{
  "best_metric": 0.4384986460208893,
  "best_model_checkpoint": "rm_ckpt_refine_reward-p100-w0.3-s0/checkpoint-5600",
  "epoch": 3.0239065337980464,
  "eval_steps": 200,
  "global_step": 5600,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.01,
      "learning_rate": 4.9999999999999996e-06,
      "loss": 0.6508,
      "step": 10
    },
    {
      "epoch": 0.01,
      "learning_rate": 6.505149978319905e-06,
      "loss": 0.5542,
      "step": 20
    },
    {
      "epoch": 0.02,
      "learning_rate": 7.385606273598311e-06,
      "loss": 0.5076,
      "step": 30
    },
    {
      "epoch": 0.02,
      "learning_rate": 8.010299956639811e-06,
      "loss": 0.4815,
      "step": 40
    },
    {
      "epoch": 0.03,
      "learning_rate": 8.494850021680093e-06,
      "loss": 0.4759,
      "step": 50
    },
    {
      "epoch": 0.03,
      "learning_rate": 8.890756251918216e-06,
      "loss": 0.4733,
      "step": 60
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.225490200071284e-06,
      "loss": 0.4675,
      "step": 70
    },
    {
      "epoch": 0.04,
      "learning_rate": 9.515449934959717e-06,
      "loss": 0.4774,
      "step": 80
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.771212547196623e-06,
      "loss": 0.4607,
      "step": 90
    },
    {
      "epoch": 0.05,
      "learning_rate": 9.999999999999999e-06,
      "loss": 0.4592,
      "step": 100
    },
    {
      "epoch": 0.06,
      "learning_rate": 1e-05,
      "loss": 0.4606,
      "step": 110
    },
    {
      "epoch": 0.06,
      "learning_rate": 1e-05,
      "loss": 0.4583,
      "step": 120
    },
    {
      "epoch": 0.07,
      "learning_rate": 1e-05,
      "loss": 0.4752,
      "step": 130
    },
    {
      "epoch": 0.08,
      "learning_rate": 1e-05,
      "loss": 0.4545,
      "step": 140
    },
    {
      "epoch": 0.08,
      "learning_rate": 1e-05,
      "loss": 0.4627,
      "step": 150
    },
    {
      "epoch": 0.09,
      "learning_rate": 1e-05,
      "loss": 0.4486,
      "step": 160
    },
    {
      "epoch": 0.09,
      "learning_rate": 1e-05,
      "loss": 0.4438,
      "step": 170
    },
    {
      "epoch": 0.1,
      "learning_rate": 1e-05,
      "loss": 0.4466,
      "step": 180
    },
    {
      "epoch": 0.1,
      "learning_rate": 1e-05,
      "loss": 0.4539,
      "step": 190
    },
    {
      "epoch": 0.11,
      "learning_rate": 1e-05,
      "loss": 0.4468,
      "step": 200
    },
    {
      "epoch": 0.11,
      "eval_val_accuracy": 0.6277666666666667,
      "eval_val_loss": 0.4563558101654053,
      "eval_val_runtime": 2251.273,
      "eval_val_samples_per_second": 4.442,
      "eval_val_steps_per_second": 0.555,
      "step": 200
    },
    {
      "epoch": 0.11,
      "eval_test_accuracy": 0.63045,
      "eval_test_loss": 0.45287254452705383,
      "eval_test_runtime": 2252.3574,
      "eval_test_samples_per_second": 4.44,
      "eval_test_steps_per_second": 0.555,
      "step": 200
    },
    {
      "epoch": 0.11,
      "learning_rate": 1e-05,
      "loss": 0.444,
      "step": 210
    },
    {
      "epoch": 0.12,
      "learning_rate": 1e-05,
      "loss": 0.4494,
      "step": 220
    },
    {
      "epoch": 0.12,
      "learning_rate": 1e-05,
      "loss": 0.4629,
      "step": 230
    },
    {
      "epoch": 0.13,
      "learning_rate": 1e-05,
      "loss": 0.4515,
      "step": 240
    },
    {
      "epoch": 0.13,
      "learning_rate": 1e-05,
      "loss": 0.4487,
      "step": 250
    },
    {
      "epoch": 0.14,
      "learning_rate": 1e-05,
      "loss": 0.4577,
      "step": 260
    },
    {
      "epoch": 0.15,
      "learning_rate": 1e-05,
      "loss": 0.4549,
      "step": 270
    },
    {
      "epoch": 0.15,
      "learning_rate": 1e-05,
      "loss": 0.4392,
      "step": 280
    },
    {
      "epoch": 0.16,
      "learning_rate": 1e-05,
      "loss": 0.4553,
      "step": 290
    },
    {
      "epoch": 0.16,
      "learning_rate": 1e-05,
      "loss": 0.4616,
      "step": 300
    },
    {
      "epoch": 0.17,
      "learning_rate": 1e-05,
      "loss": 0.4474,
      "step": 310
    },
    {
      "epoch": 0.17,
      "learning_rate": 1e-05,
      "loss": 0.4597,
      "step": 320
    },
    {
      "epoch": 0.18,
      "learning_rate": 1e-05,
      "loss": 0.4446,
      "step": 330
    },
    {
      "epoch": 0.18,
      "learning_rate": 1e-05,
      "loss": 0.4495,
      "step": 340
    },
    {
