120370370372e-05,
      "loss": 0.1798,
      "step": 2270
    },
    {
      "epoch": 23.75,
      "learning_rate": 2.2569444444444447e-05,
      "loss": 0.1832,
      "step": 2280
    },
    {
      "epoch": 23.85,
      "learning_rate": 2.242476851851852e-05,
      "loss": 0.1656,
      "step": 2290
    },
    {
      "epoch": 23.96,
      "learning_rate": 2.2280092592592593e-05,
      "loss": 0.1293,
      "step": 2300
    },
    {
      "epoch": 24.0,
      "eval_accuracy": 0.6036363636363636,
      "eval_loss": 2.6427228450775146,
      "eval_runtime": 3.9185,
      "eval_samples_per_second": 140.361,
      "eval_steps_per_second": 8.932,
      "step": 2304
    },
    {
      "epoch": 24.06,
      "learning_rate": 2.2135416666666668e-05,
      "loss": 0.1278,
      "step": 2310
    },
    {
      "epoch": 24.17,
      "learning_rate": 2.1990740740740743e-05,
      "loss": 0.1154,
      "step": 2320
    },
    {
      "epoch": 24.27,
      "learning_rate": 2.1846064814814815e-05,
      "loss": 0.1542,
      "step": 2330
    },
    {
      "epoch": 24.38,
      "learning_rate": 2.170138888888889e-05,
      "loss": 0.2238,
      "step": 2340
    },
    {
      "epoch": 24.48,
      "learning_rate": 2.1556712962962965e-05,
      "loss": 0.16,
      "step": 2350
    },
    {
      "epoch": 24.58,
      "learning_rate": 2.141203703703704e-05,
      "loss": 0.1713,
      "step": 2360
    },
    {
      "epoch": 24.69,
      "learning_rate": 2.126736111111111e-05,
      "loss": 0.1392,
      "step": 2370
    },
    {
      "epoch": 24.79,
      "learning_rate": 2.1122685185185186e-05,
      "loss": 0.1586,
      "step": 2380
    },
    {
      "epoch": 24.9,
      "learning_rate": 2.097800925925926e-05,
      "loss": 0.0984,
      "step": 2390
    },
    {
      "epoch": 25.0,
      "learning_rate": 2.0833333333333336e-05,
      "loss": 0.1354,
      "step": 2400
    },
    {
      "epoch": 25.0,
      "eval_accuracy": 0.5527272727272727,
      "eval_loss": 3.410457134246826,
      "eval_runtime": 3.9002,
      "eval_samples_per_second": 141.019,
      "eval_steps_per_second": 8.974,
      "step": 2400
    },
    {
      "epoch": 25.1,
      "learning_rate": 2.0688657407407408e-05,
      "loss": 0.1183,
      "step": 2410
    },
    {
      "epoch": 25.21,
      "learning_rate": 2.0543981481481483e-05,
      "loss": 0.1221,
      "step": 2420
    },
    {
      "epoch": 25.31,
      "learning_rate": 2.0399305555555557e-05,
      "loss": 0.1588,
      "step": 2430
    },
    {
      "epoch": 25.42,
      "learning_rate": 2.0254629629629632e-05,
      "loss": 0.1581,
      "step": 2440
    },
    {
      "epoch": 25.52,
      "learning_rate": 2.0109953703703704e-05,
      "loss": 0.1372,
      "step": 2450
    },
    {
      "epoch": 25.62,
      "learning_rate": 1.996527777777778e-05,
      "loss": 0.1165,
      "step": 2460
    },
    {
      "epoch": 25.73,
      "learning_rate": 1.9820601851851854e-05,
      "loss": 0.124,
      "step": 2470
    },
    {
      "epoch": 25.83,
      "learning_rate": 1.967592592592593e-05,
      "loss": 0.1127,
      "step": 2480
    },
    {
      "epoch": 25.94,
      "learning_rate": 1.953125e-05,
      "loss": 0.1447,
      "step": 2490
    },
    {
      "epoch": 26.0,
      "eval_accuracy": 0.6127272727272727,
      "eval_loss": 2.545973062515259,
      "eval_runtime": 4.0556,
      "eval_samples_per_second": 135.614,
