{
  "best_metric": 0.9375,
  "best_model_checkpoint": "./resnet50/checkpoint-2436",
  "epoch": 30.0,
  "eval_steps": 500,
  "global_step": 2520,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.11904761904761904,
      "grad_norm": 0.48410382866859436,
      "learning_rate": 7.936507936507937e-07,
      "loss": 1.3866,
      "step": 10
    },
    {
      "epoch": 0.23809523809523808,
      "grad_norm": 0.4765987694263458,
      "learning_rate": 1.5873015873015873e-06,
      "loss": 1.3861,
      "step": 20
    },
    {
      "epoch": 0.35714285714285715,
      "grad_norm": 0.43654772639274597,
      "learning_rate": 2.380952380952381e-06,
      "loss": 1.3859,
      "step": 30
    },
    {
      "epoch": 0.47619047619047616,
      "grad_norm": 0.5733897686004639,
      "learning_rate": 3.1746031746031746e-06,
      "loss": 1.3855,
      "step": 40
    },
    {
      "epoch": 0.5952380952380952,
      "grad_norm": 0.457210898399353,
      "learning_rate": 3.968253968253968e-06,
      "loss": 1.385,
      "step": 50
    },
    {
      "epoch": 0.7142857142857143,
      "grad_norm": 0.5107349157333374,
      "learning_rate": 4.761904761904762e-06,
      "loss": 1.3843,
      "step": 60
    },
    {
      "epoch": 0.8333333333333334,
      "grad_norm": 0.6200265288352966,
      "learning_rate": 5.555555555555557e-06,
      "loss": 1.3832,
      "step": 70
    },
    {
      "epoch": 0.9523809523809523,
      "grad_norm": 0.6969241499900818,
      "learning_rate": 6.349206349206349e-06,
      "loss": 1.382,
      "step": 80
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.58375,
      "eval_loss": 1.3806676864624023,
      "eval_runtime": 5.5764,
      "eval_samples_per_second": 143.462,
      "eval_steps_per_second": 17.933,
      "step": 84
    },
    {
      "epoch": 1.0714285714285714,
      "grad_norm": 0.616682231426239,
      "learning_rate": 7.1428571428571436e-06,
      "loss": 1.3805,
      "step": 90
    },
    {
      "epoch": 1.1904761904761905,
      "grad_norm": 0.774887204170227,
      "learning_rate": 7.936507936507936e-06,
      "loss": 1.3786,
      "step": 100
    },
    {
      "epoch": 1.3095238095238095,
      "grad_norm": 0.862555205821991,
      "learning_rate": 8.730158730158731e-06,
      "loss": 1.3754,
      "step": 110
    },
    {
      "epoch": 1.4285714285714286,
      "grad_norm": 0.7880147695541382,
      "learning_rate": 9.523809523809525e-06,
      "loss": 1.3714,
      "step": 120
    },
    {
      "epoch": 1.5476190476190477,
      "grad_norm": 0.9495492577552795,
      "learning_rate": 1.031746031746032e-05,
      "loss": 1.3663,
      "step": 130
    },
    {
      "epoch": 1.6666666666666665,
      "grad_norm": 1.0720769166946411,
      "learning_rate": 1.1111111111111113e-05,
      "loss": 1.3595,
      "step": 140
    },
    {
      "epoch": 1.7857142857142856,
      "grad_norm": 1.4009815454483032,
      "learning_rate": 1.1904761904761905e-05,
      "loss": 1.3506,
      "step": 150
    },
    {
      "epoch": 1.9047619047619047,
      "grad_norm": 1.4374886751174927,
      "learning_rate": 1.2698412698412699e-05,
      "loss": 1.3401,
      "step": 160
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.44875,
      "eval_loss": 1.31219482421875,
      "eval_runtime": 5.4479,
      "eval_samples_per_second": 146.845,
      "eval_steps_per_second": 18.356,
      "step": 168
    },
    {
      "epoch": 2.0238095238095237,
      "grad_norm": 1.4388487339019775,
      "learning_rate": 1.3492063492063494e-05,
      "loss": 1.3256,
      "step": 170
    },
    {
      "epoch": 2.142857142857143,
      "grad_norm": 2.0602617263793945,
      "learning_rate": 1.4285714285714287e-05,
      "loss": 1.3085,
      "step": 180
    },
    {
      "epoch": 2.261904761904762,
      "grad_norm": 2.5268349647521973,
      "learning_rate": 1.507936507936508e-05,
      "loss": 1.2787,
      "step": 190
    },
    {
      "epoch": 2.380952380952381,
      "grad_norm": 2.7516028881073,
      "learning_rate": 1.5873015873015872e-05,
      "loss": 1.2603,
      "step": 200
    },
    {
      "epoch": 2.5,
      "grad_norm": 2.6878774166107178,
      "learning_rate": 1.6666666666666667e-05,
      "loss": 1.2355,
      "step": 210
    },
    {
      "epoch": 2.619047619047619,
      "grad_norm": 1.6872291564941406,
      "learning_rate": 1.7460317460317463e-05,
      "loss": 1.201,
      "step": 220
    },
    {
      "epoch": 2.738095238095238,
      "grad_norm": 1.6558163166046143,
      "learning_rate": 1.8253968253968254e-05,
      "loss": 1.18,
      "step": 230
    },
    {
      "epoch": 2.857142857142857,
      "grad_norm": 2.9151134490966797,
      "learning_rate": 1.904761904761905e-05,
      "loss": 1.1485,
      "step": 240
    },
    {
      "epoch": 2.9761904761904763,
      "grad_norm": 1.5142790079116821,
      "learning_rate": 1.9841269841269845e-05,
      "loss": 1.1534,
      "step": 250
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.5,
      "eval_loss": 1.121809482574463,
      "eval_runtime": 5.3391,
      "eval_samples_per_second": 149.838,
      "eval_steps_per_second": 18.73,
      "step": 252
    },
    {
      "epoch": 3.0952380952380953,
      "grad_norm": 1.396998643875122,
      "learning_rate": 1.9938271604938272e-05,
      "loss": 1.1074,
      "step": 260
    },
    {
      "epoch": 3.2142857142857144,
      "grad_norm": 1.8247098922729492,
      "learning_rate": 1.9850088183421517e-05,
      "loss": 1.1146,
      "step": 270
    },
    {
      "epoch": 3.3333333333333335,
      "grad_norm": 3.359285831451416,
      "learning_rate": 1.9761904761904763e-05,
      "loss": 1.0917,
      "step": 280
    },
    {
      "epoch": 3.4523809523809526,
      "grad_norm": 2.1544318199157715,
      "learning_rate": 1.9673721340388008e-05,
      "loss": 1.0922,
      "step": 290
    },
    {
      "epoch": 3.571428571428571,
      "grad_norm": 1.6087664365768433,
      "learning_rate": 1.9585537918871253e-05,
      "loss": 1.076,
      "step": 300
    },
    {
      "epoch": 3.6904761904761907,
      "grad_norm": 2.3240392208099365,
      "learning_rate": 1.94973544973545e-05,
      "loss": 1.0453,
      "step": 310
    },
    {
      "epoch": 3.8095238095238093,
      "grad_norm": 2.1408257484436035,
      "learning_rate": 1.9409171075837744e-05,
      "loss": 1.0366,
