{
  "best_metric": 0.78125,
  "best_model_checkpoint": "Swin-dmae-DA5-N-Colab\\checkpoint-1361",
  "epoch": 116.36363636363636,
  "eval_steps": 500,
  "global_step": 2880,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.4,
      "learning_rate": 1.3888888888888892e-06,
      "loss": 6.7613,
      "step": 10
    },
    {
      "epoch": 0.81,
      "learning_rate": 2.7777777777777783e-06,
      "loss": 6.8626,
      "step": 20
    },
    {
      "epoch": 0.97,
      "eval_accuracy": 0.125,
      "eval_loss": 7.802581787109375,
      "eval_runtime": 0.4701,
      "eval_samples_per_second": 68.069,
      "eval_steps_per_second": 4.254,
      "step": 24
    },
    {
      "epoch": 1.21,
      "learning_rate": 4.166666666666667e-06,
      "loss": 6.8425,
      "step": 30
    },
    {
      "epoch": 1.62,
      "learning_rate": 5.555555555555557e-06,
      "loss": 6.6392,
      "step": 40
    },
    {
      "epoch": 1.98,
      "eval_accuracy": 0.125,
      "eval_loss": 7.5823211669921875,
      "eval_runtime": 0.4001,
      "eval_samples_per_second": 79.982,
      "eval_steps_per_second": 4.999,
      "step": 49
    },
    {
      "epoch": 2.02,
      "learning_rate": 6.944444444444445e-06,
      "loss": 6.5894,
      "step": 50
    },
    {
      "epoch": 2.42,
      "learning_rate": 8.333333333333334e-06,
      "loss": 6.3711,
      "step": 60
    },
    {
      "epoch": 2.83,
      "learning_rate": 9.722222222222223e-06,
      "loss": 6.3514,
      "step": 70
    },
    {
      "epoch": 2.99,
      "eval_accuracy": 0.125,
      "eval_loss": 6.665298938751221,
      "eval_runtime": 0.4046,
      "eval_samples_per_second": 79.091,
      "eval_steps_per_second": 4.943,
      "step": 74
    },
    {
      "epoch": 3.23,
      "learning_rate": 1.1111111111111113e-05,
      "loss": 5.8289,
      "step": 80
    },
    {
      "epoch": 3.64,
      "learning_rate": 1.25e-05,
      "loss": 5.3109,
      "step": 90
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.125,
      "eval_loss": 5.279791355133057,
      "eval_runtime": 0.4241,
      "eval_samples_per_second": 75.455,
      "eval_steps_per_second": 4.716,
      "step": 99
    },
    {
      "epoch": 4.04,
      "learning_rate": 1.388888888888889e-05,
      "loss": 4.9844,
      "step": 100
    },
    {
      "epoch": 4.44,
      "learning_rate": 1.5277777777777777e-05,
      "loss": 4.5015,
      "step": 110
    },
    {
      "epoch": 4.85,
      "learning_rate": 1.6666666666666667e-05,
      "loss": 3.7476,
      "step": 120
    },
    {
      "epoch": 4.97,
      "eval_accuracy": 0.125,
      "eval_loss": 3.618558168411255,
      "eval_runtime": 0.4101,
      "eval_samples_per_second": 78.031,
      "eval_steps_per_second": 4.877,
      "step": 123
    },
    {
      "epoch": 5.25,
      "learning_rate": 1.8055555555555558e-05,
      "loss": 3.1884,
      "step": 130
    },
    {
      "epoch": 5.66,
      "learning_rate": 1.9444444444444445e-05,
      "loss": 2.7138,
      "step": 140
    },
    {
      "epoch": 5.98,
      "eval_accuracy": 0.125,
      "eval_loss": 1.937683343887329,
      "eval_runtime": 0.4036,
      "eval_samples_per_second": 79.286,
      "eval_steps_per_second": 4.955,
      "step": 148
    },
    {
      "epoch": 6.06,
      "learning_rate": 2.0833333333333336e-05,
      "loss": 2.1007,
      "step": 150
    },
    {
      "epoch": 6.46,
      "learning_rate": 2.2222222222222227e-05,
      "loss": 1.7134,
      "step": 160
    },
    {
      "epoch": 6.87,
      "learning_rate": 2.3611111111111114e-05,
      "loss": 1.4116,
      "step": 170
    },
    {
      "epoch": 6.99,
      "eval_accuracy": 0.125,
      "eval_loss": 1.4680219888687134,
      "eval_runtime": 0.3951,
      "eval_samples_per_second": 80.993,
      "eval_steps_per_second": 5.062,
      "step": 173
    },
    {
      "epoch": 7.27,
      "learning_rate": 2.5e-05,
      "loss": 1.4216,
      "step": 180
    },
    {
      "epoch": 7.68,
      "learning_rate": 2.6388888888888892e-05,
      "loss": 1.3932,
      "step": 190
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.5,
      "eval_loss": 1.3818670511245728,
      "eval_runtime": 0.4901,
      "eval_samples_per_second": 65.291,
      "eval_steps_per_second": 4.081,
      "step": 198
    },
    {
      "epoch": 8.08,
      "learning_rate": 2.777777777777778e-05,
      "loss": 1.3888,
      "step": 200
    },
    {
      "epoch": 8.48,
      "learning_rate": 2.9166666666666666e-05,
      "loss": 1.3077,
      "step": 210
    },
    {
      "epoch": 8.89,
      "learning_rate": 3.0555555555555554e-05,
      "loss": 1.2566,
      "step": 220
    },
    {
      "epoch": 8.97,
      "eval_accuracy": 0.15625,
      "eval_loss": 1.591198205947876,
      "eval_runtime": 0.4171,
      "eval_samples_per_second": 76.721,
      "eval_steps_per_second": 4.795,
      "step": 222
    },
    {
      "epoch": 9.29,
      "learning_rate": 3.194444444444445e-05,
      "loss": 1.1679,
      "step": 230
    },
    {
      "epoch": 9.7,
      "learning_rate": 3.3333333333333335e-05,
      "loss": 1.1332,
      "step": 240
    },
    {
      "epoch": 9.98,
      "eval_accuracy": 0.4375,
      "eval_loss": 1.3336095809936523,
      "eval_runtime": 0.4176,
      "eval_samples_per_second": 76.628,
      "eval_steps_per_second": 4.789,
      "step": 247
    },
    {
      "epoch": 10.1,
      "learning_rate": 3.472222222222223e-05,
      "loss": 1.0699,
      "step": 250
    },
    {
      "epoch": 10.51,
      "learning_rate": 3.6111111111111116e-05,
      "loss": 0.9928,
      "step": 260
    },
    {
      "epoch": 10.91,
      "learning_rate": 3.7500000000000003e-05,
      "loss": 0.9511,
      "step": 270
    },
    {
      "epoch": 10.99,
      "eval_accuracy": 0.28125,
      "eval_loss": 1.1492191553115845,
      "eval_runtime": 0.4006,
      "eval_samples_per_second": 79.881,
      "eval_steps_per_second": 4.993,
      "step": 272
    },
    {
      "epoch": 11.31,
      "learning_rate": 3.888888888888889e-05,
      "loss": 0.9286,
      "step": 280
    },
    {
      "epoch": 11.72,
      "learning_rate": 3.996913580246914e-05,
      "loss": 0.8905,
      "step": 290
    },
    {
      "epoch": 12.0,
      "eval_accuracy": 0.53125,
      "eval_loss": 1.0571255683898926,
      "eval_runtime": 0.4566,
      "eval_samples_per_second": 70.082,
      "eval_steps_per_second": 4.38,
      "step": 297
    },
    {
      "epoch": 12.12,
      "learning_rate": 3.981481481481482e-05,
      "loss": 0.8031,
      "step": 300
    },
    {
      "epoch": 12.53,
      "learning_rate": 3.9660493827160496e-05,
      "loss": 0.8253,
      "step": 310
    },
    {
      "epoch": 12.93,
      "learning_rate": 3.950617283950617e-05,
      "loss": 0.8317,
      "step": 320
    },
    {
      "epoch": 12.97,
      "eval_accuracy": 0.65625,
      "eval_loss": 0.8473750948905945,
      "eval_runtime": 0.4201,
      "eval_samples_per_second": 76.173,
      "eval_steps_per_second": 4.761,
      "step": 321
    },
    {
      "epoch": 13.33,
      "learning_rate": 3.935185185185186e-05,
      "loss": 0.7445,
      "step": 330
    },
    {
      "epoch": 13.74,
      "learning_rate": 3.9197530864197535e-05,
      "loss": 0.6611,
      "step": 340
    },
    {
      "epoch": 13.98,
      "eval_accuracy": 0.71875,
      "eval_loss": 0.7519997954368591,
      "eval_runtime": 0.4141,
      "eval_samples_per_second": 77.277,
      "eval_steps_per_second": 4.83,
      "step": 346
    },
    {
      "epoch": 14.14,
      "learning_rate": 3.904320987654321e-05,
      "loss": 0.6514,
      "step": 350
    },
    {
      "epoch": 14.55,
      "learning_rate": 3.888888888888889e-05,
      "loss": 0.6022,
      "step": 360
    },
    {
      "epoch": 14.95,
      "learning_rate": 3.8734567901234575e-05,
      "loss": 0.5683,
      "step": 370
    },
    {