      "epoch": 0.19,
      "learning_rate": 1e-05,
      "loss": 0.4504,
      "step": 350
    },
    {
      "epoch": 0.19,
      "learning_rate": 1e-05,
      "loss": 0.4559,
      "step": 360
    },
    {
      "epoch": 0.2,
      "learning_rate": 1e-05,
      "loss": 0.4461,
      "step": 370
    },
    {
      "epoch": 0.21,
      "learning_rate": 1e-05,
      "loss": 0.4527,
      "step": 380
    },
    {
      "epoch": 0.21,
      "learning_rate": 1e-05,
      "loss": 0.4519,
      "step": 390
    },
    {
      "epoch": 0.22,
      "learning_rate": 1e-05,
      "loss": 0.4406,
      "step": 400
    },
    {
      "epoch": 0.22,
      "eval_val_accuracy": 0.6309166666666667,
      "eval_val_loss": 0.44894179701805115,
      "eval_val_runtime": 2253.5504,
      "eval_val_samples_per_second": 4.437,
      "eval_val_steps_per_second": 0.555,
      "step": 400
    },
    {
      "epoch": 0.22,
      "eval_test_accuracy": 0.6305666666666667,
      "eval_test_loss": 0.44730547070503235,
      "eval_test_runtime": 2254.5876,
      "eval_test_samples_per_second": 4.435,
      "eval_test_steps_per_second": 0.554,
      "step": 400
    },
    {
      "epoch": 0.22,
      "learning_rate": 1e-05,
      "loss": 0.4422,
      "step": 410
    },
    {
      "epoch": 0.23,
      "learning_rate": 1e-05,
      "loss": 0.4533,
      "step": 420
    },
    {
      "epoch": 0.23,
      "learning_rate": 1e-05,
      "loss": 0.4474,
      "step": 430
    },
    {
      "epoch": 0.24,
      "learning_rate": 1e-05,
      "loss": 0.4468,
      "step": 440
    },
    {
      "epoch": 0.24,
      "learning_rate": 1e-05,
      "loss": 0.4583,
      "step": 450
    },
    {
      "epoch": 0.25,
      "learning_rate": 1e-05,
      "loss": 0.4528,
      "step": 460
    },
    {
      "epoch": 0.25,
      "learning_rate": 1e-05,
      "loss": 0.4508,
      "step": 470
    },
    {
      "epoch": 0.26,
      "learning_rate": 1e-05,
      "loss": 0.4403,
      "step": 480
    },
    {
      "epoch": 0.26,
      "learning_rate": 1e-05,
      "loss": 0.4411,
      "step": 490
    },
    {
      "epoch": 0.27,
      "learning_rate": 1e-05,
      "loss": 0.4353,
      "step": 500
    },
    {
      "epoch": 0.28,
      "learning_rate": 1e-05,
      "loss": 0.461,
      "step": 510
    },
    {
      "epoch": 0.28,
      "learning_rate": 1e-05,
      "loss": 0.4444,
      "step": 520
    },
    {
      "epoch": 0.29,
      "learning_rate": 1e-05,
      "loss": 0.4468,
      "step": 530
    },
    {
      "epoch": 0.29,
      "learning_rate": 1e-05,
      "loss": 0.4522,
      "step": 540
    },
    {
      "epoch": 0.3,
      "learning_rate": 1e-05,
      "loss": 0.4504,
      "step": 550
    },
    {
      "epoch": 0.3,
      "learning_rate": 1e-05,
      "loss": 0.4512,
      "step": 560
    },
    {
      "epoch": 0.31,
      "learning_rate": 1e-05,
      "loss": 0.4414,
      "step": 570
    },
    {
      "epoch": 0.31,
      "learning_rate": 1e-05,
      "loss": 0.4563,
      "step": 580
    },
    {
      "epoch": 0.32,
      "learning_rate": 1e-05,
      "loss": 0.4537,
      "step": 590
    },
    {
      "epoch": 0.32,
      "learning_rate": 1e-05,
      "loss": 0.4362,
      "step": 600
    },
    {
      "epoch": 0.32,
      "eval_val_accuracy": 0.63345,
      "eval_val_loss": 0.4456494152545929,
      "eval_val_runtime": 2252.9912,
      "eval_val_samples_per_second": 4.439,
      "eval_val_steps_per_second": 0.555,
      "step": 600
    },
    {
      "epoch": 0.32,
      "eval_test_accuracy": 0.635,
      "eval_test_loss": 0.4436740279197693,
      "eval_test_runtime": 2254.1983,
      "eval_test_samples_per_second": 4.436,
      "eval_test_steps_per_second": 0.555,
      "step": 600
    },
    {
      "epoch": 0.33,
      "learning_rate": 1e-05,
      "loss": 0.4569,
      "step": 610
    },
    {
      "epoch": 0.33,
      "learning_rate": 1e-05,
      "loss": 0.4399,
      "step": 620
    },
    {
      "epoch": 0.34,
      "learning_rate": 1e-05,
      "loss": 0.4458,
      "step": 630
    },
    {
      "epoch": 0.35,
      "learning_rate": 1e-05,
      "loss": 0.4463,
      "step": 640
    },
    {
      "epoch": 0.35,
      "learning_rate": 1e-05,
      "loss": 0.4443,
      "step": 650
    },
    {
      "epoch": 0.36,
      "learning_rate": 1e-05,
      "loss": 0.4456,
      "step": 660
    },
    {
      "epoch": 0.36,
      "learning_rate": 1e-05,
      "loss": 0.4428,