      "step": 320
    },
    {
      "epoch": 3.928571428571429,
      "grad_norm": 4.0489583015441895,
      "learning_rate": 1.932098765432099e-05,
      "loss": 1.0343,
      "step": 330
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.67625,
      "eval_loss": 1.0234503746032715,
      "eval_runtime": 5.3529,
      "eval_samples_per_second": 149.453,
      "eval_steps_per_second": 18.682,
      "step": 336
    },
    {
      "epoch": 4.0476190476190474,
      "grad_norm": 3.3416860103607178,
      "learning_rate": 1.9232804232804235e-05,
      "loss": 1.0142,
      "step": 340
    },
    {
      "epoch": 4.166666666666667,
      "grad_norm": 2.0407114028930664,
      "learning_rate": 1.914462081128748e-05,
      "loss": 0.9883,
      "step": 350
    },
    {
      "epoch": 4.285714285714286,
      "grad_norm": 1.7006970643997192,
      "learning_rate": 1.9056437389770726e-05,
      "loss": 0.9741,
      "step": 360
    },
    {
      "epoch": 4.404761904761905,
      "grad_norm": 2.021171808242798,
      "learning_rate": 1.8968253968253968e-05,
      "loss": 0.9754,
      "step": 370
    },
    {
      "epoch": 4.523809523809524,
      "grad_norm": 4.418124675750732,
      "learning_rate": 1.8880070546737216e-05,
      "loss": 0.9428,
      "step": 380
    },
    {
      "epoch": 4.642857142857143,
      "grad_norm": 2.6121156215667725,
      "learning_rate": 1.8791887125220462e-05,
      "loss": 0.9401,
      "step": 390
    },
    {
      "epoch": 4.761904761904762,
      "grad_norm": 2.1322250366210938,
      "learning_rate": 1.8703703703703707e-05,
      "loss": 0.9409,
      "step": 400
    },
    {
      "epoch": 4.880952380952381,
      "grad_norm": 1.571338415145874,
      "learning_rate": 1.861552028218695e-05,
      "loss": 0.9062,
      "step": 410
    },
    {
      "epoch": 5.0,
      "grad_norm": Infinity,
      "learning_rate": 1.853615520282187e-05,
      "loss": 0.9117,
      "step": 420
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.7175,
      "eval_loss": 0.9020006656646729,
      "eval_runtime": 5.4379,
      "eval_samples_per_second": 147.116,
      "eval_steps_per_second": 18.389,
      "step": 420
    },
    {
      "epoch": 5.119047619047619,
      "grad_norm": 4.1433868408203125,
      "learning_rate": 1.8447971781305116e-05,
      "loss": 0.8893,
      "step": 430
    },
    {
      "epoch": 5.238095238095238,
      "grad_norm": 2.4193081855773926,
      "learning_rate": 1.835978835978836e-05,
      "loss": 0.8883,
      "step": 440
    },
    {
      "epoch": 5.357142857142857,
      "grad_norm": 3.0261456966400146,
      "learning_rate": 1.8271604938271607e-05,
      "loss": 0.8375,
      "step": 450
    },
    {
      "epoch": 5.476190476190476,
      "grad_norm": 4.649360656738281,
      "learning_rate": 1.8183421516754852e-05,
      "loss": 0.8617,
      "step": 460
    },
    {
      "epoch": 5.595238095238095,
      "grad_norm": 2.217705011367798,
      "learning_rate": 1.8095238095238097e-05,
      "loss": 0.8377,
      "step": 470
    },
    {
      "epoch": 5.714285714285714,
      "grad_norm": 3.888078212738037,
      "learning_rate": 1.8007054673721343e-05,
      "loss": 0.8027,
      "step": 480
    },
    {
      "epoch": 5.833333333333333,
      "grad_norm": 3.240481376647949,
      "learning_rate": 1.7918871252204585e-05,
      "loss": 0.8258,
      "step": 490
    },
    {
      "epoch": 5.9523809523809526,
      "grad_norm": 2.535261631011963,
      "learning_rate": 1.783068783068783e-05,
      "loss": 0.8169,
      "step": 500
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.7925,
      "eval_loss": 0.8129910230636597,
      "eval_runtime": 5.2312,
      "eval_samples_per_second": 152.927,
      "eval_steps_per_second": 19.116,
      "step": 504
    },
    {
      "epoch": 6.071428571428571,
      "grad_norm": 3.822288751602173,
      "learning_rate": 1.774250440917108e-05,
      "loss": 0.7932,
      "step": 510
    },
    {
      "epoch": 6.190476190476191,
      "grad_norm": 2.448084592819214,
      "learning_rate": 1.7654320987654324e-05,
      "loss": 0.7807,
      "step": 520
    },
    {
      "epoch": 6.309523809523809,
      "grad_norm": 4.276307106018066,
      "learning_rate": 1.7566137566137566e-05,
      "loss": 0.7736,
      "step": 530
    },
    {
      "epoch": 6.428571428571429,
      "grad_norm": 3.5429012775421143,
      "learning_rate": 1.747795414462081e-05,
      "loss": 0.7453,
      "step": 540
    },
    {
      "epoch": 6.5476190476190474,
      "grad_norm": 2.4782090187072754,
      "learning_rate": 1.7389770723104057e-05,
      "loss": 0.7327,
      "step": 550
    },
    {
      "epoch": 6.666666666666667,
      "grad_norm": 5.342073440551758,
      "learning_rate": 1.7301587301587302e-05,
      "loss": 0.7247,
      "step": 560
    },
    {
      "epoch": 6.785714285714286,
      "grad_norm": 10.06567668914795,
      "learning_rate": 1.7213403880070548e-05,
      "loss": 0.7088,
      "step": 570
    },
    {
      "epoch": 6.904761904761905,
      "grad_norm": 3.217452049255371,
      "learning_rate": 1.713403880070547e-05,
      "loss": 0.7058,
      "step": 580
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.815,
      "eval_loss": 0.7232482433319092,
      "eval_runtime": 5.4248,
      "eval_samples_per_second": 147.471,
      "eval_steps_per_second": 18.434,
      "step": 588
    },
    {
      "epoch": 7.023809523809524,
      "grad_norm": 4.111032009124756,
      "learning_rate": 1.7045855379188714e-05,
      "loss": 0.6987,
      "step": 590
    },
    {
      "epoch": 7.142857142857143,
      "grad_norm": 2.667541027069092,
      "learning_rate": 1.695767195767196e-05,
      "loss": 0.684,
      "step": 600
    },
    {
      "epoch": 7.261904761904762,
      "grad_norm": 8.819822311401367,
      "learning_rate": 1.6869488536155205e-05,
      "loss": 0.674,
      "step": 610
    },
    {
      "epoch": 7.380952380952381,
      "grad_norm": 3.9701991081237793,
      "learning_rate": 1.6781305114638447e-05,
      "loss": 0.6199,
      "step": 620
    },
    {
      "epoch": 7.5,
      "grad_norm": 8.835700035095215,
      "learning_rate": 1.6693121693121696e-05,
      "loss": 0.6241,
      "step": 630
    },
    {
      "epoch": 7.619047619047619,
      "grad_norm": 3.7322373390197754,
      "learning_rate": 1.660493827160494e-05,