      "epoch": 14.99,
      "eval_accuracy": 0.75,
      "eval_loss": 0.6655727624893188,
      "eval_runtime": 0.4031,
      "eval_samples_per_second": 79.386,
      "eval_steps_per_second": 4.962,
      "step": 371
    },
    {
      "epoch": 15.35,
      "learning_rate": 3.8580246913580246e-05,
      "loss": 0.5743,
      "step": 380
    },
    {
      "epoch": 15.76,
      "learning_rate": 3.842592592592593e-05,
      "loss": 0.569,
      "step": 390
    },
    {
      "epoch": 16.0,
      "eval_accuracy": 0.53125,
      "eval_loss": 0.8109104037284851,
      "eval_runtime": 0.4166,
      "eval_samples_per_second": 76.812,
      "eval_steps_per_second": 4.801,
      "step": 396
    },
    {
      "epoch": 16.16,
      "learning_rate": 3.827160493827161e-05,
      "loss": 0.4743,
      "step": 400
    },
    {
      "epoch": 16.57,
      "learning_rate": 3.8117283950617286e-05,
      "loss": 0.502,
      "step": 410
    },
    {
      "epoch": 16.97,
      "learning_rate": 3.7962962962962964e-05,
      "loss": 0.4702,
      "step": 420
    },
    {
      "epoch": 16.97,
      "eval_accuracy": 0.625,
      "eval_loss": 0.703583836555481,
      "eval_runtime": 0.4076,
      "eval_samples_per_second": 78.509,
      "eval_steps_per_second": 4.907,
      "step": 420
    },
    {
      "epoch": 17.37,
      "learning_rate": 3.780864197530865e-05,
      "loss": 0.4441,
      "step": 430
    },
    {
      "epoch": 17.78,
      "learning_rate": 3.7654320987654326e-05,
      "loss": 0.4244,
      "step": 440
    },
    {
      "epoch": 17.98,
      "eval_accuracy": 0.65625,
      "eval_loss": 0.8169162273406982,
      "eval_runtime": 0.4216,
      "eval_samples_per_second": 75.9,
      "eval_steps_per_second": 4.744,
      "step": 445
    },
    {
      "epoch": 18.18,
      "learning_rate": 3.7500000000000003e-05,
      "loss": 0.4277,
      "step": 450
    },
    {
      "epoch": 18.59,
      "learning_rate": 3.734567901234568e-05,
      "loss": 0.3584,
      "step": 460
    },
    {
      "epoch": 18.99,
      "learning_rate": 3.719135802469136e-05,
      "loss": 0.3483,
      "step": 470
    },
    {
      "epoch": 18.99,
      "eval_accuracy": 0.71875,
      "eval_loss": 0.7075902223587036,
      "eval_runtime": 0.4306,
      "eval_samples_per_second": 74.312,
      "eval_steps_per_second": 4.645,
      "step": 470
    },
    {
      "epoch": 19.39,
      "learning_rate": 3.703703703703704e-05,
      "loss": 0.3005,
      "step": 480
    },
    {
      "epoch": 19.8,
      "learning_rate": 3.688271604938272e-05,
      "loss": 0.3853,
      "step": 490
    },
    {
      "epoch": 20.0,
      "eval_accuracy": 0.71875,
      "eval_loss": 0.8644444942474365,
      "eval_runtime": 0.4716,
      "eval_samples_per_second": 67.852,
      "eval_steps_per_second": 4.241,
      "step": 495
    },
    {
      "epoch": 20.2,
      "learning_rate": 3.67283950617284e-05,
      "loss": 0.307,
      "step": 500
    },
    {
      "epoch": 20.61,
      "learning_rate": 3.6574074074074076e-05,
      "loss": 0.3038,
      "step": 510
    },
    {
      "epoch": 20.97,
      "eval_accuracy": 0.71875,
      "eval_loss": 0.86528080701828,
      "eval_runtime": 0.4181,
      "eval_samples_per_second": 76.538,
      "eval_steps_per_second": 4.784,
      "step": 519
    },
    {
      "epoch": 21.01,
      "learning_rate": 3.641975308641976e-05,
      "loss": 0.3084,
      "step": 520
    },
    {
      "epoch": 21.41,
      "learning_rate": 3.626543209876543e-05,
      "loss": 0.2967,
      "step": 530
    },
    {
      "epoch": 21.82,
      "learning_rate": 3.6111111111111116e-05,
      "loss": 0.2885,
      "step": 540
    },
    {
      "epoch": 21.98,
      "eval_accuracy": 0.71875,
      "eval_loss": 1.0443750619888306,
      "eval_runtime": 0.4421,
      "eval_samples_per_second": 72.382,
      "eval_steps_per_second": 4.524,
      "step": 544
    },
    {
      "epoch": 22.22,
      "learning_rate": 3.5956790123456794e-05,
      "loss": 0.2644,
      "step": 550
    },
    {
      "epoch": 22.63,
      "learning_rate": 3.580246913580247e-05,
      "loss": 0.2014,
      "step": 560
    },
    {
      "epoch": 22.99,
      "eval_accuracy": 0.59375,
      "eval_loss": 1.06843900680542,
      "eval_runtime": 0.3991,
      "eval_samples_per_second": 80.182,
      "eval_steps_per_second": 5.011,
      "step": 569
    },
    {
      "epoch": 23.03,
      "learning_rate": 3.564814814814815e-05,
      "loss": 0.309,
      "step": 570
    },
    {
      "epoch": 23.43,
      "learning_rate": 3.5493827160493834e-05,
      "loss": 0.2515,
      "step": 580
    },
    {
      "epoch": 23.84,
      "learning_rate": 3.533950617283951e-05,
      "loss": 0.2764,
      "step": 590
    },
    {
      "epoch": 24.0,
      "eval_accuracy": 0.65625,
      "eval_loss": 1.1421856880187988,
      "eval_runtime": 0.4621,
      "eval_samples_per_second": 69.248,
      "eval_steps_per_second": 4.328,
      "step": 594
    },
    {
      "epoch": 24.24,
      "learning_rate": 3.518518518518519e-05,
      "loss": 0.2519,
      "step": 600
    },
    {
      "epoch": 24.65,
      "learning_rate": 3.503086419753087e-05,
      "loss": 0.2493,
      "step": 610
    },
    {
      "epoch": 24.97,
      "eval_accuracy": 0.6875,
      "eval_loss": 1.102545976638794,
      "eval_runtime": 0.4256,
      "eval_samples_per_second": 75.187,
      "eval_steps_per_second": 4.699,
      "step": 618
    },
    {
      "epoch": 25.05,
      "learning_rate": 3.4876543209876545e-05,
      "loss": 0.2863,
      "step": 620
    },
    {
      "epoch": 25.45,
      "learning_rate": 3.472222222222223e-05,
      "loss": 0.2352,
      "step": 630
    },
    {
      "epoch": 25.86,
      "learning_rate": 3.45679012345679e-05,
      "loss": 0.2754,
      "step": 640
    },
    {
      "epoch": 25.98,
      "eval_accuracy": 0.71875,
      "eval_loss": 1.0370571613311768,
      "eval_runtime": 0.4276,
      "eval_samples_per_second": 74.835,
      "eval_steps_per_second": 4.677,
      "step": 643
    },
    {
      "epoch": 26.26,
      "learning_rate": 3.4413580246913584e-05,
      "loss": 0.2549,
      "step": 650
    },
    {
      "epoch": 26.67,
      "learning_rate": 3.425925925925926e-05,
      "loss": 0.1793,
      "step": 660
    },
    {
      "epoch": 26.99,
      "eval_accuracy": 0.65625,
      "eval_loss": 1.1623690128326416,
      "eval_runtime": 0.4181,
      "eval_samples_per_second": 76.534,
      "eval_steps_per_second": 4.783,
      "step": 668
    },
    {
      "epoch": 27.07,
      "learning_rate": 3.410493827160494e-05,
      "loss": 0.2305,
      "step": 670
    },
    {
      "epoch": 27.47,
      "learning_rate": 3.395061728395062e-05,
      "loss": 0.1836,
      "step": 680
    },
    {
      "epoch": 27.88,
      "learning_rate": 3.37962962962963e-05,
      "loss": 0.1971,
      "step": 690
    },
    {
      "epoch": 28.0,
      "eval_accuracy": 0.6875,
      "eval_loss": 1.3177158832550049,
      "eval_runtime": 0.3941,
      "eval_samples_per_second": 81.2,
      "eval_steps_per_second": 5.075,
      "step": 693
    },
    {
      "epoch": 28.28,
      "learning_rate": 3.364197530864198e-05,
      "loss": 0.1514,
      "step": 700
    },
    {
      "epoch": 28.69,
      "learning_rate": 3.348765432098766e-05,
      "loss": 0.1881,
      "step": 710
    },
    {
      "epoch": 28.97,
      "eval_accuracy": 0.6875,
      "eval_loss": 1.2813467979431152,
      "eval_runtime": 0.3961,
      "eval_samples_per_second": 80.79,
      "eval_steps_per_second": 5.049,
      "step": 717
    },
    {
      "epoch": 29.09,
      "learning_rate": 3.3333333333333335e-05,
      "loss": 0.1854,
      "step": 720
    },
    {
      "epoch": 29.49,
      "learning_rate": 3.317901234567901e-05,
      "loss": 0.2045,
      "step": 730
    },
    {
      "epoch": 29.9,