      "step": 670
    },
    {
      "epoch": 0.37,
      "learning_rate": 1e-05,
      "loss": 0.4427,
      "step": 680
    },
    {
      "epoch": 0.37,
      "learning_rate": 1e-05,
      "loss": 0.446,
      "step": 690
    },
    {
      "epoch": 0.38,
      "learning_rate": 1e-05,
      "loss": 0.4337,
      "step": 700
    },
    {
      "epoch": 0.38,
      "learning_rate": 1e-05,
      "loss": 0.4447,
      "step": 710
    },
    {
      "epoch": 0.39,
      "learning_rate": 1e-05,
      "loss": 0.449,
      "step": 720
    },
    {
      "epoch": 0.39,
      "learning_rate": 1e-05,
      "loss": 0.4445,
      "step": 730
    },
    {
      "epoch": 0.4,
      "learning_rate": 1e-05,
      "loss": 0.4401,
      "step": 740
    },
    {
      "epoch": 0.4,
      "learning_rate": 1e-05,
      "loss": 0.4252,
      "step": 750
    },
    {
      "epoch": 0.41,
      "learning_rate": 1e-05,
      "loss": 0.4466,
      "step": 760
    },
    {
      "epoch": 0.42,
      "learning_rate": 1e-05,
      "loss": 0.4405,
      "step": 770
    },
    {
      "epoch": 0.42,
      "learning_rate": 1e-05,
      "loss": 0.4384,
      "step": 780
    },
    {
      "epoch": 0.43,
      "learning_rate": 1e-05,
      "loss": 0.4574,
      "step": 790
    },
    {
      "epoch": 0.43,
      "learning_rate": 1e-05,
      "loss": 0.43,
      "step": 800
    },
    {
      "epoch": 0.43,
      "eval_val_accuracy": 0.6336833333333333,
      "eval_val_loss": 0.44466710090637207,
      "eval_val_runtime": 2251.9412,
      "eval_val_samples_per_second": 4.441,
      "eval_val_steps_per_second": 0.555,
      "step": 800
    },
    {
      "epoch": 0.43,
      "eval_test_accuracy": 0.6322333333333333,
      "eval_test_loss": 0.4431970715522766,
      "eval_test_runtime": 2253.1961,
      "eval_test_samples_per_second": 4.438,
      "eval_test_steps_per_second": 0.555,
      "step": 800
    },
    {
      "epoch": 0.44,
      "learning_rate": 1e-05,
      "loss": 0.4342,
      "step": 810
    },
    {
      "epoch": 0.44,
      "learning_rate": 1e-05,
      "loss": 0.4482,
      "step": 820
    },
    {
      "epoch": 0.45,
      "learning_rate": 1e-05,
      "loss": 0.4284,
      "step": 830
    },
    {
      "epoch": 0.45,
      "learning_rate": 1e-05,
      "loss": 0.4504,
      "step": 840
    },
    {
      "epoch": 0.46,
      "learning_rate": 1e-05,
      "loss": 0.4512,
      "step": 850
    },
    {
      "epoch": 0.46,
      "learning_rate": 1e-05,
      "loss": 0.443,
      "step": 860
    },
    {
      "epoch": 0.47,
      "learning_rate": 1e-05,
      "loss": 0.4628,
      "step": 870
    },
    {
      "epoch": 0.48,
      "learning_rate": 1e-05,
      "loss": 0.4486,
      "step": 880
    },
    {
      "epoch": 0.48,
      "learning_rate": 1e-05,
      "loss": 0.4527,
      "step": 890
    },
    {
      "epoch": 0.49,
      "learning_rate": 1e-05,
      "loss": 0.443,
      "step": 900
    },
    {
      "epoch": 0.49,
      "learning_rate": 1e-05,
      "loss": 0.4498,
      "step": 910
    },
    {
      "epoch": 0.5,
      "learning_rate": 1e-05,
      "loss": 0.4401,
      "step": 920
    },
    {
      "epoch": 0.5,
      "learning_rate": 1e-05,
      "loss": 0.4374,
      "step": 930
    },
    {
      "epoch": 0.51,
      "learning_rate": 1e-05,
      "loss": 0.4405,
      "step": 940
    },
    {
      "epoch": 0.51,
      "learning_rate": 1e-05,
      "loss": 0.4402,
      "step": 950
    },
    {
      "epoch": 0.52,
      "learning_rate": 1e-05,
      "loss": 0.4382,
      "step": 960
    },
    {
      "epoch": 0.52,
      "learning_rate": 1e-05,
      "loss": 0.4495,
      "step": 970
    },
    {
      "epoch": 0.53,
      "learning_rate": 1e-05,
      "loss": 0.4468,
      "step": 980
    },
    {
      "epoch": 0.53,
      "learning_rate": 1e-05,
      "loss": 0.4355,
      "step": 990
    },
    {
      "epoch": 0.54,
      "learning_rate": 1e-05,
      "loss": 0.4512,
      "step": 1000
    },
    {
      "epoch": 0.54,
      "eval_val_accuracy": 0.63365,
      "eval_val_loss": 0.44436272978782654,
      "eval_val_runtime": 2248.8627,
      "eval_val_samples_per_second": 4.447,
      "eval_val_steps_per_second": 0.556,
      "step": 1000
    },
    {
      "epoch": 0.54,
      "eval_test_accuracy": 0.63305,
      "eval_test_loss": 0.4426984488964081,
      "eval_test_runtime": 2250.0282,
      "eval_test_samples_per_second": 4.444,