      "eval_steps_per_second": 8.63,
      "step": 2496
    },
    {
      "epoch": 26.04,
      "learning_rate": 1.9386574074074075e-05,
      "loss": 0.1788,
      "step": 2500
    },
    {
      "epoch": 26.15,
      "learning_rate": 1.924189814814815e-05,
      "loss": 0.1482,
      "step": 2510
    },
    {
      "epoch": 26.25,
      "learning_rate": 1.9097222222222222e-05,
      "loss": 0.126,
      "step": 2520
    },
    {
      "epoch": 26.35,
      "learning_rate": 1.8952546296296297e-05,
      "loss": 0.1427,
      "step": 2530
    },
    {
      "epoch": 26.46,
      "learning_rate": 1.8807870370370372e-05,
      "loss": 0.1603,
      "step": 2540
    },
    {
      "epoch": 26.56,
      "learning_rate": 1.8663194444444447e-05,
      "loss": 0.1391,
      "step": 2550
    },
    {
      "epoch": 26.67,
      "learning_rate": 1.8518518518518518e-05,
      "loss": 0.1036,
      "step": 2560
    },
    {
      "epoch": 26.77,
      "learning_rate": 1.8373842592592593e-05,
      "loss": 0.1336,
      "step": 2570
    },
    {
      "epoch": 26.88,
      "learning_rate": 1.8229166666666668e-05,
      "loss": 0.1348,
      "step": 2580
    },
    {
      "epoch": 26.98,
      "learning_rate": 1.8084490740740743e-05,
      "loss": 0.0995,
      "step": 2590
    },
    {
      "epoch": 27.0,
      "eval_accuracy": 0.5854545454545454,
      "eval_loss": 2.9865081310272217,
      "eval_runtime": 3.9958,
      "eval_samples_per_second": 137.646,
      "eval_steps_per_second": 8.759,
      "step": 2592
    },
    {
      "epoch": 27.08,
      "learning_rate": 1.7939814814814815e-05,
      "loss": 0.1093,
      "step": 2600
    },
    {
      "epoch": 27.19,
      "learning_rate": 1.779513888888889e-05,
      "loss": 0.1454,
      "step": 2610
    },
    {
      "epoch": 27.29,
      "learning_rate": 1.7650462962962965e-05,
      "loss": 0.1408,
      "step": 2620
    },
    {
      "epoch": 27.4,
      "learning_rate": 1.750578703703704e-05,
      "loss": 0.1415,
      "step": 2630
    },
    {
      "epoch": 27.5,
      "learning_rate": 1.736111111111111e-05,
      "loss": 0.1032,
      "step": 2640
    },
    {
      "epoch": 27.6,
      "learning_rate": 1.7216435185185186e-05,
      "loss": 0.1551,
      "step": 2650
    },
    {
      "epoch": 27.71,
      "learning_rate": 1.707175925925926e-05,
      "loss": 0.084,
      "step": 2660
    },
    {
      "epoch": 27.81,
      "learning_rate": 1.6927083333333336e-05,
      "loss": 0.1088,
      "step": 2670
    },
    {
      "epoch": 27.92,
      "learning_rate": 1.6782407407407408e-05,
      "loss": 0.1369,
      "step": 2680
    },
    {
      "epoch": 28.0,
      "eval_accuracy": 0.5545454545454546,
      "eval_loss": 3.5280685424804688,
      "eval_runtime": 3.88,
      "eval_samples_per_second": 141.752,
      "eval_steps_per_second": 9.021,
      "step": 2688
    },
    {
      "epoch": 28.02,
      "learning_rate": 1.6637731481481482e-05,
      "loss": 0.14,
      "step": 2690
    },
    {
      "epoch": 28.12,
      "learning_rate": 1.6493055555555557e-05,
      "loss": 0.1198,
      "step": 2700
    },
    {
      "epoch": 28.23,
      "learning_rate": 1.6348379629629632e-05,
      "loss": 0.1132,
      "step": 2710
    },
    {
      "epoch": 28.33,
      "learning_rate": 1.6203703703703704e-05,
      "loss": 0.1526,
      "step": 2720
    },
    {