      "learning_rate": 3.30246913580247e-05,
      "loss": 0.167,
      "step": 740
    },
    {
      "epoch": 29.98,
      "eval_accuracy": 0.625,
      "eval_loss": 1.5564466714859009,
      "eval_runtime": 0.4011,
      "eval_samples_per_second": 79.783,
      "eval_steps_per_second": 4.986,
      "step": 742
    },
    {
      "epoch": 30.3,
      "learning_rate": 3.2870370370370375e-05,
      "loss": 0.1761,
      "step": 750
    },
    {
      "epoch": 30.71,
      "learning_rate": 3.271604938271605e-05,
      "loss": 0.1872,
      "step": 760
    },
    {
      "epoch": 30.99,
      "eval_accuracy": 0.71875,
      "eval_loss": 1.3761742115020752,
      "eval_runtime": 0.4136,
      "eval_samples_per_second": 77.369,
      "eval_steps_per_second": 4.836,
      "step": 767
    },
    {
      "epoch": 31.11,
      "learning_rate": 3.256172839506173e-05,
      "loss": 0.1697,
      "step": 770
    },
    {
      "epoch": 31.52,
      "learning_rate": 3.2407407407407415e-05,
      "loss": 0.1806,
      "step": 780
    },
    {
      "epoch": 31.92,
      "learning_rate": 3.2253086419753086e-05,
      "loss": 0.1374,
      "step": 790
    },
    {
      "epoch": 32.0,
      "eval_accuracy": 0.625,
      "eval_loss": 1.4406580924987793,
      "eval_runtime": 0.4051,
      "eval_samples_per_second": 78.994,
      "eval_steps_per_second": 4.937,
      "step": 792
    },
    {
      "epoch": 32.32,
      "learning_rate": 3.209876543209877e-05,
      "loss": 0.1848,
      "step": 800
    },
    {
      "epoch": 32.73,
      "learning_rate": 3.194444444444445e-05,
      "loss": 0.1841,
      "step": 810
    },
    {
      "epoch": 32.97,
      "eval_accuracy": 0.6875,
      "eval_loss": 1.4038010835647583,
      "eval_runtime": 0.4176,
      "eval_samples_per_second": 76.627,
      "eval_steps_per_second": 4.789,
      "step": 816
    },
    {
      "epoch": 33.13,
      "learning_rate": 3.1790123456790125e-05,
      "loss": 0.1365,
      "step": 820
    },
    {
      "epoch": 33.54,
      "learning_rate": 3.16358024691358e-05,
      "loss": 0.1309,
      "step": 830
    },
    {
      "epoch": 33.94,
      "learning_rate": 3.148148148148149e-05,
      "loss": 0.167,
      "step": 840
    },
    {
      "epoch": 33.98,
      "eval_accuracy": 0.6875,
      "eval_loss": 1.3768980503082275,
      "eval_runtime": 0.4141,
      "eval_samples_per_second": 77.277,
      "eval_steps_per_second": 4.83,
      "step": 841
    },
    {
      "epoch": 34.34,
      "learning_rate": 3.1327160493827165e-05,
      "loss": 0.1689,
      "step": 850
    },
    {
      "epoch": 34.75,
      "learning_rate": 3.117283950617284e-05,
      "loss": 0.1614,
      "step": 860
    },
    {
      "epoch": 34.99,
      "eval_accuracy": 0.65625,
      "eval_loss": 1.5350613594055176,
      "eval_runtime": 0.4051,
      "eval_samples_per_second": 78.994,
      "eval_steps_per_second": 4.937,
      "step": 866
    },
    {
      "epoch": 35.15,
      "learning_rate": 3.101851851851852e-05,
      "loss": 0.1439,
      "step": 870
    },
    {
      "epoch": 35.56,
      "learning_rate": 3.08641975308642e-05,
      "loss": 0.1706,
      "step": 880
    },
    {
      "epoch": 35.96,
      "learning_rate": 3.070987654320988e-05,
      "loss": 0.1835,
      "step": 890
    },
    {
      "epoch": 36.0,
      "eval_accuracy": 0.6875,
      "eval_loss": 1.4466042518615723,
      "eval_runtime": 0.4396,
      "eval_samples_per_second": 72.793,
      "eval_steps_per_second": 4.55,
      "step": 891
    },
    {
      "epoch": 36.36,
      "learning_rate": 3.0555555555555554e-05,
      "loss": 0.1558,
      "step": 900
    },
    {
      "epoch": 36.77,
      "learning_rate": 3.0401234567901238e-05,
      "loss": 0.1917,
      "step": 910
    },
    {
      "epoch": 36.97,
      "eval_accuracy": 0.75,
      "eval_loss": 1.349318265914917,
      "eval_runtime": 0.4031,
      "eval_samples_per_second": 79.387,
      "eval_steps_per_second": 4.962,
      "step": 915
    },
    {
      "epoch": 37.17,
      "learning_rate": 3.0246913580246916e-05,
      "loss": 0.1761,
      "step": 920
    },
    {
      "epoch": 37.58,
      "learning_rate": 3.0092592592592593e-05,
      "loss": 0.1861,
      "step": 930
    },
    {
      "epoch": 37.98,
      "learning_rate": 2.9938271604938275e-05,
      "loss": 0.1171,
      "step": 940
    },
    {
      "epoch": 37.98,
      "eval_accuracy": 0.75,
      "eval_loss": 1.4756131172180176,
      "eval_runtime": 0.4051,
      "eval_samples_per_second": 78.994,
      "eval_steps_per_second": 4.937,
      "step": 940
    },
    {
      "epoch": 38.38,
      "learning_rate": 2.9783950617283956e-05,
      "loss": 0.1234,
      "step": 950
    },
    {
      "epoch": 38.79,
      "learning_rate": 2.962962962962963e-05,
      "loss": 0.163,
      "step": 960
    },
    {
      "epoch": 38.99,
      "eval_accuracy": 0.6875,
      "eval_loss": 1.4373202323913574,
      "eval_runtime": 0.4221,
      "eval_samples_per_second": 75.809,
      "eval_steps_per_second": 4.738,
      "step": 965
    },
    {
      "epoch": 39.19,
      "learning_rate": 2.947530864197531e-05,
      "loss": 0.1452,
      "step": 970
    },
    {
      "epoch": 39.6,
      "learning_rate": 2.9320987654320992e-05,
      "loss": 0.088,
      "step": 980
    },
    {
      "epoch": 40.0,
      "learning_rate": 2.9166666666666666e-05,
      "loss": 0.1688,
      "step": 990
    },
    {
      "epoch": 40.0,
      "eval_accuracy": 0.75,
      "eval_loss": 1.4082034826278687,
      "eval_runtime": 0.4786,
      "eval_samples_per_second": 66.859,
      "eval_steps_per_second": 4.179,
      "step": 990
    },
    {
      "epoch": 40.4,
      "learning_rate": 2.9012345679012347e-05,
      "loss": 0.1359,
      "step": 1000
    },
    {
      "epoch": 40.81,
      "learning_rate": 2.885802469135803e-05,
      "loss": 0.1318,
      "step": 1010
    },
    {
      "epoch": 40.97,
      "eval_accuracy": 0.6875,
      "eval_loss": 1.5907220840454102,
      "eval_runtime": 0.4061,
      "eval_samples_per_second": 78.797,
      "eval_steps_per_second": 4.925,
      "step": 1014
    },
    {
      "epoch": 41.21,
      "learning_rate": 2.8703703703703706e-05,
      "loss": 0.1547,
      "step": 1020
    },
    {
      "epoch": 41.62,
      "learning_rate": 2.8549382716049384e-05,
      "loss": 0.1107,
      "step": 1030
    },
    {
      "epoch": 41.98,
      "eval_accuracy": 0.6875,
      "eval_loss": 1.7461936473846436,
      "eval_runtime": 0.4081,
      "eval_samples_per_second": 78.414,
      "eval_steps_per_second": 4.901,
      "step": 1039
    },
    {
      "epoch": 42.02,
      "learning_rate": 2.8395061728395065e-05,
      "loss": 0.1144,
      "step": 1040
    },
    {
      "epoch": 42.42,
      "learning_rate": 2.8240740740740743e-05,
      "loss": 0.1239,
      "step": 1050
    },
    {
      "epoch": 42.83,
      "learning_rate": 2.8086419753086424e-05,
      "loss": 0.1064,
      "step": 1060
    },
    {
      "epoch": 42.99,
      "eval_accuracy": 0.5625,
      "eval_loss": 1.8704116344451904,
      "eval_runtime": 0.4081,
      "eval_samples_per_second": 78.414,
      "eval_steps_per_second": 4.901,
      "step": 1064
    },
    {
      "epoch": 43.23,
      "learning_rate": 2.79320987654321e-05,
      "loss": 0.0978,
      "step": 1070
    },
    {
      "epoch": 43.64,
      "learning_rate": 2.777777777777778e-05,
      "loss": 0.1423,
      "step": 1080
    },
    {
      "epoch": 44.0,
      "eval_accuracy": 0.5625,
      "eval_loss": 1.715477705001831,
      "eval_runtime": 0.4051,
      "eval_samples_per_second": 78.994,
      "eval_steps_per_second": 4.937,
      "step": 1089
    },
    {
      "epoch": 44.04,
      "learning_rate": 2.762345679012346e-05,
      "loss": 0.1031,
      "step": 1090
    },
    {
      "epoch": 44.44,