      "eval_test_steps_per_second": 0.556,
      "step": 1000
    },
    {
      "epoch": 0.55,
      "learning_rate": 1e-05,
      "loss": 0.4372,
      "step": 1010
    },
    {
      "epoch": 0.55,
      "learning_rate": 1e-05,
      "loss": 0.4413,
      "step": 1020
    },
    {
      "epoch": 0.56,
      "learning_rate": 1e-05,
      "loss": 0.451,
      "step": 1030
    },
    {
      "epoch": 0.56,
      "learning_rate": 1e-05,
      "loss": 0.4368,
      "step": 1040
    },
    {
      "epoch": 0.57,
      "learning_rate": 1e-05,
      "loss": 0.4362,
      "step": 1050
    },
    {
      "epoch": 0.57,
      "learning_rate": 1e-05,
      "loss": 0.4417,
      "step": 1060
    },
    {
      "epoch": 0.58,
      "learning_rate": 1e-05,
      "loss": 0.4511,
      "step": 1070
    },
    {
      "epoch": 0.58,
      "learning_rate": 1e-05,
      "loss": 0.4462,
      "step": 1080
    },
    {
      "epoch": 0.59,
      "learning_rate": 1e-05,
      "loss": 0.4551,
      "step": 1090
    },
    {
      "epoch": 0.59,
      "learning_rate": 1e-05,
      "loss": 0.4345,
      "step": 1100
    },
    {
      "epoch": 0.6,
      "learning_rate": 1e-05,
      "loss": 0.452,
      "step": 1110
    },
    {
      "epoch": 0.6,
      "learning_rate": 1e-05,
      "loss": 0.4318,
      "step": 1120
    },
    {
      "epoch": 0.61,
      "learning_rate": 1e-05,
      "loss": 0.4419,
      "step": 1130
    },
    {
      "epoch": 0.62,
      "learning_rate": 1e-05,
      "loss": 0.4423,
      "step": 1140
    },
    {
      "epoch": 0.62,
      "learning_rate": 1e-05,
      "loss": 0.4391,
      "step": 1150
    },
    {
      "epoch": 0.63,
      "learning_rate": 1e-05,
      "loss": 0.441,
      "step": 1160
    },
    {
      "epoch": 0.63,
      "learning_rate": 1e-05,
      "loss": 0.4419,
      "step": 1170
    },
    {
      "epoch": 0.64,
      "learning_rate": 1e-05,
      "loss": 0.4412,
      "step": 1180
    },
    {
      "epoch": 0.64,
      "learning_rate": 1e-05,
      "loss": 0.4345,
      "step": 1190
    },
    {
      "epoch": 0.65,
      "learning_rate": 1e-05,
      "loss": 0.4446,
      "step": 1200
    },
    {
      "epoch": 0.65,
      "eval_val_accuracy": 0.63345,
      "eval_val_loss": 0.44255268573760986,
      "eval_val_runtime": 2248.3014,
      "eval_val_samples_per_second": 4.448,
      "eval_val_steps_per_second": 0.556,
      "step": 1200
    },
    {
      "epoch": 0.65,
      "eval_test_accuracy": 0.6339333333333333,
      "eval_test_loss": 0.44228971004486084,
      "eval_test_runtime": 2250.9332,
      "eval_test_samples_per_second": 4.443,
      "eval_test_steps_per_second": 0.555,
      "step": 1200
    },
    {
      "epoch": 0.65,
      "learning_rate": 1e-05,
      "loss": 0.453,
      "step": 1210
    },
    {
      "epoch": 0.66,
      "learning_rate": 1e-05,
      "loss": 0.4354,
      "step": 1220
    },
    {
      "epoch": 0.66,
      "learning_rate": 1e-05,
      "loss": 0.4456,
      "step": 1230
    },
    {
      "epoch": 0.67,
      "learning_rate": 1e-05,
      "loss": 0.4245,
      "step": 1240
    },
    {
      "epoch": 0.67,
      "learning_rate": 1e-05,
      "loss": 0.4333,
      "step": 1250
    },
    {
      "epoch": 0.68,
      "learning_rate": 1e-05,
      "loss": 0.4512,
      "step": 1260
    },
    {
      "epoch": 0.69,
      "learning_rate": 1e-05,
      "loss": 0.4458,
      "step": 1270
    },
    {
      "epoch": 0.69,
      "learning_rate": 1e-05,
      "loss": 0.4413,
      "step": 1280
    },
    {
      "epoch": 0.7,
      "learning_rate": 1e-05,
      "loss": 0.4423,
      "step": 1290
    },
    {
      "epoch": 0.7,
      "learning_rate": 1e-05,
      "loss": 0.4481,
      "step": 1300
    },
    {
      "epoch": 0.71,
      "learning_rate": 1e-05,
      "loss": 0.4325,
      "step": 1310
    },
    {
      "epoch": 0.71,
      "learning_rate": 1e-05,
      "loss": 0.4299,
      "step": 1320
    },
    {
      "epoch": 0.72,
      "learning_rate": 1e-05,
      "loss": 0.4437,
      "step": 1330
    },
    {
      "epoch": 0.72,
      "learning_rate": 1e-05,
      "loss": 0.4454,
      "step": 1340
    },
    {
      "epoch": 0.73,
      "learning_rate": 1e-05,
      "loss": 0.441,
      "step": 1350
    },
    {
      "epoch": 0.73,
      "learning_rate": 1e-05,
      "loss": 0.4476,
      "step": 1360
    },
    {
      "epoch": 0.74,
      "learning_rate": 1e-05,
      "loss": 0.4453,