      "epoch": 28.44,
      "learning_rate": 1.605902777777778e-05,
      "loss": 0.1114,
      "step": 2730
    },
    {
      "epoch": 28.54,
      "learning_rate": 1.5914351851851854e-05,
      "loss": 0.1157,
      "step": 2740
    },
    {
      "epoch": 28.65,
      "learning_rate": 1.576967592592593e-05,
      "loss": 0.1322,
      "step": 2750
    },
    {
      "epoch": 28.75,
      "learning_rate": 1.5625e-05,
      "loss": 0.0794,
      "step": 2760
    },
    {
      "epoch": 28.85,
      "learning_rate": 1.5480324074074072e-05,
      "loss": 0.1354,
      "step": 2770
    },
    {
      "epoch": 28.96,
      "learning_rate": 1.533564814814815e-05,
      "loss": 0.1238,
      "step": 2780
    },
    {
      "epoch": 29.0,
      "eval_accuracy": 0.6018181818181818,
      "eval_loss": 2.816072702407837,
      "eval_runtime": 3.8975,
      "eval_samples_per_second": 141.115,
      "eval_steps_per_second": 8.98,
      "step": 2784
    },
    {
      "epoch": 29.06,
      "learning_rate": 1.5190972222222222e-05,
      "loss": 0.1339,
      "step": 2790
    },
    {
      "epoch": 29.17,
      "learning_rate": 1.5046296296296297e-05,
      "loss": 0.098,
      "step": 2800
    },
    {
      "epoch": 29.27,
      "learning_rate": 1.490162037037037e-05,
      "loss": 0.1411,
      "step": 2810
    },
    {
      "epoch": 29.38,
      "learning_rate": 1.4756944444444445e-05,
      "loss": 0.0686,
      "step": 2820
    },
    {
      "epoch": 29.48,
      "learning_rate": 1.4612268518518518e-05,
      "loss": 0.1029,
      "step": 2830
    },
    {
      "epoch": 29.58,
      "learning_rate": 1.4467592592592593e-05,
      "loss": 0.1361,
      "step": 2840
    },
    {
      "epoch": 29.69,
      "learning_rate": 1.4322916666666666e-05,
      "loss": 0.0927,
      "step": 2850
    },
    {
      "epoch": 29.79,
      "learning_rate": 1.4178240740740741e-05,
      "loss": 0.1329,
      "step": 2860
    },
    {
      "epoch": 29.9,
      "learning_rate": 1.4033564814814815e-05,
      "loss": 0.1146,
      "step": 2870
    },
    {
      "epoch": 30.0,
      "learning_rate": 1.388888888888889e-05,
      "loss": 0.1256,
      "step": 2880
    },
    {
      "epoch": 30.0,
      "eval_accuracy": 0.5490909090909091,
      "eval_loss": 3.491748332977295,
      "eval_runtime": 3.8988,
      "eval_samples_per_second": 141.071,
      "eval_steps_per_second": 8.977,
      "step": 2880
    },
    {
      "epoch": 30.1,
      "learning_rate": 1.3744212962962963e-05,
      "loss": 0.1046,
      "step": 2890
    },
    {
      "epoch": 30.21,
      "learning_rate": 1.3599537037037038e-05,
      "loss": 0.1199,
      "step": 2900
    },
    {
      "epoch": 30.31,
      "learning_rate": 1.3454861111111111e-05,
      "loss": 0.1524,
      "step": 2910
    },
    {
      "epoch": 30.42,
      "learning_rate": 1.3310185185185186e-05,
      "loss": 0.1064,
      "step": 2920
    },
    {
      "epoch": 30.52,
      "learning_rate": 1.316550925925926e-05,
      "loss": 0.1227,
      "step": 2930
    },
    {
      "epoch": 30.62,
      "learning_rate": 1.3020833333333334e-05,
      "loss": 0.1043,
      "step": 2940
    },
    {
      "epoch": 30.73,
      "learning_rate": 1.2876157407407407e-05,
      "loss": 0.1002,
      "step": 2950
    },
    {
      "epoch": 30.83,
      "learning_rate": 1.2731481481481482e-05,
      "loss": 0.1153,
      "step": 2960
    },