      "learning_rate": 2.746913580246914e-05,
      "loss": 0.1473,
      "step": 1100
    },
    {
      "epoch": 44.85,
      "learning_rate": 2.7314814814814816e-05,
      "loss": 0.082,
      "step": 1110
    },
    {
      "epoch": 44.97,
      "eval_accuracy": 0.71875,
      "eval_loss": 1.5552070140838623,
      "eval_runtime": 0.3991,
      "eval_samples_per_second": 80.182,
      "eval_steps_per_second": 5.011,
      "step": 1113
    },
    {
      "epoch": 45.25,
      "learning_rate": 2.7160493827160497e-05,
      "loss": 0.1206,
      "step": 1120
    },
    {
      "epoch": 45.66,
      "learning_rate": 2.7006172839506174e-05,
      "loss": 0.1012,
      "step": 1130
    },
    {
      "epoch": 45.98,
      "eval_accuracy": 0.6875,
      "eval_loss": 1.4189940690994263,
      "eval_runtime": 0.4136,
      "eval_samples_per_second": 77.37,
      "eval_steps_per_second": 4.836,
      "step": 1138
    },
    {
      "epoch": 46.06,
      "learning_rate": 2.6851851851851852e-05,
      "loss": 0.1124,
      "step": 1140
    },
    {
      "epoch": 46.46,
      "learning_rate": 2.6697530864197533e-05,
      "loss": 0.114,
      "step": 1150
    },
    {
      "epoch": 46.87,
      "learning_rate": 2.654320987654321e-05,
      "loss": 0.1001,
      "step": 1160
    },
    {
      "epoch": 46.99,
      "eval_accuracy": 0.71875,
      "eval_loss": 1.6800808906555176,
      "eval_runtime": 0.4021,
      "eval_samples_per_second": 79.584,
      "eval_steps_per_second": 4.974,
      "step": 1163
    },
    {
      "epoch": 47.27,
      "learning_rate": 2.6388888888888892e-05,
      "loss": 0.1309,
      "step": 1170
    },
    {
      "epoch": 47.68,
      "learning_rate": 2.623456790123457e-05,
      "loss": 0.1037,
      "step": 1180
    },
    {
      "epoch": 48.0,
      "eval_accuracy": 0.71875,
      "eval_loss": 1.686426043510437,
      "eval_runtime": 0.4541,
      "eval_samples_per_second": 70.469,
      "eval_steps_per_second": 4.404,
      "step": 1188
    },
    {
      "epoch": 48.08,
      "learning_rate": 2.6080246913580247e-05,
      "loss": 0.0909,
      "step": 1190
    },
    {
      "epoch": 48.48,
      "learning_rate": 2.5925925925925928e-05,
      "loss": 0.1296,
      "step": 1200
    },
    {
      "epoch": 48.89,
      "learning_rate": 2.577160493827161e-05,
      "loss": 0.1089,
      "step": 1210
    },
    {
      "epoch": 48.97,
      "eval_accuracy": 0.6875,
      "eval_loss": 1.5225051641464233,
      "eval_runtime": 0.4051,
      "eval_samples_per_second": 78.995,
      "eval_steps_per_second": 4.937,
      "step": 1212
    },
    {
      "epoch": 49.29,
      "learning_rate": 2.5617283950617284e-05,
      "loss": 0.1184,
      "step": 1220
    },
    {
      "epoch": 49.7,
      "learning_rate": 2.5462962962962965e-05,
      "loss": 0.0835,
      "step": 1230
    },
    {
      "epoch": 49.98,
      "eval_accuracy": 0.6875,
      "eval_loss": 1.979781150817871,
      "eval_runtime": 0.4161,
      "eval_samples_per_second": 76.903,
      "eval_steps_per_second": 4.806,
      "step": 1237
    },
    {
      "epoch": 50.1,
      "learning_rate": 2.5308641975308646e-05,
      "loss": 0.1079,
      "step": 1240
    },
    {
      "epoch": 50.51,
      "learning_rate": 2.515432098765432e-05,
      "loss": 0.1204,
      "step": 1250
    },
    {
      "epoch": 50.91,
      "learning_rate": 2.5e-05,
      "loss": 0.0818,
      "step": 1260
    },
    {
      "epoch": 50.99,
      "eval_accuracy": 0.65625,
      "eval_loss": 1.726827621459961,
      "eval_runtime": 0.4051,
      "eval_samples_per_second": 78.995,
      "eval_steps_per_second": 4.937,
      "step": 1262
    },
    {
      "epoch": 51.31,
      "learning_rate": 2.4845679012345682e-05,
      "loss": 0.0759,
      "step": 1270
    },
    {
      "epoch": 51.72,
      "learning_rate": 2.469135802469136e-05,
      "loss": 0.1134,
      "step": 1280
    },
    {
      "epoch": 52.0,
      "eval_accuracy": 0.75,
      "eval_loss": 1.599599838256836,
      "eval_runtime": 0.4061,
      "eval_samples_per_second": 78.8,
      "eval_steps_per_second": 4.925,
      "step": 1287
    },
    {
      "epoch": 52.12,
      "learning_rate": 2.4537037037037038e-05,
      "loss": 0.0988,
      "step": 1290
    },
    {
      "epoch": 52.53,
      "learning_rate": 2.438271604938272e-05,
      "loss": 0.0851,
      "step": 1300
    },
    {
      "epoch": 52.93,
      "learning_rate": 2.4228395061728396e-05,
      "loss": 0.1115,
      "step": 1310
    },
    {
      "epoch": 52.97,
      "eval_accuracy": 0.65625,
      "eval_loss": 1.7280734777450562,
      "eval_runtime": 0.4181,
      "eval_samples_per_second": 76.538,
      "eval_steps_per_second": 4.784,
      "step": 1311
    },
    {
      "epoch": 53.33,
      "learning_rate": 2.4074074074074077e-05,
      "loss": 0.0995,
      "step": 1320
    },
    {
      "epoch": 53.74,
      "learning_rate": 2.391975308641976e-05,
      "loss": 0.0929,
      "step": 1330
    },
    {
      "epoch": 53.98,
      "eval_accuracy": 0.75,
      "eval_loss": 1.6346489191055298,
      "eval_runtime": 0.4061,
      "eval_samples_per_second": 78.8,
      "eval_steps_per_second": 4.925,
      "step": 1336
    },
    {
      "epoch": 54.14,
      "learning_rate": 2.3765432098765433e-05,
      "loss": 0.1187,
      "step": 1340
    },
    {
      "epoch": 54.55,
      "learning_rate": 2.3611111111111114e-05,
      "loss": 0.1477,
      "step": 1350
    },
    {
      "epoch": 54.95,
      "learning_rate": 2.3456790123456795e-05,
      "loss": 0.0909,
      "step": 1360
    },
    {
      "epoch": 54.99,
      "eval_accuracy": 0.78125,
      "eval_loss": 1.4369856119155884,
      "eval_runtime": 0.4041,
      "eval_samples_per_second": 79.19,
      "eval_steps_per_second": 4.949,
      "step": 1361
    },
    {
      "epoch": 55.35,
      "learning_rate": 2.330246913580247e-05,
      "loss": 0.1106,
      "step": 1370
    },
    {
      "epoch": 55.76,
      "learning_rate": 2.314814814814815e-05,
      "loss": 0.1076,
      "step": 1380
    },
    {
      "epoch": 56.0,
      "eval_accuracy": 0.78125,
      "eval_loss": 1.550971508026123,
      "eval_runtime": 0.4091,
      "eval_samples_per_second": 78.222,
      "eval_steps_per_second": 4.889,
      "step": 1386
    },
    {
      "epoch": 56.16,
      "learning_rate": 2.299382716049383e-05,
      "loss": 0.1301,
      "step": 1390
    },
    {
      "epoch": 56.57,
      "learning_rate": 2.2839506172839506e-05,
      "loss": 0.0807,
      "step": 1400
    },
    {
      "epoch": 56.97,
      "learning_rate": 2.2685185185185187e-05,
      "loss": 0.0948,
      "step": 1410
    },
    {
      "epoch": 56.97,
      "eval_accuracy": 0.75,
      "eval_loss": 1.6382544040679932,
      "eval_runtime": 0.4751,
      "eval_samples_per_second": 67.353,
      "eval_steps_per_second": 4.21,
      "step": 1410
    },
    {
      "epoch": 57.37,
      "learning_rate": 2.2530864197530865e-05,
      "loss": 0.0961,
      "step": 1420
    },
    {
      "epoch": 57.78,
      "learning_rate": 2.2376543209876546e-05,
      "loss": 0.0914,
      "step": 1430
    },
    {
      "epoch": 57.98,
      "eval_accuracy": 0.6875,
      "eval_loss": 1.6937541961669922,
      "eval_runtime": 0.3996,
      "eval_samples_per_second": 80.08,
      "eval_steps_per_second": 5.005,
      "step": 1435
    },
    {
      "epoch": 58.18,
      "learning_rate": 2.2222222222222227e-05,
      "loss": 0.0716,
      "step": 1440
    },
    {
      "epoch": 58.59,
      "learning_rate": 2.20679012345679e-05,
      "loss": 0.0911,
      "step": 1450
    },
    {
      "epoch": 58.99,
      "learning_rate": 2.1913580246913582e-05,
      "loss": 0.0598,
      "step": 1460
    },
    {
      "epoch": 58.99,
      "eval_accuracy": 0.75,
      "eval_loss": 1.6290702819824219,
      "eval_runtime": 0.3996,