      "loss": 0.6087,
      "step": 640
    },
    {
      "epoch": 7.738095238095238,
      "grad_norm": 4.442748069763184,
      "learning_rate": 1.6516754850088187e-05,
      "loss": 0.589,
      "step": 650
    },
    {
      "epoch": 7.857142857142857,
      "grad_norm": 3.6890928745269775,
      "learning_rate": 1.642857142857143e-05,
      "loss": 0.5804,
      "step": 660
    },
    {
      "epoch": 7.976190476190476,
      "grad_norm": 7.095566749572754,
      "learning_rate": 1.6340388007054674e-05,
      "loss": 0.5556,
      "step": 670
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.83375,
      "eval_loss": 0.58147794008255,
      "eval_runtime": 6.0471,
      "eval_samples_per_second": 132.295,
      "eval_steps_per_second": 16.537,
      "step": 672
    },
    {
      "epoch": 8.095238095238095,
      "grad_norm": 4.2159318923950195,
      "learning_rate": 1.625220458553792e-05,
      "loss": 0.577,
      "step": 680
    },
    {
      "epoch": 8.214285714285714,
      "grad_norm": 3.634282350540161,
      "learning_rate": 1.6164021164021168e-05,
      "loss": 0.5226,
      "step": 690
    },
    {
      "epoch": 8.333333333333334,
      "grad_norm": 3.7204463481903076,
      "learning_rate": 1.607583774250441e-05,
      "loss": 0.5397,
      "step": 700
    },
    {
      "epoch": 8.452380952380953,
      "grad_norm": 8.489058494567871,
      "learning_rate": 1.5987654320987655e-05,
      "loss": 0.515,
      "step": 710
    },
    {
      "epoch": 8.571428571428571,
      "grad_norm": 8.01997184753418,
      "learning_rate": 1.58994708994709e-05,
      "loss": 0.5017,
      "step": 720
    },
    {
      "epoch": 8.69047619047619,
      "grad_norm": 5.835849761962891,
      "learning_rate": 1.5811287477954146e-05,
      "loss": 0.5008,
      "step": 730
    },
    {
      "epoch": 8.80952380952381,
      "grad_norm": 10.542766571044922,
      "learning_rate": 1.572310405643739e-05,
      "loss": 0.476,
      "step": 740
    },
    {
      "epoch": 8.928571428571429,
      "grad_norm": 7.684937953948975,
      "learning_rate": 1.5634920634920637e-05,
      "loss": 0.4527,
      "step": 750
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.86625,
      "eval_loss": 0.4813559651374817,
      "eval_runtime": 6.5617,
      "eval_samples_per_second": 121.92,
      "eval_steps_per_second": 15.24,
      "step": 756
    },
    {
      "epoch": 9.047619047619047,
      "grad_norm": 6.5577569007873535,
      "learning_rate": 1.5546737213403882e-05,
      "loss": 0.4875,
      "step": 760
    },
    {
      "epoch": 9.166666666666666,
      "grad_norm": 6.667921543121338,
      "learning_rate": 1.5458553791887128e-05,
      "loss": 0.4581,
      "step": 770
    },
    {
      "epoch": 9.285714285714286,
      "grad_norm": 5.33371114730835,
      "learning_rate": 1.537037037037037e-05,
      "loss": 0.4269,
      "step": 780
    },
    {
      "epoch": 9.404761904761905,
      "grad_norm": 4.12121057510376,
      "learning_rate": 1.5282186948853618e-05,
      "loss": 0.4173,
      "step": 790
    },
    {
      "epoch": 9.523809523809524,
      "grad_norm": 7.042411804199219,
      "learning_rate": 1.5194003527336862e-05,
      "loss": 0.4312,
      "step": 800
    },
    {
      "epoch": 9.642857142857142,
      "grad_norm": 6.229002952575684,
      "learning_rate": 1.5105820105820109e-05,
      "loss": 0.421,
      "step": 810
    },
    {
      "epoch": 9.761904761904763,
      "grad_norm": 9.01046085357666,
      "learning_rate": 1.5017636684303351e-05,
      "loss": 0.4182,
      "step": 820
    },
    {
      "epoch": 9.880952380952381,
      "grad_norm": 9.333113670349121,
      "learning_rate": 1.4929453262786598e-05,
      "loss": 0.4065,
      "step": 830
    },
    {
      "epoch": 10.0,
      "grad_norm": 5.7903876304626465,
      "learning_rate": 1.4841269841269843e-05,
      "loss": 0.3994,
      "step": 840
    },
    {
      "epoch": 10.0,
      "eval_accuracy": 0.86875,
      "eval_loss": 0.43535658717155457,
      "eval_runtime": 6.424,
      "eval_samples_per_second": 124.534,
      "eval_steps_per_second": 15.567,
      "step": 840
    },
    {
      "epoch": 10.119047619047619,
      "grad_norm": 4.672136306762695,
      "learning_rate": 1.4753086419753087e-05,
      "loss": 0.3692,
      "step": 850
    },
    {
      "epoch": 10.238095238095237,
      "grad_norm": 9.825118064880371,
      "learning_rate": 1.4664902998236332e-05,
      "loss": 0.3946,
      "step": 860
    },
    {
      "epoch": 10.357142857142858,
      "grad_norm": 10.594053268432617,
      "learning_rate": 1.4576719576719578e-05,
      "loss": 0.3836,
      "step": 870
    },
    {
      "epoch": 10.476190476190476,
      "grad_norm": 6.456384658813477,
      "learning_rate": 1.4488536155202823e-05,
      "loss": 0.3922,
      "step": 880
    },
    {
      "epoch": 10.595238095238095,
      "grad_norm": 7.883987903594971,
      "learning_rate": 1.4400352733686067e-05,
      "loss": 0.3767,
      "step": 890
    },
    {
      "epoch": 10.714285714285714,
      "grad_norm": 3.752121686935425,
      "learning_rate": 1.4312169312169312e-05,
      "loss": 0.3731,
      "step": 900
    },
    {
      "epoch": 10.833333333333334,
      "grad_norm": 5.878788948059082,
      "learning_rate": 1.422398589065256e-05,
      "loss": 0.3581,
      "step": 910
    },
    {
      "epoch": 10.952380952380953,
      "grad_norm": 10.106928825378418,
      "learning_rate": 1.4135802469135805e-05,
      "loss": 0.3399,
      "step": 920
    },
    {
      "epoch": 11.0,
      "eval_accuracy": 0.89,
      "eval_loss": 0.3747052252292633,
      "eval_runtime": 6.1766,
      "eval_samples_per_second": 129.52,
      "eval_steps_per_second": 16.19,
      "step": 924
    },
    {
      "epoch": 11.071428571428571,
      "grad_norm": 13.221107482910156,
      "learning_rate": 1.4047619047619048e-05,
      "loss": 0.3512,
      "step": 930
    },
    {
      "epoch": 11.19047619047619,
      "grad_norm": 5.013436317443848,
      "learning_rate": 1.3959435626102294e-05,
      "loss": 0.3217,
      "step": 940
    },
    {
      "epoch": 11.30952380952381,
      "grad_norm": 4.782650470733643,
      "learning_rate": 1.3871252204585539e-05,
      "loss": 0.3603,
      "step": 950
    },
    {
      "epoch": 11.428571428571429,