    {
      "epoch": 30.94,
      "learning_rate": 1.2586805555555556e-05,
      "loss": 0.1064,
      "step": 2970
    },
    {
      "epoch": 31.0,
      "eval_accuracy": 0.58,
      "eval_loss": 3.065882682800293,
      "eval_runtime": 3.8866,
      "eval_samples_per_second": 141.514,
      "eval_steps_per_second": 9.005,
      "step": 2976
    },
    {
      "epoch": 31.04,
      "learning_rate": 1.244212962962963e-05,
      "loss": 0.1036,
      "step": 2980
    },
    {
      "epoch": 31.15,
      "learning_rate": 1.2297453703703704e-05,
      "loss": 0.1149,
      "step": 2990
    },
    {
      "epoch": 31.25,
      "learning_rate": 1.2152777777777779e-05,
      "loss": 0.1135,
      "step": 3000
    },
    {
      "epoch": 31.35,
      "learning_rate": 1.2008101851851852e-05,
      "loss": 0.0999,
      "step": 3010
    },
    {
      "epoch": 31.46,
      "learning_rate": 1.1863425925925927e-05,
      "loss": 0.1174,
      "step": 3020
    },
    {
      "epoch": 31.56,
      "learning_rate": 1.171875e-05,
      "loss": 0.1161,
      "step": 3030
    },
    {
      "epoch": 31.67,
      "learning_rate": 1.1574074074074075e-05,
      "loss": 0.1123,
      "step": 3040
    },
    {
      "epoch": 31.77,
      "learning_rate": 1.1429398148148148e-05,
      "loss": 0.1167,
      "step": 3050
    },
    {
      "epoch": 31.88,
      "learning_rate": 1.1284722222222223e-05,
      "loss": 0.0939,
      "step": 3060
    },
    {
      "epoch": 31.98,
      "learning_rate": 1.1140046296296297e-05,
      "loss": 0.1333,
      "step": 3070
    },
    {
      "epoch": 32.0,
      "eval_accuracy": 0.5472727272727272,
      "eval_loss": 3.5971763134002686,
      "eval_runtime": 3.9815,
      "eval_samples_per_second": 138.137,
      "eval_steps_per_second": 8.791,
      "step": 3072
    },
    {
      "epoch": 32.08,
      "learning_rate": 1.0995370370370372e-05,
      "loss": 0.1048,
      "step": 3080
    },
    {
      "epoch": 32.19,
      "learning_rate": 1.0850694444444445e-05,
      "loss": 0.118,
      "step": 3090
    },
    {
      "epoch": 32.29,
      "learning_rate": 1.070601851851852e-05,
      "loss": 0.1366,
      "step": 3100
    },
    {
      "epoch": 32.4,
      "learning_rate": 1.0561342592592593e-05,
      "loss": 0.0787,
      "step": 3110
    },
    {
      "epoch": 32.5,
      "learning_rate": 1.0416666666666668e-05,
      "loss": 0.1368,
      "step": 3120
    },
    {
      "epoch": 32.6,
      "learning_rate": 1.0271990740740741e-05,
      "loss": 0.1137,
      "step": 3130
    },
    {
      "epoch": 32.71,
      "learning_rate": 1.0127314814814816e-05,
      "loss": 0.1161,
      "step": 3140
    },
    {
      "epoch": 32.81,
      "learning_rate": 9.98263888888889e-06,
      "loss": 0.1096,
      "step": 3150
    },
    {
      "epoch": 32.92,
      "learning_rate": 9.837962962962964e-06,
      "loss": 0.1134,
      "step": 3160
    },
    {
      "epoch": 33.0,
      "eval_accuracy": 0.54,
      "eval_loss": 3.611628293991089,
      "eval_runtime": 3.8603,
      "eval_samples_per_second": 142.477,
      "eval_steps_per_second": 9.067,
      "step": 3168
    },
    {
      "epoch": 33.02,
      "learning_rate": 9.693287037037038e-06,
      "loss": 0.0905,
      "step": 3170
    },
    {
      "epoch": 33.12,
      "learning_rate": 9.548611111111111e-06,
      "loss": 0.1007,
      "step": 3180
    },
    {