      "eval_samples_per_second": 80.081,
      "eval_steps_per_second": 5.005,
      "step": 1460
    },
    {
      "epoch": 59.39,
      "learning_rate": 2.1759259259259263e-05,
      "loss": 0.087,
      "step": 1470
    },
    {
      "epoch": 59.8,
      "learning_rate": 2.1604938271604937e-05,
      "loss": 0.0769,
      "step": 1480
    },
    {
      "epoch": 60.0,
      "eval_accuracy": 0.75,
      "eval_loss": 1.6593661308288574,
      "eval_runtime": 0.4171,
      "eval_samples_per_second": 76.721,
      "eval_steps_per_second": 4.795,
      "step": 1485
    },
    {
      "epoch": 60.2,
      "learning_rate": 2.145061728395062e-05,
      "loss": 0.0964,
      "step": 1490
    },
    {
      "epoch": 60.61,
      "learning_rate": 2.12962962962963e-05,
      "loss": 0.0894,
      "step": 1500
    },
    {
      "epoch": 60.97,
      "eval_accuracy": 0.78125,
      "eval_loss": 1.630151629447937,
      "eval_runtime": 0.4706,
      "eval_samples_per_second": 67.996,
      "eval_steps_per_second": 4.25,
      "step": 1509
    },
    {
      "epoch": 61.01,
      "learning_rate": 2.1141975308641974e-05,
      "loss": 0.0664,
      "step": 1510
    },
    {
      "epoch": 61.41,
      "learning_rate": 2.0987654320987655e-05,
      "loss": 0.0701,
      "step": 1520
    },
    {
      "epoch": 61.82,
      "learning_rate": 2.0833333333333336e-05,
      "loss": 0.0999,
      "step": 1530
    },
    {
      "epoch": 61.98,
      "eval_accuracy": 0.71875,
      "eval_loss": 1.656170129776001,
      "eval_runtime": 0.7332,
      "eval_samples_per_second": 43.646,
      "eval_steps_per_second": 2.728,
      "step": 1534
    },
    {
      "epoch": 62.22,
      "learning_rate": 2.0679012345679014e-05,
      "loss": 0.0559,
      "step": 1540
    },
    {
      "epoch": 62.63,
      "learning_rate": 2.0524691358024695e-05,
      "loss": 0.0759,
      "step": 1550
    },
    {
      "epoch": 62.99,
      "eval_accuracy": 0.75,
      "eval_loss": 1.598868489265442,
      "eval_runtime": 0.4032,
      "eval_samples_per_second": 79.364,
      "eval_steps_per_second": 4.96,
      "step": 1559
    },
    {
      "epoch": 63.03,
      "learning_rate": 2.0370370370370372e-05,
      "loss": 0.1008,
      "step": 1560
    },
    {
      "epoch": 63.43,
      "learning_rate": 2.021604938271605e-05,
      "loss": 0.06,
      "step": 1570
    },
    {
      "epoch": 63.84,
      "learning_rate": 2.006172839506173e-05,
      "loss": 0.102,
      "step": 1580
    },
    {
      "epoch": 64.0,
      "eval_accuracy": 0.78125,
      "eval_loss": 1.660170078277588,
      "eval_runtime": 0.4806,
      "eval_samples_per_second": 66.581,
      "eval_steps_per_second": 4.161,
      "step": 1584
    },
    {
      "epoch": 64.24,
      "learning_rate": 1.990740740740741e-05,
      "loss": 0.0836,
      "step": 1590
    },
    {
      "epoch": 64.65,
      "learning_rate": 1.9753086419753087e-05,
      "loss": 0.0864,
      "step": 1600
    },
    {
      "epoch": 64.97,
      "eval_accuracy": 0.78125,
      "eval_loss": 1.7385599613189697,
      "eval_runtime": 0.4181,
      "eval_samples_per_second": 76.538,
      "eval_steps_per_second": 4.784,
      "step": 1608
    },
    {
      "epoch": 65.05,
      "learning_rate": 1.9598765432098768e-05,
      "loss": 0.076,
      "step": 1610
    },
    {
      "epoch": 65.45,
      "learning_rate": 1.9444444444444445e-05,
      "loss": 0.0763,
      "step": 1620
    },
    {
      "epoch": 65.86,
      "learning_rate": 1.9290123456790123e-05,
      "loss": 0.0722,
      "step": 1630
    },
    {
      "epoch": 65.98,
      "eval_accuracy": 0.71875,
      "eval_loss": 2.0494906902313232,
      "eval_runtime": 0.4086,
      "eval_samples_per_second": 78.317,
      "eval_steps_per_second": 4.895,
      "step": 1633
    },
    {
      "epoch": 66.26,
      "learning_rate": 1.9135802469135804e-05,
      "loss": 0.0929,
      "step": 1640
    },
    {
      "epoch": 66.67,
      "learning_rate": 1.8981481481481482e-05,
      "loss": 0.0956,
      "step": 1650
    },
    {
      "epoch": 66.99,
      "eval_accuracy": 0.6875,
      "eval_loss": 1.9748592376708984,
      "eval_runtime": 0.4206,
      "eval_samples_per_second": 76.081,
      "eval_steps_per_second": 4.755,
      "step": 1658
    },
    {
      "epoch": 67.07,
      "learning_rate": 1.8827160493827163e-05,
      "loss": 0.071,
      "step": 1660
    },
    {
      "epoch": 67.47,
      "learning_rate": 1.867283950617284e-05,
      "loss": 0.0844,
      "step": 1670
    },
    {
      "epoch": 67.88,
      "learning_rate": 1.851851851851852e-05,
      "loss": 0.0698,
      "step": 1680
    },
    {
      "epoch": 68.0,
      "eval_accuracy": 0.65625,
      "eval_loss": 2.0089621543884277,
      "eval_runtime": 0.4176,
      "eval_samples_per_second": 76.628,
      "eval_steps_per_second": 4.789,
      "step": 1683
    },
    {
      "epoch": 68.28,
      "learning_rate": 1.83641975308642e-05,
      "loss": 0.0837,
      "step": 1690
    },
    {
      "epoch": 68.69,
      "learning_rate": 1.820987654320988e-05,
      "loss": 0.0635,
      "step": 1700
    },
    {
      "epoch": 68.97,
      "eval_accuracy": 0.625,
      "eval_loss": 2.160045862197876,
      "eval_runtime": 0.4306,
      "eval_samples_per_second": 74.314,
      "eval_steps_per_second": 4.645,
      "step": 1707
    },
    {
      "epoch": 69.09,
      "learning_rate": 1.8055555555555558e-05,
      "loss": 0.0719,
      "step": 1710
    },
    {
      "epoch": 69.49,
      "learning_rate": 1.7901234567901236e-05,
      "loss": 0.0656,
      "step": 1720
    },
    {
      "epoch": 69.9,
      "learning_rate": 1.7746913580246917e-05,
      "loss": 0.0726,
      "step": 1730
    },
    {
      "epoch": 69.98,
      "eval_accuracy": 0.75,
      "eval_loss": 1.8476567268371582,
      "eval_runtime": 0.4076,
      "eval_samples_per_second": 78.508,
      "eval_steps_per_second": 4.907,
      "step": 1732
    },
    {
      "epoch": 70.3,
      "learning_rate": 1.7592592592592595e-05,
      "loss": 0.0651,
      "step": 1740
    },
    {
      "epoch": 70.71,
      "learning_rate": 1.7438271604938272e-05,
      "loss": 0.0905,
      "step": 1750
    },
    {
      "epoch": 70.99,
      "eval_accuracy": 0.71875,
      "eval_loss": 1.9970349073410034,
      "eval_runtime": 0.4801,
      "eval_samples_per_second": 66.652,
      "eval_steps_per_second": 4.166,
      "step": 1757
    },
    {
      "epoch": 71.11,
      "learning_rate": 1.728395061728395e-05,
      "loss": 0.0705,
      "step": 1760
    },
    {
      "epoch": 71.52,
      "learning_rate": 1.712962962962963e-05,
      "loss": 0.053,
      "step": 1770
    },
    {
      "epoch": 71.92,
      "learning_rate": 1.697530864197531e-05,
      "loss": 0.0955,
      "step": 1780
    },
    {
      "epoch": 72.0,
      "eval_accuracy": 0.75,
      "eval_loss": 1.9001073837280273,
      "eval_runtime": 0.4716,
      "eval_samples_per_second": 67.852,
      "eval_steps_per_second": 4.241,
      "step": 1782
    },
    {
      "epoch": 72.32,
      "learning_rate": 1.682098765432099e-05,
      "loss": 0.0943,
      "step": 1790
    },
    {
      "epoch": 72.73,
      "learning_rate": 1.6666666666666667e-05,
      "loss": 0.0614,
      "step": 1800
    },
    {
      "epoch": 72.97,
      "eval_accuracy": 0.65625,
      "eval_loss": 1.9346568584442139,
      "eval_runtime": 0.4161,
      "eval_samples_per_second": 76.906,
      "eval_steps_per_second": 4.807,
      "step": 1806
    },
    {
      "epoch": 73.13,
      "learning_rate": 1.651234567901235e-05,
      "loss": 0.0652,
      "step": 1810
    },
    {
      "epoch": 73.54,
      "learning_rate": 1.6358024691358026e-05,
      "loss": 0.0634,
      "step": 1820
    },
    {
      "epoch": 73.94,
      "learning_rate": 1.6203703703703707e-05,
      "loss": 0.0721,
      "step": 1830