      "step": 1370
    },
    {
      "epoch": 0.75,
      "learning_rate": 1e-05,
      "loss": 0.4423,
      "step": 1380
    },
    {
      "epoch": 0.75,
      "learning_rate": 1e-05,
      "loss": 0.441,
      "step": 1390
    },
    {
      "epoch": 0.76,
      "learning_rate": 1e-05,
      "loss": 0.4334,
      "step": 1400
    },
    {
      "epoch": 0.76,
      "eval_val_accuracy": 0.6357666666666667,
      "eval_val_loss": 0.44135814905166626,
      "eval_val_runtime": 2249.0131,
      "eval_val_samples_per_second": 4.446,
      "eval_val_steps_per_second": 0.556,
      "step": 1400
    },
    {
      "epoch": 0.76,
      "eval_test_accuracy": 0.6341,
      "eval_test_loss": 0.44054901599884033,
      "eval_test_runtime": 2250.0171,
      "eval_test_samples_per_second": 4.444,
      "eval_test_steps_per_second": 0.556,
      "step": 1400
    },
    {
      "epoch": 0.76,
      "learning_rate": 1e-05,
      "loss": 0.4378,
      "step": 1410
    },
    {
      "epoch": 0.77,
      "learning_rate": 1e-05,
      "loss": 0.4412,
      "step": 1420
    },
    {
      "epoch": 0.77,
      "learning_rate": 1e-05,
      "loss": 0.4442,
      "step": 1430
    },
    {
      "epoch": 0.78,
      "learning_rate": 1e-05,
      "loss": 0.4378,
      "step": 1440
    },
    {
      "epoch": 0.78,
      "learning_rate": 1e-05,
      "loss": 0.4368,
      "step": 1450
    },
    {
      "epoch": 0.79,
      "learning_rate": 1e-05,
      "loss": 0.4356,
      "step": 1460
    },
    {
      "epoch": 0.79,
      "learning_rate": 1e-05,
      "loss": 0.4337,
      "step": 1470
    },
    {
      "epoch": 0.8,
      "learning_rate": 1e-05,
      "loss": 0.4421,
      "step": 1480
    },
    {
      "epoch": 0.8,
      "learning_rate": 1e-05,
      "loss": 0.4304,
      "step": 1490
    },
    {
      "epoch": 0.81,
      "learning_rate": 1e-05,
      "loss": 0.4348,
      "step": 1500
    },
    {
      "epoch": 0.82,
      "learning_rate": 1e-05,
      "loss": 0.4479,
      "step": 1510
    },
    {
      "epoch": 0.82,
      "learning_rate": 1e-05,
      "loss": 0.4435,
      "step": 1520
    },
    {
      "epoch": 0.83,
      "learning_rate": 1e-05,
      "loss": 0.4375,
      "step": 1530
    },
    {
      "epoch": 0.83,
      "learning_rate": 1e-05,
      "loss": 0.4242,
      "step": 1540
    },
    {
      "epoch": 0.84,
      "learning_rate": 1e-05,
      "loss": 0.4406,
      "step": 1550
    },
    {
      "epoch": 0.84,
      "learning_rate": 1e-05,
      "loss": 0.4508,
      "step": 1560
    },
    {
      "epoch": 0.85,
      "learning_rate": 1e-05,
      "loss": 0.4397,
      "step": 1570
    },
    {
      "epoch": 0.85,
      "learning_rate": 1e-05,
      "loss": 0.4435,
      "step": 1580
    },
    {
      "epoch": 0.86,
      "learning_rate": 1e-05,
      "loss": 0.4516,
      "step": 1590
    },
    {
      "epoch": 0.86,
      "learning_rate": 1e-05,
      "loss": 0.4377,
      "step": 1600
    },
    {
      "epoch": 0.86,
      "eval_val_accuracy": 0.6362833333333333,
      "eval_val_loss": 0.4415554702281952,
      "eval_val_runtime": 2247.1988,
      "eval_val_samples_per_second": 4.45,
      "eval_val_steps_per_second": 0.556,
      "step": 1600
    },
    {
      "epoch": 0.86,
      "eval_test_accuracy": 0.6348833333333334,
      "eval_test_loss": 0.44132542610168457,
      "eval_test_runtime": 2248.5935,
      "eval_test_samples_per_second": 4.447,
      "eval_test_steps_per_second": 0.556,
      "step": 1600
    },
    {
      "epoch": 0.87,
      "learning_rate": 1e-05,
      "loss": 0.4415,
      "step": 1610
    },
    {
      "epoch": 0.87,
      "learning_rate": 1e-05,
      "loss": 0.4393,
      "step": 1620
    },
    {
      "epoch": 0.88,
      "learning_rate": 1e-05,
      "loss": 0.4422,
      "step": 1630
    },
    {
      "epoch": 0.89,
      "learning_rate": 1e-05,
      "loss": 0.4322,
      "step": 1640
    },
    {
      "epoch": 0.89,
      "learning_rate": 1e-05,
      "loss": 0.4327,
      "step": 1650
    },
    {
      "epoch": 0.9,
      "learning_rate": 1e-05,
      "loss": 0.4317,
      "step": 1660
    },
    {
      "epoch": 0.9,
      "learning_rate": 1e-05,
      "loss": 0.443,
      "step": 1670
    },
    {
      "epoch": 0.91,
      "learning_rate": 1e-05,
      "loss": 0.4488,
      "step": 1680
    },
    {
      "epoch": 0.91,
      "learning_rate": 1e-05,
      "loss": 0.4273,
      "step": 1690