      "epoch": 33.23,
      "learning_rate": 9.403935185185186e-06,
      "loss": 0.1342,
      "step": 3190
    },
    {
      "epoch": 33.33,
      "learning_rate": 9.259259259259259e-06,
      "loss": 0.1072,
      "step": 3200
    },
    {
      "epoch": 33.44,
      "learning_rate": 9.114583333333334e-06,
      "loss": 0.0824,
      "step": 3210
    },
    {
      "epoch": 33.54,
      "learning_rate": 8.969907407407407e-06,
      "loss": 0.123,
      "step": 3220
    },
    {
      "epoch": 33.65,
      "learning_rate": 8.825231481481482e-06,
      "loss": 0.1083,
      "step": 3230
    },
    {
      "epoch": 33.75,
      "learning_rate": 8.680555555555556e-06,
      "loss": 0.1073,
      "step": 3240
    },
    {
      "epoch": 33.85,
      "learning_rate": 8.53587962962963e-06,
      "loss": 0.0854,
      "step": 3250
    },
    {
      "epoch": 33.96,
      "learning_rate": 8.391203703703704e-06,
      "loss": 0.0831,
      "step": 3260
    },
    {
      "epoch": 34.0,
      "eval_accuracy": 0.5509090909090909,
      "eval_loss": 3.530831813812256,
      "eval_runtime": 3.9226,
      "eval_samples_per_second": 140.214,
      "eval_steps_per_second": 8.923,
      "step": 3264
    },
    {
      "epoch": 34.06,
      "learning_rate": 8.246527777777779e-06,
      "loss": 0.1373,
      "step": 3270
    },
    {
      "epoch": 34.17,
      "learning_rate": 8.101851851851852e-06,
      "loss": 0.0997,
      "step": 3280
    },
    {
      "epoch": 34.27,
      "learning_rate": 7.957175925925927e-06,
      "loss": 0.0959,
      "step": 3290
    },
    {
      "epoch": 34.38,
      "learning_rate": 7.8125e-06,
      "loss": 0.0999,
      "step": 3300
    },
    {
      "epoch": 34.48,
      "learning_rate": 7.667824074074075e-06,
      "loss": 0.0585,
      "step": 3310
    },
    {
      "epoch": 34.58,
      "learning_rate": 7.523148148148148e-06,
      "loss": 0.0815,
      "step": 3320
    },
    {
      "epoch": 34.69,
      "learning_rate": 7.3784722222222225e-06,
      "loss": 0.073,
      "step": 3330
    },
    {
      "epoch": 34.79,
      "learning_rate": 7.2337962962962966e-06,
      "loss": 0.1152,
      "step": 3340
    },
    {
      "epoch": 34.9,
      "learning_rate": 7.089120370370371e-06,
      "loss": 0.0926,
      "step": 3350
    },
    {
      "epoch": 35.0,
      "learning_rate": 6.944444444444445e-06,
      "loss": 0.1035,
      "step": 3360
    },
    {
      "epoch": 35.0,
      "eval_accuracy": 0.5581818181818182,
      "eval_loss": 3.4789416790008545,
      "eval_runtime": 3.9255,
      "eval_samples_per_second": 140.108,
      "eval_steps_per_second": 8.916,
      "step": 3360
    },
    {
      "epoch": 35.1,
      "learning_rate": 6.799768518518519e-06,
      "loss": 0.1094,
      "step": 3370
    },
    {
      "epoch": 35.21,
      "learning_rate": 6.655092592592593e-06,
      "loss": 0.1014,
      "step": 3380
    },
    {
      "epoch": 35.31,
      "learning_rate": 6.510416666666667e-06,
      "loss": 0.0975,
      "step": 3390
    },
    {
      "epoch": 35.42,
      "learning_rate": 6.365740740740741e-06,
      "loss": 0.0769,
      "step": 3400
    },
    {
      "epoch": 35.52,
      "learning_rate": 6.221064814814815e-06,
      "loss": 0.0742,
      "step": 3410
    },
    {
      "epoch": 35.62,
      "learning_rate": 6.076388888888889e-06,
      "loss": 0.0994,
      "step": 3420
    },
    {