      "grad_norm": 7.278000354766846,
      "learning_rate": 1.3783068783068784e-05,
      "loss": 0.3193,
      "step": 960
    },
    {
      "epoch": 11.547619047619047,
      "grad_norm": 13.20319652557373,
      "learning_rate": 1.3694885361552028e-05,
      "loss": 0.3429,
      "step": 970
    },
    {
      "epoch": 11.666666666666666,
      "grad_norm": 10.208937644958496,
      "learning_rate": 1.3606701940035273e-05,
      "loss": 0.3391,
      "step": 980
    },
    {
      "epoch": 11.785714285714286,
      "grad_norm": 8.55504322052002,
      "learning_rate": 1.351851851851852e-05,
      "loss": 0.3286,
      "step": 990
    },
    {
      "epoch": 11.904761904761905,
      "grad_norm": 9.174301147460938,
      "learning_rate": 1.3430335097001766e-05,
      "loss": 0.3157,
      "step": 1000
    },
    {
      "epoch": 12.0,
      "eval_accuracy": 0.895,
      "eval_loss": 0.33855053782463074,
      "eval_runtime": 5.5685,
      "eval_samples_per_second": 143.664,
      "eval_steps_per_second": 17.958,
      "step": 1008
    },
    {
      "epoch": 12.023809523809524,
      "grad_norm": 9.196775436401367,
      "learning_rate": 1.334215167548501e-05,
      "loss": 0.32,
      "step": 1010
    },
    {
      "epoch": 12.142857142857142,
      "grad_norm": 7.491143703460693,
      "learning_rate": 1.3253968253968255e-05,
      "loss": 0.3211,
      "step": 1020
    },
    {
      "epoch": 12.261904761904763,
      "grad_norm": 10.411942481994629,
      "learning_rate": 1.31657848324515e-05,
      "loss": 0.31,
      "step": 1030
    },
    {
      "epoch": 12.380952380952381,
      "grad_norm": 7.7979021072387695,
      "learning_rate": 1.3077601410934746e-05,
      "loss": 0.301,
      "step": 1040
    },
    {
      "epoch": 12.5,
      "grad_norm": 16.540781021118164,
      "learning_rate": 1.298941798941799e-05,
      "loss": 0.2879,
      "step": 1050
    },
    {
      "epoch": 12.619047619047619,
      "grad_norm": 8.34303092956543,
      "learning_rate": 1.2901234567901235e-05,
      "loss": 0.2837,
      "step": 1060
    },
    {
      "epoch": 12.738095238095237,
      "grad_norm": 5.1471405029296875,
      "learning_rate": 1.2813051146384482e-05,
      "loss": 0.2943,
      "step": 1070
    },
    {
      "epoch": 12.857142857142858,
      "grad_norm": 12.514649391174316,
      "learning_rate": 1.2724867724867727e-05,
      "loss": 0.3033,
      "step": 1080
    },
    {
      "epoch": 12.976190476190476,
      "grad_norm": 14.337882041931152,
      "learning_rate": 1.263668430335097e-05,
      "loss": 0.3094,
      "step": 1090
    },
    {
      "epoch": 13.0,
      "eval_accuracy": 0.90375,
      "eval_loss": 0.3166072368621826,
      "eval_runtime": 5.6029,
      "eval_samples_per_second": 142.783,
      "eval_steps_per_second": 17.848,
      "step": 1092
    },
    {
      "epoch": 13.095238095238095,
      "grad_norm": 7.664865493774414,
      "learning_rate": 1.2548500881834216e-05,
      "loss": 0.3027,
      "step": 1100
    },
    {
      "epoch": 13.214285714285714,
      "grad_norm": 7.5357489585876465,
      "learning_rate": 1.2460317460317461e-05,
      "loss": 0.303,
      "step": 1110
    },
    {
      "epoch": 13.333333333333334,
      "grad_norm": 10.726656913757324,
      "learning_rate": 1.2372134038800707e-05,
      "loss": 0.2897,
      "step": 1120
    },
    {
      "epoch": 13.452380952380953,
      "grad_norm": 12.0068359375,
      "learning_rate": 1.228395061728395e-05,
      "loss": 0.2949,
      "step": 1130
    },
    {
      "epoch": 13.571428571428571,
      "grad_norm": 9.883105278015137,
      "learning_rate": 1.2195767195767196e-05,
      "loss": 0.2573,
      "step": 1140
    },
    {
      "epoch": 13.69047619047619,
      "grad_norm": 9.254776000976562,
      "learning_rate": 1.2107583774250443e-05,
      "loss": 0.2838,
      "step": 1150
    },
    {
      "epoch": 13.80952380952381,
      "grad_norm": 6.0406880378723145,
      "learning_rate": 1.2019400352733688e-05,
      "loss": 0.2558,
      "step": 1160
    },
    {
      "epoch": 13.928571428571429,
      "grad_norm": 6.664582252502441,
      "learning_rate": 1.1931216931216932e-05,
      "loss": 0.2839,
      "step": 1170
    },
    {
      "epoch": 14.0,
      "eval_accuracy": 0.90125,
      "eval_loss": 0.31680428981781006,
      "eval_runtime": 5.4208,
      "eval_samples_per_second": 147.579,
      "eval_steps_per_second": 18.447,
      "step": 1176
    },
    {
      "epoch": 14.047619047619047,
      "grad_norm": 11.964251518249512,
      "learning_rate": 1.1843033509700177e-05,
      "loss": 0.2688,
      "step": 1180
    },
    {
      "epoch": 14.166666666666666,
      "grad_norm": 6.548354148864746,
      "learning_rate": 1.1754850088183423e-05,
      "loss": 0.2545,
      "step": 1190
    },
    {
      "epoch": 14.285714285714286,
      "grad_norm": 14.916826248168945,
      "learning_rate": 1.1666666666666668e-05,
      "loss": 0.2618,
      "step": 1200
    },
    {
      "epoch": 14.404761904761905,
      "grad_norm": 15.267316818237305,
      "learning_rate": 1.1578483245149912e-05,
      "loss": 0.2442,
      "step": 1210
    },
    {
      "epoch": 14.523809523809524,
      "grad_norm": 8.463184356689453,
      "learning_rate": 1.1490299823633157e-05,
      "loss": 0.2618,
      "step": 1220
    },
    {
      "epoch": 14.642857142857142,
      "grad_norm": 13.68342113494873,
      "learning_rate": 1.1402116402116404e-05,
      "loss": 0.2447,
      "step": 1230
    },
    {
      "epoch": 14.761904761904763,
      "grad_norm": 4.8457112312316895,
      "learning_rate": 1.131393298059965e-05,
      "loss": 0.2361,
      "step": 1240
    },
    {
      "epoch": 14.880952380952381,
      "grad_norm": 6.4722185134887695,
      "learning_rate": 1.1225749559082893e-05,
      "loss": 0.2519,
      "step": 1250
    },
    {
      "epoch": 15.0,
      "grad_norm": 15.844517707824707,
      "learning_rate": 1.1137566137566138e-05,
      "loss": 0.2658,
      "step": 1260
    },
    {
      "epoch": 15.0,
      "eval_accuracy": 0.91,
      "eval_loss": 0.2803078591823578,
      "eval_runtime": 5.3029,
      "eval_samples_per_second": 150.862,
      "eval_steps_per_second": 18.858,
      "step": 1260
    },
    {
      "epoch": 15.119047619047619,
      "grad_norm": 20.211238861083984,
      "learning_rate": 1.1049382716049384e-05,