    },
    {
      "epoch": 0.92,
      "learning_rate": 1e-05,
      "loss": 0.4536,
      "step": 1700
    },
    {
      "epoch": 0.92,
      "learning_rate": 1e-05,
      "loss": 0.4339,
      "step": 1710
    },
    {
      "epoch": 0.93,
      "learning_rate": 1e-05,
      "loss": 0.4438,
      "step": 1720
    },
    {
      "epoch": 0.93,
      "learning_rate": 1e-05,
      "loss": 0.4379,
      "step": 1730
    },
    {
      "epoch": 0.94,
      "learning_rate": 1e-05,
      "loss": 0.4366,
      "step": 1740
    },
    {
      "epoch": 0.94,
      "learning_rate": 1e-05,
      "loss": 0.4363,
      "step": 1750
    },
    {
      "epoch": 0.95,
      "learning_rate": 1e-05,
      "loss": 0.4349,
      "step": 1760
    },
    {
      "epoch": 0.96,
      "learning_rate": 1e-05,
      "loss": 0.4407,
      "step": 1770
    },
    {
      "epoch": 0.96,
      "learning_rate": 1e-05,
      "loss": 0.4393,
      "step": 1780
    },
    {
      "epoch": 0.97,
      "learning_rate": 1e-05,
      "loss": 0.4441,
      "step": 1790
    },
    {
      "epoch": 0.97,
      "learning_rate": 1e-05,
      "loss": 0.4451,
      "step": 1800
    },
    {
      "epoch": 0.97,
      "eval_val_accuracy": 0.6365,
      "eval_val_loss": 0.44033148884773254,
      "eval_val_runtime": 2244.905,
      "eval_val_samples_per_second": 4.455,
      "eval_val_steps_per_second": 0.557,
      "step": 1800
    },
    {
      "epoch": 0.97,
      "eval_test_accuracy": 0.6357833333333333,
      "eval_test_loss": 0.43971341848373413,
      "eval_test_runtime": 2247.1088,
      "eval_test_samples_per_second": 4.45,
      "eval_test_steps_per_second": 0.556,
      "step": 1800
    },
    {
      "epoch": 0.98,
      "learning_rate": 1e-05,
      "loss": 0.4389,
      "step": 1810
    },
    {
      "epoch": 0.98,
      "learning_rate": 1e-05,
      "loss": 0.4406,
      "step": 1820
    },
    {
      "epoch": 0.99,
      "learning_rate": 1e-05,
      "loss": 0.4404,
      "step": 1830
    },
    {
      "epoch": 0.99,
      "learning_rate": 1e-05,
      "loss": 0.4373,
      "step": 1840
    },
    {
      "epoch": 1.0,
      "learning_rate": 1e-05,
      "loss": 0.4456,
      "step": 1850
    },
    {
      "epoch": 1.0,
      "learning_rate": 1e-05,
      "loss": 0.4326,
      "step": 1860
    },
    {
      "epoch": 1.01,
      "learning_rate": 1e-05,
      "loss": 0.4331,
      "step": 1870
    },
    {
      "epoch": 1.02,
      "learning_rate": 1e-05,
      "loss": 0.4411,
      "step": 1880
    },
    {
      "epoch": 1.02,
      "learning_rate": 1e-05,
      "loss": 0.4307,
      "step": 1890
    },
    {
      "epoch": 1.03,
      "learning_rate": 1e-05,
      "loss": 0.4266,
      "step": 1900
    },
    {
      "epoch": 1.03,
      "learning_rate": 1e-05,
      "loss": 0.4289,
      "step": 1910
    },
    {
      "epoch": 1.04,
      "learning_rate": 1e-05,
      "loss": 0.422,
      "step": 1920
    },
    {
      "epoch": 1.04,
      "learning_rate": 1e-05,
      "loss": 0.4327,
      "step": 1930
    },
    {
      "epoch": 1.05,
      "learning_rate": 1e-05,
      "loss": 0.4197,
      "step": 1940
    },
    {
      "epoch": 1.05,
      "learning_rate": 1e-05,
      "loss": 0.4245,
      "step": 1950
    },
    {
      "epoch": 1.06,
      "learning_rate": 1e-05,
      "loss": 0.4184,
      "step": 1960
    },
    {
      "epoch": 1.06,
      "learning_rate": 1e-05,
      "loss": 0.4218,
      "step": 1970
    },
    {
      "epoch": 1.07,
      "learning_rate": 1e-05,
      "loss": 0.425,
      "step": 1980
    },
    {
      "epoch": 1.07,
      "learning_rate": 1e-05,
      "loss": 0.4213,
      "step": 1990
    },
    {
      "epoch": 1.08,
      "learning_rate": 1e-05,
      "loss": 0.418,
      "step": 2000
    },
    {
      "epoch": 1.08,
      "eval_val_accuracy": 0.6367333333333333,
      "eval_val_loss": 0.4390324354171753,
      "eval_val_runtime": 2247.499,
      "eval_val_samples_per_second": 4.449,
      "eval_val_steps_per_second": 0.556,
      "step": 2000
    },
    {
      "epoch": 1.08,
      "eval_test_accuracy": 0.63745,
      "eval_test_loss": 0.4380366802215576,
      "eval_test_runtime": 2248.656,
      "eval_test_samples_per_second": 4.447,
      "eval_test_steps_per_second": 0.556,
      "step": 2000
    },
    {
      "epoch": 1.09,
      "learning_rate": 1e-05,
      "loss": 0.4133,
      "step": 2010
    },
    {
      "epoch": 1.09,