      "epoch": 35.73,
      "learning_rate": 5.9317129629629635e-06,
      "loss": 0.0755,
      "step": 3430
    },
    {
      "epoch": 35.83,
      "learning_rate": 5.787037037037038e-06,
      "loss": 0.0833,
      "step": 3440
    },
    {
      "epoch": 35.94,
      "learning_rate": 5.642361111111112e-06,
      "loss": 0.0957,
      "step": 3450
    },
    {
      "epoch": 36.0,
      "eval_accuracy": 0.5509090909090909,
      "eval_loss": 3.6358087062835693,
      "eval_runtime": 3.9255,
      "eval_samples_per_second": 140.108,
      "eval_steps_per_second": 8.916,
      "step": 3456
    },
    {
      "epoch": 36.04,
      "learning_rate": 5.497685185185186e-06,
      "loss": 0.1083,
      "step": 3460
    },
    {
      "epoch": 36.15,
      "learning_rate": 5.35300925925926e-06,
      "loss": 0.0927,
      "step": 3470
    },
    {
      "epoch": 36.25,
      "learning_rate": 5.208333333333334e-06,
      "loss": 0.0876,
      "step": 3480
    },
    {
      "epoch": 36.35,
      "learning_rate": 5.063657407407408e-06,
      "loss": 0.0921,
      "step": 3490
    },
    {
      "epoch": 36.46,
      "learning_rate": 4.918981481481482e-06,
      "loss": 0.0909,
      "step": 3500
    },
    {
      "epoch": 36.56,
      "learning_rate": 4.7743055555555555e-06,
      "loss": 0.0927,
      "step": 3510
    },
    {
      "epoch": 36.67,
      "learning_rate": 4.6296296296296296e-06,
      "loss": 0.1031,
      "step": 3520
    },
    {
      "epoch": 36.77,
      "learning_rate": 4.484953703703704e-06,
      "loss": 0.1017,
      "step": 3530
    },
    {
      "epoch": 36.88,
      "learning_rate": 4.340277777777778e-06,
      "loss": 0.064,
      "step": 3540
    },
    {
      "epoch": 36.98,
      "learning_rate": 4.195601851851852e-06,
      "loss": 0.0764,
      "step": 3550
    },
    {
      "epoch": 37.0,
      "eval_accuracy": 0.5709090909090909,
      "eval_loss": 3.363851547241211,
      "eval_runtime": 3.9482,
      "eval_samples_per_second": 139.304,
      "eval_steps_per_second": 8.865,
      "step": 3552
    },
    {
      "epoch": 37.08,
      "learning_rate": 4.050925925925926e-06,
      "loss": 0.1104,
      "step": 3560
    },
    {
      "epoch": 37.19,
      "learning_rate": 3.90625e-06,
      "loss": 0.0785,
      "step": 3570
    },
    {
      "epoch": 37.29,
      "learning_rate": 3.761574074074074e-06,
      "loss": 0.0777,
      "step": 3580
    },
    {
      "epoch": 37.4,
      "learning_rate": 3.6168981481481483e-06,
      "loss": 0.0696,
      "step": 3590
    },
    {
      "epoch": 37.5,
      "learning_rate": 3.4722222222222224e-06,
      "loss": 0.0789,
      "step": 3600
    },
    {
      "epoch": 37.6,
      "learning_rate": 3.3275462962962965e-06,
      "loss": 0.0921,
      "step": 3610
    },
    {
      "epoch": 37.71,
      "learning_rate": 3.1828703703703706e-06,
      "loss": 0.0932,
      "step": 3620
    },
    {
      "epoch": 37.81,
      "learning_rate": 3.0381944444444447e-06,
      "loss": 0.0614,
      "step": 3630
    },
    {
      "epoch": 37.92,
      "learning_rate": 2.893518518518519e-06,
      "loss": 0.072,
      "step": 3640
    },
    {
      "epoch": 38.0,
      "eval_accuracy": 0.5563636363636364,
      "eval_loss": 3.5639150142669678,
      "eval_runtime": 3.9387,
      "eval_samples_per_second": 139.642,
      "eval_steps_per_second": 8.886,
      "step": 3648
    },