      "loss": 0.2386,
      "step": 1270
    },
    {
      "epoch": 15.238095238095237,
      "grad_norm": 6.853188991546631,
      "learning_rate": 1.0961199294532629e-05,
      "loss": 0.2326,
      "step": 1280
    },
    {
      "epoch": 15.357142857142858,
      "grad_norm": 8.602685928344727,
      "learning_rate": 1.0873015873015873e-05,
      "loss": 0.2553,
      "step": 1290
    },
    {
      "epoch": 15.476190476190476,
      "grad_norm": 4.291443824768066,
      "learning_rate": 1.0784832451499118e-05,
      "loss": 0.2434,
      "step": 1300
    },
    {
      "epoch": 15.595238095238095,
      "grad_norm": 6.306275367736816,
      "learning_rate": 1.0696649029982365e-05,
      "loss": 0.2277,
      "step": 1310
    },
    {
      "epoch": 15.714285714285714,
      "grad_norm": 5.804137706756592,
      "learning_rate": 1.060846560846561e-05,
      "loss": 0.2248,
      "step": 1320
    },
    {
      "epoch": 15.833333333333334,
      "grad_norm": 13.754738807678223,
      "learning_rate": 1.0520282186948854e-05,
      "loss": 0.2313,
      "step": 1330
    },
    {
      "epoch": 15.952380952380953,
      "grad_norm": 8.410860061645508,
      "learning_rate": 1.04320987654321e-05,
      "loss": 0.2331,
      "step": 1340
    },
    {
      "epoch": 16.0,
      "eval_accuracy": 0.9125,
      "eval_loss": 0.26975059509277344,
      "eval_runtime": 5.4078,
      "eval_samples_per_second": 147.934,
      "eval_steps_per_second": 18.492,
      "step": 1344
    },
    {
      "epoch": 16.071428571428573,
      "grad_norm": 9.222555160522461,
      "learning_rate": 1.0352733686067021e-05,
      "loss": 0.2394,
      "step": 1350
    },
    {
      "epoch": 16.19047619047619,
      "grad_norm": 9.493171691894531,
      "learning_rate": 1.0264550264550266e-05,
      "loss": 0.2213,
      "step": 1360
    },
    {
      "epoch": 16.30952380952381,
      "grad_norm": 4.863091945648193,
      "learning_rate": 1.0176366843033512e-05,
      "loss": 0.2447,
      "step": 1370
    },
    {
      "epoch": 16.428571428571427,
      "grad_norm": 9.915203094482422,
      "learning_rate": 1.0088183421516755e-05,
      "loss": 0.2369,
      "step": 1380
    },
    {
      "epoch": 16.547619047619047,
      "grad_norm": 7.485960960388184,
      "learning_rate": 1e-05,
      "loss": 0.213,
      "step": 1390
    },
    {
      "epoch": 16.666666666666668,
      "grad_norm": 5.043095111846924,
      "learning_rate": 9.911816578483246e-06,
      "loss": 0.2035,
      "step": 1400
    },
    {
      "epoch": 16.785714285714285,
      "grad_norm": 12.636078834533691,
      "learning_rate": 9.823633156966492e-06,
      "loss": 0.2312,
      "step": 1410
    },
    {
      "epoch": 16.904761904761905,
      "grad_norm": 13.119421005249023,
      "learning_rate": 9.735449735449735e-06,
      "loss": 0.2271,
      "step": 1420
    },
    {
      "epoch": 17.0,
      "eval_accuracy": 0.90625,
      "eval_loss": 0.2662568986415863,
      "eval_runtime": 5.5739,
      "eval_samples_per_second": 143.526,
      "eval_steps_per_second": 17.941,
      "step": 1428
    },
    {
      "epoch": 17.023809523809526,
      "grad_norm": 19.145736694335938,
      "learning_rate": 9.64726631393298e-06,
      "loss": 0.2089,
      "step": 1430
    },
    {
      "epoch": 17.142857142857142,
      "grad_norm": 13.143487930297852,
      "learning_rate": 9.559082892416226e-06,
      "loss": 0.2036,
      "step": 1440
    },
    {
      "epoch": 17.261904761904763,
      "grad_norm": 13.245809555053711,
      "learning_rate": 9.470899470899471e-06,
      "loss": 0.1902,
      "step": 1450
    },
    {
      "epoch": 17.38095238095238,
      "grad_norm": 15.815811157226562,
      "learning_rate": 9.382716049382717e-06,
      "loss": 0.2234,
      "step": 1460
    },
    {
      "epoch": 17.5,
      "grad_norm": 8.950321197509766,
      "learning_rate": 9.294532627865962e-06,
      "loss": 0.2258,
      "step": 1470
    },
    {
      "epoch": 17.61904761904762,
      "grad_norm": 6.937575817108154,
      "learning_rate": 9.206349206349207e-06,
      "loss": 0.212,
      "step": 1480
    },
    {
      "epoch": 17.738095238095237,
      "grad_norm": 9.044116020202637,
      "learning_rate": 9.118165784832453e-06,
      "loss": 0.2084,
      "step": 1490
    },
    {
      "epoch": 17.857142857142858,
      "grad_norm": 9.224007606506348,
      "learning_rate": 9.029982363315696e-06,
      "loss": 0.2036,
      "step": 1500
    },
    {
      "epoch": 17.976190476190474,
      "grad_norm": 5.575143814086914,
      "learning_rate": 8.941798941798942e-06,
      "loss": 0.1895,
      "step": 1510
    },
    {
      "epoch": 18.0,
      "eval_accuracy": 0.92375,
      "eval_loss": 0.2639971375465393,
      "eval_runtime": 5.9356,
      "eval_samples_per_second": 134.779,
      "eval_steps_per_second": 16.847,
      "step": 1512
    },
    {
      "epoch": 18.095238095238095,
      "grad_norm": 10.496304512023926,
      "learning_rate": 8.853615520282187e-06,
      "loss": 0.1908,
      "step": 1520
    },
    {
      "epoch": 18.214285714285715,
      "grad_norm": 8.395342826843262,
      "learning_rate": 8.765432098765432e-06,
      "loss": 0.202,
      "step": 1530
    },
    {
      "epoch": 18.333333333333332,
      "grad_norm": 6.879480838775635,
      "learning_rate": 8.677248677248678e-06,
      "loss": 0.1825,
      "step": 1540
    },
    {
      "epoch": 18.452380952380953,
      "grad_norm": 6.519415855407715,
      "learning_rate": 8.589065255731923e-06,
      "loss": 0.2025,
      "step": 1550
    },
    {
      "epoch": 18.571428571428573,
      "grad_norm": 14.108360290527344,
      "learning_rate": 8.500881834215169e-06,
      "loss": 0.2156,
      "step": 1560
    },
    {
      "epoch": 18.69047619047619,
      "grad_norm": 8.98913288116455,
      "learning_rate": 8.412698412698414e-06,
      "loss": 0.1895,
      "step": 1570
    },
    {
      "epoch": 18.80952380952381,
      "grad_norm": 7.803181171417236,
      "learning_rate": 8.324514991181658e-06,
      "loss": 0.1829,
      "step": 1580
    },
    {
      "epoch": 18.928571428571427,
      "grad_norm": 6.913029193878174,
      "learning_rate": 8.236331569664903e-06,
      "loss": 0.1914,
      "step": 1590
    },
    {
      "epoch": 19.0,
      "eval_accuracy": 0.9225,