    },
    {
      "epoch": 73.98,
      "eval_accuracy": 0.6875,
      "eval_loss": 1.900674819946289,
      "eval_runtime": 0.4776,
      "eval_samples_per_second": 66.996,
      "eval_steps_per_second": 4.187,
      "step": 1831
    },
    {
      "epoch": 74.34,
      "learning_rate": 1.6049382716049385e-05,
      "loss": 0.0633,
      "step": 1840
    },
    {
      "epoch": 74.75,
      "learning_rate": 1.5895061728395063e-05,
      "loss": 0.0868,
      "step": 1850
    },
    {
      "epoch": 74.99,
      "eval_accuracy": 0.65625,
      "eval_loss": 2.0204198360443115,
      "eval_runtime": 0.4036,
      "eval_samples_per_second": 79.286,
      "eval_steps_per_second": 4.955,
      "step": 1856
    },
    {
      "epoch": 75.15,
      "learning_rate": 1.5740740740740744e-05,
      "loss": 0.0798,
      "step": 1860
    },
    {
      "epoch": 75.56,
      "learning_rate": 1.558641975308642e-05,
      "loss": 0.0783,
      "step": 1870
    },
    {
      "epoch": 75.96,
      "learning_rate": 1.54320987654321e-05,
      "loss": 0.0817,
      "step": 1880
    },
    {
      "epoch": 76.0,
      "eval_accuracy": 0.71875,
      "eval_loss": 1.9806559085845947,
      "eval_runtime": 0.4096,
      "eval_samples_per_second": 78.123,
      "eval_steps_per_second": 4.883,
      "step": 1881
    },
    {
      "epoch": 76.36,
      "learning_rate": 1.5277777777777777e-05,
      "loss": 0.0823,
      "step": 1890
    },
    {
      "epoch": 76.77,
      "learning_rate": 1.5123456790123458e-05,
      "loss": 0.0533,
      "step": 1900
    },
    {
      "epoch": 76.97,
      "eval_accuracy": 0.75,
      "eval_loss": 1.978171467781067,
      "eval_runtime": 0.4191,
      "eval_samples_per_second": 76.353,
      "eval_steps_per_second": 4.772,
      "step": 1905
    },
    {
      "epoch": 77.17,
      "learning_rate": 1.4969135802469137e-05,
      "loss": 0.0866,
      "step": 1910
    },
    {
      "epoch": 77.58,
      "learning_rate": 1.4814814814814815e-05,
      "loss": 0.0737,
      "step": 1920
    },
    {
      "epoch": 77.98,
      "learning_rate": 1.4660493827160496e-05,
      "loss": 0.0682,
      "step": 1930
    },
    {
      "epoch": 77.98,
      "eval_accuracy": 0.75,
      "eval_loss": 1.831963062286377,
      "eval_runtime": 0.4216,
      "eval_samples_per_second": 75.902,
      "eval_steps_per_second": 4.744,
      "step": 1930
    },
    {
      "epoch": 78.38,
      "learning_rate": 1.4506172839506174e-05,
      "loss": 0.0819,
      "step": 1940
    },
    {
      "epoch": 78.79,
      "learning_rate": 1.4351851851851853e-05,
      "loss": 0.078,
      "step": 1950
    },
    {
      "epoch": 78.99,
      "eval_accuracy": 0.71875,
      "eval_loss": 1.835146427154541,
      "eval_runtime": 0.4216,
      "eval_samples_per_second": 75.902,
      "eval_steps_per_second": 4.744,
      "step": 1955
    },
    {
      "epoch": 79.19,
      "learning_rate": 1.4197530864197532e-05,
      "loss": 0.0574,
      "step": 1960
    },
    {
      "epoch": 79.6,
      "learning_rate": 1.4043209876543212e-05,
      "loss": 0.0468,
      "step": 1970
    },
    {
      "epoch": 80.0,
      "learning_rate": 1.388888888888889e-05,
      "loss": 0.0991,
      "step": 1980
    },
    {
      "epoch": 80.0,
      "eval_accuracy": 0.71875,
      "eval_loss": 1.9694030284881592,
      "eval_runtime": 0.4798,
      "eval_samples_per_second": 66.7,
      "eval_steps_per_second": 4.169,
      "step": 1980
    },
    {
      "epoch": 80.4,
      "learning_rate": 1.373456790123457e-05,
      "loss": 0.0466,
      "step": 1990
    },
    {
      "epoch": 80.81,
      "learning_rate": 1.3580246913580248e-05,
      "loss": 0.0601,
      "step": 2000
    },
    {
      "epoch": 80.97,
      "eval_accuracy": 0.71875,
      "eval_loss": 1.8794611692428589,
      "eval_runtime": 0.4141,
      "eval_samples_per_second": 77.276,
      "eval_steps_per_second": 4.83,
      "step": 2004
    },
    {
      "epoch": 81.21,
      "learning_rate": 1.3425925925925926e-05,
      "loss": 0.0567,
      "step": 2010
    },
    {
      "epoch": 81.62,
      "learning_rate": 1.3271604938271605e-05,
      "loss": 0.072,
      "step": 2020
    },
    {
      "epoch": 81.98,
      "eval_accuracy": 0.65625,
      "eval_loss": 2.02938175201416,
      "eval_runtime": 0.4232,
      "eval_samples_per_second": 75.623,
      "eval_steps_per_second": 4.726,
      "step": 2029
    },
    {
      "epoch": 82.02,
      "learning_rate": 1.3117283950617285e-05,
      "loss": 0.073,
      "step": 2030
    },
    {
      "epoch": 82.42,
      "learning_rate": 1.2962962962962964e-05,
      "loss": 0.0788,
      "step": 2040
    },
    {
      "epoch": 82.83,
      "learning_rate": 1.2808641975308642e-05,
      "loss": 0.0746,
      "step": 2050
    },
    {
      "epoch": 82.99,
      "eval_accuracy": 0.71875,
      "eval_loss": 1.8438613414764404,
      "eval_runtime": 0.3956,
      "eval_samples_per_second": 80.89,
      "eval_steps_per_second": 5.056,
      "step": 2054
    },
    {
      "epoch": 83.23,
      "learning_rate": 1.2654320987654323e-05,
      "loss": 0.0626,
      "step": 2060
    },
    {
      "epoch": 83.64,
      "learning_rate": 1.25e-05,
      "loss": 0.0547,
      "step": 2070
    },
    {
      "epoch": 84.0,
      "eval_accuracy": 0.71875,
      "eval_loss": 1.932082176208496,
      "eval_runtime": 0.4141,
      "eval_samples_per_second": 77.273,
      "eval_steps_per_second": 4.83,
      "step": 2079
    },
    {
      "epoch": 84.04,
      "learning_rate": 1.234567901234568e-05,
      "loss": 0.0632,
      "step": 2080
    },
    {
      "epoch": 84.44,
      "learning_rate": 1.219135802469136e-05,
      "loss": 0.0478,
      "step": 2090
    },
    {
      "epoch": 84.85,
      "learning_rate": 1.2037037037037039e-05,
      "loss": 0.0497,
      "step": 2100
    },
    {
      "epoch": 84.97,
      "eval_accuracy": 0.78125,
      "eval_loss": 1.8862378597259521,
      "eval_runtime": 0.4806,
      "eval_samples_per_second": 66.582,
      "eval_steps_per_second": 4.161,
      "step": 2103
    },
    {
      "epoch": 85.25,
      "learning_rate": 1.1882716049382716e-05,
      "loss": 0.0544,
      "step": 2110
    },
    {
      "epoch": 85.66,
      "learning_rate": 1.1728395061728398e-05,
      "loss": 0.0566,
      "step": 2120
    },
    {
      "epoch": 85.98,
      "eval_accuracy": 0.65625,
      "eval_loss": 2.0067098140716553,
      "eval_runtime": 0.5932,
      "eval_samples_per_second": 53.947,
      "eval_steps_per_second": 3.372,
      "step": 2128
    },
    {
      "epoch": 86.06,
      "learning_rate": 1.1574074074074075e-05,
      "loss": 0.0557,
      "step": 2130
    },
    {
      "epoch": 86.46,
      "learning_rate": 1.1419753086419753e-05,
      "loss": 0.0872,
      "step": 2140
    },
    {
      "epoch": 86.87,
      "learning_rate": 1.1265432098765432e-05,
      "loss": 0.0353,
      "step": 2150
    },
    {
      "epoch": 86.99,
      "eval_accuracy": 0.71875,
      "eval_loss": 2.095703363418579,
      "eval_runtime": 0.4511,
      "eval_samples_per_second": 70.942,
      "eval_steps_per_second": 4.434,
      "step": 2153
    },
    {
      "epoch": 87.27,
      "learning_rate": 1.1111111111111113e-05,
      "loss": 0.0482,
      "step": 2160
    },
    {
      "epoch": 87.68,
      "learning_rate": 1.0956790123456791e-05,
      "loss": 0.0634,
      "step": 2170
    },
    {
      "epoch": 88.0,
      "eval_accuracy": 0.65625,
      "eval_loss": 2.15714168548584,
      "eval_runtime": 0.4471,
      "eval_samples_per_second": 71.57,
      "eval_steps_per_second": 4.473,
      "step": 2178
    },
    {
      "epoch": 88.08,
      "learning_rate": 1.0802469135802469e-05,
      "loss": 0.0632,
      "step": 2180
    },
    {
      "epoch": 88.48,
      "learning_rate": 1.064814814814815e-05,