      "learning_rate": 1e-05,
      "loss": 0.4144,
      "step": 2020
    },
    {
      "epoch": 1.1,
      "learning_rate": 1e-05,
      "loss": 0.419,
      "step": 2030
    },
    {
      "epoch": 1.1,
      "learning_rate": 1e-05,
      "loss": 0.4185,
      "step": 2040
    },
    {
      "epoch": 1.11,
      "learning_rate": 1e-05,
      "loss": 0.4184,
      "step": 2050
    },
    {
      "epoch": 1.11,
      "learning_rate": 1e-05,
      "loss": 0.4104,
      "step": 2060
    },
    {
      "epoch": 1.12,
      "learning_rate": 1e-05,
      "loss": 0.4131,
      "step": 2070
    },
    {
      "epoch": 1.12,
      "learning_rate": 1e-05,
      "loss": 0.4177,
      "step": 2080
    },
    {
      "epoch": 1.13,
      "learning_rate": 1e-05,
      "loss": 0.4135,
      "step": 2090
    },
    {
      "epoch": 1.13,
      "learning_rate": 1e-05,
      "loss": 0.4143,
      "step": 2100
    },
    {
      "epoch": 1.14,
      "learning_rate": 1e-05,
      "loss": 0.4249,
      "step": 2110
    },
    {
      "epoch": 1.14,
      "learning_rate": 1e-05,
      "loss": 0.4156,
      "step": 2120
    },
    {
      "epoch": 1.15,
      "learning_rate": 1e-05,
      "loss": 0.4108,
      "step": 2130
    },
    {
      "epoch": 1.16,
      "learning_rate": 1e-05,
      "loss": 0.4144,
      "step": 2140
    },
    {
      "epoch": 1.16,
      "learning_rate": 1e-05,
      "loss": 0.4192,
      "step": 2150
    },
    {
      "epoch": 1.17,
      "learning_rate": 1e-05,
      "loss": 0.4179,
      "step": 2160
    },
    {
      "epoch": 1.17,
      "learning_rate": 1e-05,
      "loss": 0.4169,
      "step": 2170
    },
    {
      "epoch": 1.18,
      "learning_rate": 1e-05,
      "loss": 0.4147,
      "step": 2180
    },
    {
      "epoch": 1.18,
      "learning_rate": 1e-05,
      "loss": 0.4187,
      "step": 2190
    },
    {
      "epoch": 1.19,
      "learning_rate": 1e-05,
      "loss": 0.4185,
      "step": 2200
    },
    {
      "epoch": 1.19,
      "eval_val_accuracy": 0.6359,
      "eval_val_loss": 0.43944188952445984,
      "eval_val_runtime": 2247.1138,
      "eval_val_samples_per_second": 4.45,
      "eval_val_steps_per_second": 0.556,
      "step": 2200
    },
    {
      "epoch": 1.19,
      "eval_test_accuracy": 0.6373,
      "eval_test_loss": 0.43821561336517334,
      "eval_test_runtime": 2248.3005,
      "eval_test_samples_per_second": 4.448,
      "eval_test_steps_per_second": 0.556,
      "step": 2200
    },
    {
      "epoch": 1.19,
      "learning_rate": 1e-05,
      "loss": 0.4218,
      "step": 2210
    },
    {
      "epoch": 1.2,
      "learning_rate": 1e-05,
      "loss": 0.416,
      "step": 2220
    },
    {
      "epoch": 1.2,
      "learning_rate": 1e-05,
      "loss": 0.4147,
      "step": 2230
    },
    {
      "epoch": 1.21,
      "learning_rate": 1e-05,
      "loss": 0.418,
      "step": 2240
    },
    {
      "epoch": 1.21,
      "learning_rate": 1e-05,
      "loss": 0.4115,
      "step": 2250
    },
    {
      "epoch": 1.22,
      "learning_rate": 1e-05,
      "loss": 0.4111,
      "step": 2260
    },
    {
      "epoch": 1.23,
      "learning_rate": 1e-05,
      "loss": 0.4124,
      "step": 2270
    },
    {
      "epoch": 1.23,
      "learning_rate": 1e-05,
      "loss": 0.4162,
      "step": 2280
    },
    {
      "epoch": 1.24,
      "learning_rate": 1e-05,
      "loss": 0.4148,
      "step": 2290
    },
    {
      "epoch": 1.24,
      "learning_rate": 1e-05,
      "loss": 0.4171,
      "step": 2300
    },
    {
      "epoch": 1.25,
      "learning_rate": 1e-05,
      "loss": 0.4151,
      "step": 2310
    },
    {
      "epoch": 1.25,
      "learning_rate": 1e-05,
      "loss": 0.4177,
      "step": 2320
    },
    {
      "epoch": 1.26,
      "learning_rate": 1e-05,
      "loss": 0.4116,
      "step": 2330
    },
    {
      "epoch": 1.26,
      "learning_rate": 1e-05,
      "loss": 0.4077,
      "step": 2340
    },
    {
      "epoch": 1.27,
      "learning_rate": 1e-05,
      "loss": 0.4139,
      "step": 2350
    },
    {
      "epoch": 1.27,
      "learning_rate": 1e-05,
      "loss": 0.4169,
      "step": 2360
    },
    {
      "epoch": 1.28,
      "learning_rate": 1e-05,
      "loss": 0.412,
      "step": 2370
    },
    {
      "epoch": 1.29,
      "learning_rate": 1e-05,
      "loss": 0.4141,
      "step": 2380
    },
    {
      "epoch": 1.29,
      "learning_rate": 1e-05,
      "loss": 0.4139,