      "eval_loss": 0.23814032971858978,
      "eval_runtime": 6.1225,
      "eval_samples_per_second": 130.665,
      "eval_steps_per_second": 16.333,
      "step": 1596
    },
    {
      "epoch": 19.047619047619047,
      "grad_norm": 8.117836952209473,
      "learning_rate": 8.148148148148148e-06,
      "loss": 0.1953,
      "step": 1600
    },
    {
      "epoch": 19.166666666666668,
      "grad_norm": 7.4035139083862305,
      "learning_rate": 8.059964726631394e-06,
      "loss": 0.1755,
      "step": 1610
    },
    {
      "epoch": 19.285714285714285,
      "grad_norm": 9.961386680603027,
      "learning_rate": 7.971781305114639e-06,
      "loss": 0.1826,
      "step": 1620
    },
    {
      "epoch": 19.404761904761905,
      "grad_norm": 3.902977705001831,
      "learning_rate": 7.883597883597884e-06,
      "loss": 0.2038,
      "step": 1630
    },
    {
      "epoch": 19.523809523809526,
      "grad_norm": 11.7597074508667,
      "learning_rate": 7.79541446208113e-06,
      "loss": 0.1755,
      "step": 1640
    },
    {
      "epoch": 19.642857142857142,
      "grad_norm": 5.295619964599609,
      "learning_rate": 7.707231040564375e-06,
      "loss": 0.1821,
      "step": 1650
    },
    {
      "epoch": 19.761904761904763,
      "grad_norm": 8.691162109375,
      "learning_rate": 7.61904761904762e-06,
      "loss": 0.1931,
      "step": 1660
    },
    {
      "epoch": 19.88095238095238,
      "grad_norm": 14.22559928894043,
      "learning_rate": 7.530864197530865e-06,
      "loss": 0.1842,
      "step": 1670
    },
    {
      "epoch": 20.0,
      "grad_norm": 19.08734130859375,
      "learning_rate": 7.4426807760141095e-06,
      "loss": 0.1741,
      "step": 1680
    },
    {
      "epoch": 20.0,
      "eval_accuracy": 0.92125,
      "eval_loss": 0.2286679595708847,
      "eval_runtime": 6.3863,
      "eval_samples_per_second": 125.269,
      "eval_steps_per_second": 15.659,
      "step": 1680
    },
    {
      "epoch": 20.11904761904762,
      "grad_norm": 11.843003273010254,
      "learning_rate": 7.354497354497355e-06,
      "loss": 0.1704,
      "step": 1690
    },
    {
      "epoch": 20.238095238095237,
      "grad_norm": 4.4936299324035645,
      "learning_rate": 7.2663139329806e-06,
      "loss": 0.1833,
      "step": 1700
    },
    {
      "epoch": 20.357142857142858,
      "grad_norm": 3.9449779987335205,
      "learning_rate": 7.178130511463846e-06,
      "loss": 0.1779,
      "step": 1710
    },
    {
      "epoch": 20.476190476190474,
      "grad_norm": 9.936880111694336,
      "learning_rate": 7.08994708994709e-06,
      "loss": 0.1789,
      "step": 1720
    },
    {
      "epoch": 20.595238095238095,
      "grad_norm": 5.483986854553223,
      "learning_rate": 7.0017636684303355e-06,
      "loss": 0.1776,
      "step": 1730
    },
    {
      "epoch": 20.714285714285715,
      "grad_norm": 8.04780101776123,
      "learning_rate": 6.913580246913581e-06,
      "loss": 0.1737,
      "step": 1740
    },
    {
      "epoch": 20.833333333333332,
      "grad_norm": 6.709885120391846,
      "learning_rate": 6.825396825396826e-06,
      "loss": 0.1855,
      "step": 1750
    },
    {
      "epoch": 20.952380952380953,
      "grad_norm": 7.882541656494141,
      "learning_rate": 6.737213403880071e-06,
      "loss": 0.1682,
      "step": 1760
    },
    {
      "epoch": 21.0,
      "eval_accuracy": 0.92625,
      "eval_loss": 0.21940070390701294,
      "eval_runtime": 5.4374,
      "eval_samples_per_second": 147.128,
      "eval_steps_per_second": 18.391,
      "step": 1764
    },
    {
      "epoch": 21.071428571428573,
      "grad_norm": 5.953088760375977,
      "learning_rate": 6.649029982363316e-06,
      "loss": 0.1741,
      "step": 1770
    },
    {
      "epoch": 21.19047619047619,
      "grad_norm": 7.106684684753418,
      "learning_rate": 6.560846560846561e-06,
      "loss": 0.1745,
      "step": 1780
    },
    {
      "epoch": 21.30952380952381,
      "grad_norm": 13.026501655578613,
      "learning_rate": 6.472663139329807e-06,
      "loss": 0.1844,
      "step": 1790
    },
    {
      "epoch": 21.428571428571427,
      "grad_norm": 3.5783817768096924,
      "learning_rate": 6.384479717813051e-06,
      "loss": 0.1636,
      "step": 1800
    },
    {
      "epoch": 21.547619047619047,
      "grad_norm": 10.339872360229492,
      "learning_rate": 6.296296296296297e-06,
      "loss": 0.1766,
      "step": 1810
    },
    {
      "epoch": 21.666666666666668,
      "grad_norm": 13.413084030151367,
      "learning_rate": 6.208112874779542e-06,
      "loss": 0.1643,
      "step": 1820
    },
    {
      "epoch": 21.785714285714285,
      "grad_norm": 6.75968074798584,
      "learning_rate": 6.119929453262787e-06,
      "loss": 0.1589,
      "step": 1830
    },
    {
      "epoch": 21.904761904761905,
      "grad_norm": 3.4033656120300293,
      "learning_rate": 6.031746031746032e-06,
      "loss": 0.1569,
      "step": 1840
    },
    {
      "epoch": 22.0,
      "eval_accuracy": 0.93375,
      "eval_loss": 0.21772493422031403,
      "eval_runtime": 5.4159,
      "eval_samples_per_second": 147.714,
      "eval_steps_per_second": 18.464,
      "step": 1848
    },
    {
      "epoch": 22.023809523809526,
      "grad_norm": 28.122211456298828,
      "learning_rate": 5.943562610229277e-06,
      "loss": 0.1648,
      "step": 1850
    },
    {
      "epoch": 22.142857142857142,
      "grad_norm": 9.448488235473633,
      "learning_rate": 5.855379188712523e-06,
      "loss": 0.1397,
      "step": 1860
    },
    {
      "epoch": 22.261904761904763,
      "grad_norm": 7.661059379577637,
      "learning_rate": 5.767195767195768e-06,
      "loss": 0.1756,
      "step": 1870
    },
    {
      "epoch": 22.38095238095238,
      "grad_norm": 10.659194946289062,
      "learning_rate": 5.6790123456790125e-06,
      "loss": 0.1588,
      "step": 1880
    },
    {
      "epoch": 22.5,
      "grad_norm": 3.608710527420044,
      "learning_rate": 5.590828924162258e-06,
      "loss": 0.1656,
      "step": 1890
    },
    {
      "epoch": 22.61904761904762,
      "grad_norm": 10.583845138549805,
      "learning_rate": 5.502645502645503e-06,
      "loss": 0.1618,
      "step": 1900
    },
    {
      "epoch": 22.738095238095237,
      "grad_norm": 8.316349029541016,
      "learning_rate": 5.4144620811287486e-06,
      "loss": 0.1541,