      "loss": 0.0504,
      "step": 2190
    },
    {
      "epoch": 88.89,
      "learning_rate": 1.0493827160493827e-05,
      "loss": 0.0477,
      "step": 2200
    },
    {
      "epoch": 88.97,
      "eval_accuracy": 0.6875,
      "eval_loss": 2.0384438037872314,
      "eval_runtime": 0.4271,
      "eval_samples_per_second": 74.924,
      "eval_steps_per_second": 4.683,
      "step": 2202
    },
    {
      "epoch": 89.29,
      "learning_rate": 1.0339506172839507e-05,
      "loss": 0.0446,
      "step": 2210
    },
    {
      "epoch": 89.7,
      "learning_rate": 1.0185185185185186e-05,
      "loss": 0.0513,
      "step": 2220
    },
    {
      "epoch": 89.98,
      "eval_accuracy": 0.75,
      "eval_loss": 1.9145632982254028,
      "eval_runtime": 0.4711,
      "eval_samples_per_second": 67.923,
      "eval_steps_per_second": 4.245,
      "step": 2227
    },
    {
      "epoch": 90.1,
      "learning_rate": 1.0030864197530866e-05,
      "loss": 0.0626,
      "step": 2230
    },
    {
      "epoch": 90.51,
      "learning_rate": 9.876543209876543e-06,
      "loss": 0.0354,
      "step": 2240
    },
    {
      "epoch": 90.91,
      "learning_rate": 9.722222222222223e-06,
      "loss": 0.0717,
      "step": 2250
    },
    {
      "epoch": 90.99,
      "eval_accuracy": 0.71875,
      "eval_loss": 1.8837898969650269,
      "eval_runtime": 0.4031,
      "eval_samples_per_second": 79.378,
      "eval_steps_per_second": 4.961,
      "step": 2252
    },
    {
      "epoch": 91.31,
      "learning_rate": 9.567901234567902e-06,
      "loss": 0.0486,
      "step": 2260
    },
    {
      "epoch": 91.72,
      "learning_rate": 9.413580246913581e-06,
      "loss": 0.0644,
      "step": 2270
    },
    {
      "epoch": 92.0,
      "eval_accuracy": 0.6875,
      "eval_loss": 1.9186292886734009,
      "eval_runtime": 0.4621,
      "eval_samples_per_second": 69.247,
      "eval_steps_per_second": 4.328,
      "step": 2277
    },
    {
      "epoch": 92.12,
      "learning_rate": 9.25925925925926e-06,
      "loss": 0.048,
      "step": 2280
    },
    {
      "epoch": 92.53,
      "learning_rate": 9.10493827160494e-06,
      "loss": 0.0711,
      "step": 2290
    },
    {
      "epoch": 92.93,
      "learning_rate": 8.950617283950618e-06,
      "loss": 0.0848,
      "step": 2300
    },
    {
      "epoch": 92.97,
      "eval_accuracy": 0.71875,
      "eval_loss": 1.882826805114746,
      "eval_runtime": 0.4741,
      "eval_samples_per_second": 67.495,
      "eval_steps_per_second": 4.218,
      "step": 2301
    },
    {
      "epoch": 93.33,
      "learning_rate": 8.796296296296297e-06,
      "loss": 0.038,
      "step": 2310
    },
    {
      "epoch": 93.74,
      "learning_rate": 8.641975308641975e-06,
      "loss": 0.0393,
      "step": 2320
    },
    {
      "epoch": 93.98,
      "eval_accuracy": 0.71875,
      "eval_loss": 1.9442168474197388,
      "eval_runtime": 0.4091,
      "eval_samples_per_second": 78.217,
      "eval_steps_per_second": 4.889,
      "step": 2326
    },
    {
      "epoch": 94.14,
      "learning_rate": 8.487654320987654e-06,
      "loss": 0.038,
      "step": 2330
    },
    {
      "epoch": 94.55,
      "learning_rate": 8.333333333333334e-06,
      "loss": 0.0703,
      "step": 2340
    },
    {
      "epoch": 94.95,
      "learning_rate": 8.179012345679013e-06,
      "loss": 0.046,
      "step": 2350
    },
    {
      "epoch": 94.99,
      "eval_accuracy": 0.71875,
      "eval_loss": 1.8865550756454468,
      "eval_runtime": 0.3931,
      "eval_samples_per_second": 81.403,
      "eval_steps_per_second": 5.088,
      "step": 2351
    },
    {
      "epoch": 95.35,
      "learning_rate": 8.024691358024692e-06,
      "loss": 0.0762,
      "step": 2360
    },
    {
      "epoch": 95.76,
      "learning_rate": 7.870370370370372e-06,
      "loss": 0.0487,
      "step": 2370
    },
    {
      "epoch": 96.0,
      "eval_accuracy": 0.6875,
      "eval_loss": 1.9787421226501465,
      "eval_runtime": 0.4221,
      "eval_samples_per_second": 75.807,
      "eval_steps_per_second": 4.738,
      "step": 2376
    },
    {
      "epoch": 96.16,
      "learning_rate": 7.71604938271605e-06,
      "loss": 0.0724,
      "step": 2380
    },
    {
      "epoch": 96.57,
      "learning_rate": 7.561728395061729e-06,
      "loss": 0.0563,
      "step": 2390
    },
    {
      "epoch": 96.97,
      "learning_rate": 7.4074074074074075e-06,
      "loss": 0.074,
      "step": 2400
    },
    {
      "epoch": 96.97,
      "eval_accuracy": 0.6875,
      "eval_loss": 2.0081026554107666,
      "eval_runtime": 0.6192,
      "eval_samples_per_second": 51.682,
      "eval_steps_per_second": 3.23,
      "step": 2400
    },
    {
      "epoch": 97.37,
      "learning_rate": 7.253086419753087e-06,
      "loss": 0.0624,
      "step": 2410
    },
    {
      "epoch": 97.78,
      "learning_rate": 7.098765432098766e-06,
      "loss": 0.0435,
      "step": 2420
    },
    {
      "epoch": 97.98,
      "eval_accuracy": 0.75,
      "eval_loss": 1.8838809728622437,
      "eval_runtime": 0.4211,
      "eval_samples_per_second": 75.988,
      "eval_steps_per_second": 4.749,
      "step": 2425
    },
    {
      "epoch": 98.18,
      "learning_rate": 6.944444444444445e-06,
      "loss": 0.0688,
      "step": 2430
    },
    {
      "epoch": 98.59,
      "learning_rate": 6.790123456790124e-06,
      "loss": 0.0441,
      "step": 2440
    },
    {
      "epoch": 98.99,
      "learning_rate": 6.635802469135803e-06,
      "loss": 0.0509,
      "step": 2450
    },
    {
      "epoch": 98.99,
      "eval_accuracy": 0.71875,
      "eval_loss": 1.9207700490951538,
      "eval_runtime": 0.4015,
      "eval_samples_per_second": 79.693,
      "eval_steps_per_second": 4.981,
      "step": 2450
    },
    {
      "epoch": 99.39,
      "learning_rate": 6.481481481481482e-06,
      "loss": 0.0557,
      "step": 2460
    },
    {
      "epoch": 99.8,
      "learning_rate": 6.3271604938271615e-06,
      "loss": 0.0571,
      "step": 2470
    },
    {
      "epoch": 100.0,
      "eval_accuracy": 0.71875,
      "eval_loss": 1.9770445823669434,
      "eval_runtime": 0.4191,
      "eval_samples_per_second": 76.363,
      "eval_steps_per_second": 4.773,
      "step": 2475
    },
    {
      "epoch": 100.2,
      "learning_rate": 6.17283950617284e-06,
      "loss": 0.0505,
      "step": 2480
    },
    {
      "epoch": 100.61,
      "learning_rate": 6.018518518518519e-06,
      "loss": 0.0327,
      "step": 2490
    },
    {
      "epoch": 100.97,
      "eval_accuracy": 0.71875,
      "eval_loss": 1.969954013824463,
      "eval_runtime": 0.5071,
      "eval_samples_per_second": 63.101,
      "eval_steps_per_second": 3.944,
      "step": 2499
    },
    {
      "epoch": 101.01,
      "learning_rate": 5.864197530864199e-06,
      "loss": 0.0472,
      "step": 2500
    },
    {
      "epoch": 101.41,
      "learning_rate": 5.7098765432098764e-06,
      "loss": 0.0512,
      "step": 2510
    },
    {
      "epoch": 101.82,
      "learning_rate": 5.555555555555557e-06,
      "loss": 0.0387,
      "step": 2520
    },
    {
      "epoch": 101.98,
      "eval_accuracy": 0.75,
      "eval_loss": 1.9250848293304443,
      "eval_runtime": 0.4241,
      "eval_samples_per_second": 75.45,
      "eval_steps_per_second": 4.716,
      "step": 2524
    },
    {
      "epoch": 102.22,
      "learning_rate": 5.401234567901234e-06,
      "loss": 0.0614,
      "step": 2530
    },
    {
      "epoch": 102.63,
      "learning_rate": 5.246913580246914e-06,
      "loss": 0.029,
      "step": 2540
    },
    {
      "epoch": 102.99,
      "eval_accuracy": 0.71875,
      "eval_loss": 1.9489786624908447,
      "eval_runtime": 0.4031,
      "eval_samples_per_second": 79.385,
      "eval_steps_per_second": 4.962,
      "step": 2549
    },
    {
      "epoch": 103.03,
      "learning_rate": 5.092592592592593e-06,