    {
      "epoch": 38.02,
      "learning_rate": 2.748842592592593e-06,
      "loss": 0.1215,
      "step": 3650
    },
    {
      "epoch": 38.12,
      "learning_rate": 2.604166666666667e-06,
      "loss": 0.0942,
      "step": 3660
    },
    {
      "epoch": 38.23,
      "learning_rate": 2.459490740740741e-06,
      "loss": 0.0525,
      "step": 3670
    },
    {
      "epoch": 38.33,
      "learning_rate": 2.3148148148148148e-06,
      "loss": 0.0871,
      "step": 3680
    },
    {
      "epoch": 38.44,
      "learning_rate": 2.170138888888889e-06,
      "loss": 0.0973,
      "step": 3690
    },
    {
      "epoch": 38.54,
      "learning_rate": 2.025462962962963e-06,
      "loss": 0.1067,
      "step": 3700
    },
    {
      "epoch": 38.65,
      "learning_rate": 1.880787037037037e-06,
      "loss": 0.1219,
      "step": 3710
    },
    {
      "epoch": 38.75,
      "learning_rate": 1.7361111111111112e-06,
      "loss": 0.0806,
      "step": 3720
    },
    {
      "epoch": 38.85,
      "learning_rate": 1.5914351851851853e-06,
      "loss": 0.0646,
      "step": 3730
    },
    {
      "epoch": 38.96,
      "learning_rate": 1.4467592592592594e-06,
      "loss": 0.0727,
      "step": 3740
    },
    {
      "epoch": 39.0,
      "eval_accuracy": 0.5581818181818182,
      "eval_loss": 3.519263744354248,
      "eval_runtime": 3.9518,
      "eval_samples_per_second": 139.177,
      "eval_steps_per_second": 8.857,
      "step": 3744
    },
    {
      "epoch": 39.06,
      "learning_rate": 1.3020833333333335e-06,
      "loss": 0.1025,
      "step": 3750
    },
    {
      "epoch": 39.17,
      "learning_rate": 1.1574074074074074e-06,
      "loss": 0.1064,
      "step": 3760
    },
    {
      "epoch": 39.27,
      "learning_rate": 1.0127314814814815e-06,
      "loss": 0.0921,
      "step": 3770
    },
    {
      "epoch": 39.38,
      "learning_rate": 8.680555555555556e-07,
      "loss": 0.0786,
      "step": 3780
    },
    {
      "epoch": 39.48,
      "learning_rate": 7.233796296296297e-07,
      "loss": 0.0767,
      "step": 3790
    },
    {
      "epoch": 39.58,
      "learning_rate": 5.787037037037037e-07,
      "loss": 0.0661,
      "step": 3800
    },
    {
      "epoch": 39.69,
      "learning_rate": 4.340277777777778e-07,
      "loss": 0.0873,
      "step": 3810
    },
    {
      "epoch": 39.79,
      "learning_rate": 2.8935185185185185e-07,
      "loss": 0.0692,
      "step": 3820
    },
    {
      "epoch": 39.9,
      "learning_rate": 1.4467592592592592e-07,
      "loss": 0.1063,
      "step": 3830
    },
    {
      "epoch": 40.0,
      "learning_rate": 0.0,
      "loss": 0.0619,
      "step": 3840
    },
    {
      "epoch": 40.0,
      "eval_accuracy": 0.5581818181818182,
      "eval_loss": 3.5836477279663086,
      "eval_runtime": 3.9696,
      "eval_samples_per_second": 138.553,
      "eval_steps_per_second": 8.817,
      "step": 3840
    },
    {
      "epoch": 40.0,
      "step": 3840,
      "total_flos": 1.9002720805055447e+19,
      "train_loss": 0.31003439707371094,
      "train_runtime": 4246.6951,
      "train_samples_per_second": 57.767,
      "train_steps_per_second": 0.904
    }
  ],
  "logging_steps": 10,
  "max_steps": 3840,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 40,
  "save_steps": 500,
  "total_flos": 1.9002720805055447e+19,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": null
}