      "step": 1910
    },
    {
      "epoch": 22.857142857142858,
      "grad_norm": 9.836575508117676,
      "learning_rate": 5.326278659611993e-06,
      "loss": 0.1588,
      "step": 1920
    },
    {
      "epoch": 22.976190476190474,
      "grad_norm": 7.263607501983643,
      "learning_rate": 5.2380952380952384e-06,
      "loss": 0.144,
      "step": 1930
    },
    {
      "epoch": 23.0,
      "eval_accuracy": 0.93375,
      "eval_loss": 0.21350684762001038,
      "eval_runtime": 5.4573,
      "eval_samples_per_second": 146.592,
      "eval_steps_per_second": 18.324,
      "step": 1932
    },
    {
      "epoch": 23.095238095238095,
      "grad_norm": 9.257414817810059,
      "learning_rate": 5.149911816578484e-06,
      "loss": 0.1624,
      "step": 1940
    },
    {
      "epoch": 23.214285714285715,
      "grad_norm": 23.810653686523438,
      "learning_rate": 5.061728395061729e-06,
      "loss": 0.1528,
      "step": 1950
    },
    {
      "epoch": 23.333333333333332,
      "grad_norm": 8.397602081298828,
      "learning_rate": 4.973544973544974e-06,
      "loss": 0.1506,
      "step": 1960
    },
    {
      "epoch": 23.452380952380953,
      "grad_norm": 3.9385476112365723,
      "learning_rate": 4.885361552028219e-06,
      "loss": 0.1427,
      "step": 1970
    },
    {
      "epoch": 23.571428571428573,
      "grad_norm": 9.915690422058105,
      "learning_rate": 4.7971781305114636e-06,
      "loss": 0.1548,
      "step": 1980
    },
    {
      "epoch": 23.69047619047619,
      "grad_norm": 10.970438003540039,
      "learning_rate": 4.708994708994709e-06,
      "loss": 0.165,
      "step": 1990
    },
    {
      "epoch": 23.80952380952381,
      "grad_norm": 12.917925834655762,
      "learning_rate": 4.620811287477954e-06,
      "loss": 0.1658,
      "step": 2000
    },
    {
      "epoch": 23.928571428571427,
      "grad_norm": 16.563697814941406,
      "learning_rate": 4.5326278659612e-06,
      "loss": 0.1581,
      "step": 2010
    },
    {
      "epoch": 24.0,
      "eval_accuracy": 0.9325,
      "eval_loss": 0.20941905677318573,
      "eval_runtime": 6.2026,
      "eval_samples_per_second": 128.978,
      "eval_steps_per_second": 16.122,
      "step": 2016
    },
    {
      "epoch": 24.047619047619047,
      "grad_norm": 5.663702487945557,
      "learning_rate": 4.444444444444444e-06,
      "loss": 0.1742,
      "step": 2020
    },
    {
      "epoch": 24.166666666666668,
      "grad_norm": 21.897550582885742,
      "learning_rate": 4.3562610229276895e-06,
      "loss": 0.1663,
      "step": 2030
    },
    {
      "epoch": 24.285714285714285,
      "grad_norm": 7.413302421569824,
      "learning_rate": 4.268077601410935e-06,
      "loss": 0.1478,
      "step": 2040
    },
    {
      "epoch": 24.404761904761905,
      "grad_norm": 9.5319242477417,
      "learning_rate": 4.17989417989418e-06,
      "loss": 0.1616,
      "step": 2050
    },
    {
      "epoch": 24.523809523809526,
      "grad_norm": 7.299759864807129,
      "learning_rate": 4.091710758377425e-06,
      "loss": 0.1601,
      "step": 2060
    },
    {
      "epoch": 24.642857142857142,
      "grad_norm": 7.582053184509277,
      "learning_rate": 4.00352733686067e-06,
      "loss": 0.1481,
      "step": 2070
    },
    {
      "epoch": 24.761904761904763,
      "grad_norm": 6.607138633728027,
      "learning_rate": 3.9153439153439155e-06,
      "loss": 0.1447,
      "step": 2080
    },
    {
      "epoch": 24.88095238095238,
      "grad_norm": 5.343249320983887,
      "learning_rate": 3.827160493827161e-06,
      "loss": 0.1366,
      "step": 2090
    },
    {
      "epoch": 25.0,
      "grad_norm": 11.388031005859375,
      "learning_rate": 3.7389770723104058e-06,
      "loss": 0.1426,
      "step": 2100
    },
    {
      "epoch": 25.0,
      "eval_accuracy": 0.935,
      "eval_loss": 0.2057761698961258,
      "eval_runtime": 6.4445,
      "eval_samples_per_second": 124.136,
      "eval_steps_per_second": 15.517,
      "step": 2100
    },
    {
      "epoch": 25.11904761904762,
      "grad_norm": 13.238511085510254,
      "learning_rate": 3.6507936507936507e-06,
      "loss": 0.1451,
      "step": 2110
    },
    {
      "epoch": 25.238095238095237,
      "grad_norm": 18.473859786987305,
      "learning_rate": 3.562610229276896e-06,
      "loss": 0.146,
      "step": 2120
    },
    {
      "epoch": 25.357142857142858,
      "grad_norm": 8.533565521240234,
      "learning_rate": 3.474426807760141e-06,
      "loss": 0.15,
      "step": 2130
    },
    {
      "epoch": 25.476190476190474,
      "grad_norm": 8.290474891662598,
      "learning_rate": 3.3862433862433864e-06,
      "loss": 0.1367,
      "step": 2140
    },
    {
      "epoch": 25.595238095238095,
      "grad_norm": 3.6419670581817627,
      "learning_rate": 3.2980599647266313e-06,
      "loss": 0.1392,
      "step": 2150
    },
    {
      "epoch": 25.714285714285715,
      "grad_norm": 16.031980514526367,
      "learning_rate": 3.2098765432098767e-06,
      "loss": 0.1496,
      "step": 2160
    },
    {
      "epoch": 25.833333333333332,
      "grad_norm": 16.360395431518555,
      "learning_rate": 3.1216931216931216e-06,
      "loss": 0.1705,
      "step": 2170
    },
    {
      "epoch": 25.952380952380953,
      "grad_norm": 7.6555280685424805,
      "learning_rate": 3.033509700176367e-06,
      "loss": 0.1409,
      "step": 2180
    },
    {
      "epoch": 26.0,
      "eval_accuracy": 0.93375,
      "eval_loss": 0.2027110904455185,
      "eval_runtime": 6.4093,
      "eval_samples_per_second": 124.818,
      "eval_steps_per_second": 15.602,
      "step": 2184
    },
    {
      "epoch": 26.071428571428573,
      "grad_norm": 13.298584938049316,
      "learning_rate": 2.945326278659612e-06,
      "loss": 0.1339,
      "step": 2190
    },
    {
      "epoch": 26.19047619047619,
      "grad_norm": 5.4013800621032715,
      "learning_rate": 2.8571428571428573e-06,
      "loss": 0.1388,
      "step": 2200
    },
    {
      "epoch": 26.30952380952381,
      "grad_norm": 6.846310138702393,
      "learning_rate": 2.768959435626102e-06,
      "loss": 0.1379,
      "step": 2210
    },
    {
      "epoch": 26.428571428571427,
      "grad_norm": 8.436290740966797,
      "learning_rate": 2.6807760141093476e-06,
      "loss": 0.1404,
      "step": 2220
    },
    {