      "loss": 0.052,
      "step": 2550
    },
    {
      "epoch": 103.43,
      "learning_rate": 4.938271604938272e-06,
      "loss": 0.0696,
      "step": 2560
    },
    {
      "epoch": 103.84,
      "learning_rate": 4.783950617283951e-06,
      "loss": 0.0478,
      "step": 2570
    },
    {
      "epoch": 104.0,
      "eval_accuracy": 0.71875,
      "eval_loss": 1.9358038902282715,
      "eval_runtime": 0.4036,
      "eval_samples_per_second": 79.283,
      "eval_steps_per_second": 4.955,
      "step": 2574
    },
    {
      "epoch": 104.24,
      "learning_rate": 4.62962962962963e-06,
      "loss": 0.0358,
      "step": 2580
    },
    {
      "epoch": 104.65,
      "learning_rate": 4.475308641975309e-06,
      "loss": 0.0587,
      "step": 2590
    },
    {
      "epoch": 104.97,
      "eval_accuracy": 0.75,
      "eval_loss": 1.9197365045547485,
      "eval_runtime": 0.4296,
      "eval_samples_per_second": 74.488,
      "eval_steps_per_second": 4.655,
      "step": 2598
    },
    {
      "epoch": 105.05,
      "learning_rate": 4.3209876543209875e-06,
      "loss": 0.0417,
      "step": 2600
    },
    {
      "epoch": 105.45,
      "learning_rate": 4.166666666666667e-06,
      "loss": 0.0644,
      "step": 2610
    },
    {
      "epoch": 105.86,
      "learning_rate": 4.012345679012346e-06,
      "loss": 0.0523,
      "step": 2620
    },
    {
      "epoch": 105.98,
      "eval_accuracy": 0.71875,
      "eval_loss": 1.9309440851211548,
      "eval_runtime": 0.4621,
      "eval_samples_per_second": 69.248,
      "eval_steps_per_second": 4.328,
      "step": 2623
    },
    {
      "epoch": 106.26,
      "learning_rate": 3.858024691358025e-06,
      "loss": 0.0365,
      "step": 2630
    },
    {
      "epoch": 106.67,
      "learning_rate": 3.7037037037037037e-06,
      "loss": 0.0581,
      "step": 2640
    },
    {
      "epoch": 106.99,
      "eval_accuracy": 0.71875,
      "eval_loss": 1.9828894138336182,
      "eval_runtime": 0.4227,
      "eval_samples_per_second": 75.712,
      "eval_steps_per_second": 4.732,
      "step": 2648
    },
    {
      "epoch": 107.07,
      "learning_rate": 3.549382716049383e-06,
      "loss": 0.0459,
      "step": 2650
    },
    {
      "epoch": 107.47,
      "learning_rate": 3.395061728395062e-06,
      "loss": 0.031,
      "step": 2660
    },
    {
      "epoch": 107.88,
      "learning_rate": 3.240740740740741e-06,
      "loss": 0.0352,
      "step": 2670
    },
    {
      "epoch": 108.0,
      "eval_accuracy": 0.6875,
      "eval_loss": 2.00472354888916,
      "eval_runtime": 0.4442,
      "eval_samples_per_second": 72.042,
      "eval_steps_per_second": 4.503,
      "step": 2673
    },
    {
      "epoch": 108.28,
      "learning_rate": 3.08641975308642e-06,
      "loss": 0.0553,
      "step": 2680
    },
    {
      "epoch": 108.69,
      "learning_rate": 2.9320987654320994e-06,
      "loss": 0.0373,
      "step": 2690
    },
    {
      "epoch": 108.97,
      "eval_accuracy": 0.71875,
      "eval_loss": 1.9897255897521973,
      "eval_runtime": 0.4256,
      "eval_samples_per_second": 75.187,
      "eval_steps_per_second": 4.699,
      "step": 2697
    },
    {
      "epoch": 109.09,
      "learning_rate": 2.7777777777777783e-06,
      "loss": 0.049,
      "step": 2700
    },
    {
      "epoch": 109.49,
      "learning_rate": 2.623456790123457e-06,
      "loss": 0.0294,
      "step": 2710
    },
    {
      "epoch": 109.9,
      "learning_rate": 2.469135802469136e-06,
      "loss": 0.0258,
      "step": 2720
    },
    {
      "epoch": 109.98,
      "eval_accuracy": 0.71875,
      "eval_loss": 1.938352346420288,
      "eval_runtime": 0.4216,
      "eval_samples_per_second": 75.898,
      "eval_steps_per_second": 4.744,
      "step": 2722
    },
    {
      "epoch": 110.3,
      "learning_rate": 2.314814814814815e-06,
      "loss": 0.0484,
      "step": 2730
    },
    {
      "epoch": 110.71,
      "learning_rate": 2.1604938271604937e-06,
      "loss": 0.039,
      "step": 2740
    },
    {
      "epoch": 110.99,
      "eval_accuracy": 0.6875,
      "eval_loss": 1.9356164932250977,
      "eval_runtime": 0.4076,
      "eval_samples_per_second": 78.509,
      "eval_steps_per_second": 4.907,
      "step": 2747
    },
    {
      "epoch": 111.11,
      "learning_rate": 2.006172839506173e-06,
      "loss": 0.0326,
      "step": 2750
    },
    {
      "epoch": 111.52,
      "learning_rate": 1.8518518518518519e-06,
      "loss": 0.066,
      "step": 2760
    },
    {
      "epoch": 111.92,
      "learning_rate": 1.697530864197531e-06,
      "loss": 0.0333,
      "step": 2770
    },
    {
      "epoch": 112.0,
      "eval_accuracy": 0.71875,
      "eval_loss": 1.980454683303833,
      "eval_runtime": 0.4112,
      "eval_samples_per_second": 77.828,
      "eval_steps_per_second": 4.864,
      "step": 2772
    },
    {
      "epoch": 112.32,
      "learning_rate": 1.54320987654321e-06,
      "loss": 0.0524,
      "step": 2780
    },
    {
      "epoch": 112.73,
      "learning_rate": 1.3888888888888892e-06,
      "loss": 0.0641,
      "step": 2790
    },
    {
      "epoch": 112.97,
      "eval_accuracy": 0.6875,
      "eval_loss": 1.9813566207885742,
      "eval_runtime": 0.5541,
      "eval_samples_per_second": 57.748,
      "eval_steps_per_second": 3.609,
      "step": 2796
    },
    {
      "epoch": 113.13,
      "learning_rate": 1.234567901234568e-06,
      "loss": 0.0505,
      "step": 2800
    },
    {
      "epoch": 113.54,
      "learning_rate": 1.0802469135802469e-06,
      "loss": 0.0222,
      "step": 2810
    },
    {
      "epoch": 113.94,
      "learning_rate": 9.259259259259259e-07,
      "loss": 0.0649,
      "step": 2820
    },
    {
      "epoch": 113.98,
      "eval_accuracy": 0.6875,
      "eval_loss": 1.972588062286377,
      "eval_runtime": 0.4431,
      "eval_samples_per_second": 72.211,
      "eval_steps_per_second": 4.513,
      "step": 2821
    },
    {
      "epoch": 114.34,
      "learning_rate": 7.71604938271605e-07,
      "loss": 0.0354,
      "step": 2830
    },
    {
      "epoch": 114.75,
      "learning_rate": 6.17283950617284e-07,
      "loss": 0.0241,
      "step": 2840
    },
    {
      "epoch": 114.99,
      "eval_accuracy": 0.6875,
      "eval_loss": 1.9736533164978027,
      "eval_runtime": 0.4011,
      "eval_samples_per_second": 79.777,
      "eval_steps_per_second": 4.986,
      "step": 2846
    },
    {
      "epoch": 115.15,
      "learning_rate": 4.6296296296296297e-07,
      "loss": 0.0418,
      "step": 2850
    },
    {
      "epoch": 115.56,
      "learning_rate": 3.08641975308642e-07,
      "loss": 0.0382,
      "step": 2860
    },
    {
      "epoch": 115.96,
      "learning_rate": 1.54320987654321e-07,
      "loss": 0.0356,
      "step": 2870
    },
    {
      "epoch": 116.0,
      "eval_accuracy": 0.6875,
      "eval_loss": 1.985576868057251,
      "eval_runtime": 0.4721,
      "eval_samples_per_second": 67.78,
      "eval_steps_per_second": 4.236,
      "step": 2871
    },
    {
      "epoch": 116.36,
      "learning_rate": 0.0,
      "loss": 0.0601,
      "step": 2880
    },
    {
      "epoch": 116.36,
      "eval_accuracy": 0.6875,
      "eval_loss": 1.9852969646453857,
      "eval_runtime": 0.4181,
      "eval_samples_per_second": 76.535,
      "eval_steps_per_second": 4.783,
      "step": 2880
    },
    {
      "epoch": 116.36,
      "step": 2880,
      "total_flos": 5.970642494234296e+18,
      "train_loss": 0.4532645735475752,
      "train_runtime": 3508.1723,
      "train_samples_per_second": 53.943,
      "train_steps_per_second": 0.821
    }
  ],
  "logging_steps": 10,
  "max_steps": 2880,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 120,
  "save_steps": 500,
  "total_flos": 5.970642494234296e+18,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": null
}