      "epoch": 26.547619047619047,
      "grad_norm": 8.173290252685547,
      "learning_rate": 2.5925925925925925e-06,
      "loss": 0.1526,
      "step": 2230
    },
    {
      "epoch": 26.666666666666668,
      "grad_norm": 7.008592128753662,
      "learning_rate": 2.504409171075838e-06,
      "loss": 0.1656,
      "step": 2240
    },
    {
      "epoch": 26.785714285714285,
      "grad_norm": 14.528263092041016,
      "learning_rate": 2.416225749559083e-06,
      "loss": 0.1403,
      "step": 2250
    },
    {
      "epoch": 26.904761904761905,
      "grad_norm": 19.173906326293945,
      "learning_rate": 2.328042328042328e-06,
      "loss": 0.1445,
      "step": 2260
    },
    {
      "epoch": 27.0,
      "eval_accuracy": 0.9275,
      "eval_loss": 0.20720510184764862,
      "eval_runtime": 5.4406,
      "eval_samples_per_second": 147.044,
      "eval_steps_per_second": 18.38,
      "step": 2268
    },
    {
      "epoch": 27.023809523809526,
      "grad_norm": 7.463446140289307,
      "learning_rate": 2.239858906525573e-06,
      "loss": 0.1377,
      "step": 2270
    },
    {
      "epoch": 27.142857142857142,
      "grad_norm": 11.728429794311523,
      "learning_rate": 2.1516754850088184e-06,
      "loss": 0.1315,
      "step": 2280
    },
    {
      "epoch": 27.261904761904763,
      "grad_norm": 4.965158939361572,
      "learning_rate": 2.0634920634920634e-06,
      "loss": 0.1466,
      "step": 2290
    },
    {
      "epoch": 27.38095238095238,
      "grad_norm": 5.408288955688477,
      "learning_rate": 1.9753086419753087e-06,
      "loss": 0.1418,
      "step": 2300
    },
    {
      "epoch": 27.5,
      "grad_norm": 5.822543621063232,
      "learning_rate": 1.887125220458554e-06,
      "loss": 0.1189,
      "step": 2310
    },
    {
      "epoch": 27.61904761904762,
      "grad_norm": 16.357952117919922,
      "learning_rate": 1.798941798941799e-06,
      "loss": 0.1595,
      "step": 2320
    },
    {
      "epoch": 27.738095238095237,
      "grad_norm": 7.142815589904785,
      "learning_rate": 1.7107583774250442e-06,
      "loss": 0.1362,
      "step": 2330
    },
    {
      "epoch": 27.857142857142858,
      "grad_norm": 3.4048264026641846,
      "learning_rate": 1.6225749559082893e-06,
      "loss": 0.1226,
      "step": 2340
    },
    {
      "epoch": 27.976190476190474,
      "grad_norm": 3.4568896293640137,
      "learning_rate": 1.5343915343915345e-06,
      "loss": 0.1472,
      "step": 2350
    },
    {
      "epoch": 28.0,
      "eval_accuracy": 0.93375,
      "eval_loss": 0.2012936919927597,
      "eval_runtime": 5.3691,
      "eval_samples_per_second": 149.0,
      "eval_steps_per_second": 18.625,
      "step": 2352
    },
    {
      "epoch": 28.095238095238095,
      "grad_norm": 25.373470306396484,
      "learning_rate": 1.4462081128747796e-06,
      "loss": 0.1368,
      "step": 2360
    },
    {
      "epoch": 28.214285714285715,
      "grad_norm": 14.415323257446289,
      "learning_rate": 1.3580246913580248e-06,
      "loss": 0.145,
      "step": 2370
    },
    {
      "epoch": 28.333333333333332,
      "grad_norm": 9.555523872375488,
      "learning_rate": 1.26984126984127e-06,
      "loss": 0.1556,
      "step": 2380
    },
    {
      "epoch": 28.452380952380953,
      "grad_norm": 13.204785346984863,
      "learning_rate": 1.181657848324515e-06,
      "loss": 0.1495,
      "step": 2390
    },
    {
      "epoch": 28.571428571428573,
      "grad_norm": 13.89194393157959,
      "learning_rate": 1.0934744268077602e-06,
      "loss": 0.1221,
      "step": 2400
    },
    {
      "epoch": 28.69047619047619,
      "grad_norm": 21.069202423095703,
      "learning_rate": 1.0052910052910054e-06,
      "loss": 0.1376,
      "step": 2410
    },
    {
      "epoch": 28.80952380952381,
      "grad_norm": 13.420437812805176,
      "learning_rate": 9.171075837742504e-07,
      "loss": 0.135,
      "step": 2420
    },
    {
      "epoch": 28.928571428571427,
      "grad_norm": 6.938393592834473,
      "learning_rate": 8.289241622574956e-07,
      "loss": 0.1329,
      "step": 2430
    },
    {
      "epoch": 29.0,
      "eval_accuracy": 0.9375,
      "eval_loss": 0.19804717600345612,
      "eval_runtime": 5.2492,
      "eval_samples_per_second": 152.406,
      "eval_steps_per_second": 19.051,
      "step": 2436
    },
    {
      "epoch": 29.047619047619047,
      "grad_norm": 6.732052803039551,
      "learning_rate": 7.407407407407407e-07,
      "loss": 0.1373,
      "step": 2440
    },
    {
      "epoch": 29.166666666666668,
      "grad_norm": 11.9608154296875,
      "learning_rate": 6.525573192239859e-07,
      "loss": 0.1288,
      "step": 2450
    },
    {
      "epoch": 29.285714285714285,
      "grad_norm": 9.937846183776855,
      "learning_rate": 5.64373897707231e-07,
      "loss": 0.1267,
      "step": 2460
    },
    {
      "epoch": 29.404761904761905,
      "grad_norm": 6.138334274291992,
      "learning_rate": 4.7619047619047623e-07,
      "loss": 0.1358,
      "step": 2470
    },
    {
      "epoch": 29.523809523809526,
      "grad_norm": 5.230373859405518,
      "learning_rate": 3.880070546737214e-07,
      "loss": 0.1473,
      "step": 2480
    },
    {
      "epoch": 29.642857142857142,
      "grad_norm": 8.851503372192383,
      "learning_rate": 2.9982363315696647e-07,
      "loss": 0.1507,
      "step": 2490
    },
    {
      "epoch": 29.761904761904763,
      "grad_norm": 5.106894493103027,
      "learning_rate": 2.1164021164021165e-07,
      "loss": 0.1262,
      "step": 2500
    },
    {
      "epoch": 29.88095238095238,
      "grad_norm": 9.42648696899414,
      "learning_rate": 1.234567901234568e-07,
      "loss": 0.1397,
      "step": 2510
    },
    {
      "epoch": 30.0,
      "grad_norm": 8.953765869140625,
      "learning_rate": 3.527336860670194e-08,
      "loss": 0.132,
      "step": 2520
    },
    {
      "epoch": 30.0,
      "eval_accuracy": 0.93625,
      "eval_loss": 0.1968877911567688,
      "eval_runtime": 5.803,
      "eval_samples_per_second": 137.859,
      "eval_steps_per_second": 17.232,
      "step": 2520
    }
  ],
  "logging_steps": 10,
  "max_steps": 2520,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 30,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 1.019477021294592e+19,
  "train_batch_size": 192,
  "trial_name": null,
  "trial_params": null
}