      "step": 2390
    },
    {
      "epoch": 1.3,
      "learning_rate": 1e-05,
      "loss": 0.4079,
      "step": 2400
    },
    {
      "epoch": 1.3,
      "eval_val_accuracy": 0.6370666666666667,
      "eval_val_loss": 0.4390869140625,
      "eval_val_runtime": 2247.8072,
      "eval_val_samples_per_second": 4.449,
      "eval_val_steps_per_second": 0.556,
      "step": 2400
    },
    {
      "epoch": 1.3,
      "eval_test_accuracy": 0.63715,
      "eval_test_loss": 0.43794676661491394,
      "eval_test_runtime": 2249.4828,
      "eval_test_samples_per_second": 4.445,
      "eval_test_steps_per_second": 0.556,
      "step": 2400
    },
    {
      "epoch": 1.3,
      "learning_rate": 1e-05,
      "loss": 0.4165,
      "step": 2410
    },
    {
      "epoch": 1.31,
      "learning_rate": 1e-05,
      "loss": 0.4086,
      "step": 2420
    },
    {
      "epoch": 1.31,
      "learning_rate": 1e-05,
      "loss": 0.4191,
      "step": 2430
    },
    {
      "epoch": 1.32,
      "learning_rate": 1e-05,
      "loss": 0.4193,
      "step": 2440
    },
    {
      "epoch": 1.32,
      "learning_rate": 1e-05,
      "loss": 0.4115,
      "step": 2450
    },
    {
      "epoch": 1.33,
      "learning_rate": 1e-05,
      "loss": 0.4166,
      "step": 2460
    },
    {
      "epoch": 1.33,
      "learning_rate": 1e-05,
      "loss": 0.4152,
      "step": 2470
    },
    {
      "epoch": 1.34,
      "learning_rate": 1e-05,
      "loss": 0.4096,
      "step": 2480
    },
    {
      "epoch": 1.34,
      "learning_rate": 1e-05,
      "loss": 0.4123,
      "step": 2490
    },
    {
      "epoch": 1.35,
      "learning_rate": 1e-05,
      "loss": 0.4045,
      "step": 2500
    },
    {
      "epoch": 1.36,
      "learning_rate": 1e-05,
      "loss": 0.4128,
      "step": 2510
    },
    {
      "epoch": 1.36,
      "learning_rate": 1e-05,
      "loss": 0.41,
      "step": 2520
    },
    {
      "epoch": 1.37,
      "learning_rate": 1e-05,
      "loss": 0.4089,
      "step": 2530
    },
    {
      "epoch": 1.37,
      "learning_rate": 1e-05,
      "loss": 0.4104,
      "step": 2540
    },
    {
      "epoch": 1.38,
      "learning_rate": 1e-05,
      "loss": 0.4089,
      "step": 2550
    },
    {
      "epoch": 1.38,
      "learning_rate": 1e-05,
      "loss": 0.4113,
      "step": 2560
    },
    {
      "epoch": 1.39,
      "learning_rate": 1e-05,
      "loss": 0.4145,
      "step": 2570
    },
    {
      "epoch": 1.39,
      "learning_rate": 1e-05,
      "loss": 0.4093,
      "step": 2580
    },
    {
      "epoch": 1.4,
      "learning_rate": 1e-05,
      "loss": 0.4101,
      "step": 2590
    },
    {
      "epoch": 1.4,
      "learning_rate": 1e-05,
      "loss": 0.4017,
      "step": 2600
    },
    {
      "epoch": 1.4,
      "eval_val_accuracy": 0.6358666666666667,
      "eval_val_loss": 0.4389342665672302,
      "eval_val_runtime": 2245.5574,
      "eval_val_samples_per_second": 4.453,
      "eval_val_steps_per_second": 0.557,
      "step": 2600
    },
    {
      "epoch": 1.4,
      "eval_test_accuracy": 0.6374833333333333,
      "eval_test_loss": 0.43796294927597046,
      "eval_test_runtime": 2246.6417,
      "eval_test_samples_per_second": 4.451,
      "eval_test_steps_per_second": 0.556,
      "step": 2600
    },
    {
      "epoch": 1.41,
      "learning_rate": 1e-05,
      "loss": 0.415,
      "step": 2610
    },
    {
      "epoch": 1.41,
      "learning_rate": 1e-05,
      "loss": 0.409,
      "step": 2620
    },
    {
      "epoch": 1.42,
      "learning_rate": 1e-05,
      "loss": 0.4026,
      "step": 2630
    },
    {
      "epoch": 1.43,
      "learning_rate": 1e-05,
      "loss": 0.4136,
      "step": 2640
    },
    {
      "epoch": 1.43,
      "learning_rate": 1e-05,
      "loss": 0.408,
      "step": 2650
    },
    {
      "epoch": 1.44,
      "learning_rate": 1e-05,
      "loss": 0.4063,
      "step": 2660
    },
    {
      "epoch": 1.44,
      "learning_rate": 1e-05,
      "loss": 0.4079,
      "step": 2670
    },
    {
      "epoch": 1.45,
      "learning_rate": 1e-05,
      "loss": 0.4052,
      "step": 2680
    },
    {
      "epoch": 1.45,
      "learning_rate": 1e-05,
      "loss": 0.4142,
      "step": 2690
    },
    {
      "epoch": 1.46,
      "learning_rate": 1e-05,
      "loss": 0.4079,
      "step": 2700
    },
    {
      "epoch": 1.46,
      "learning_rate": 1e-05,
      "loss": 0.4108,
      "step": 2710
    },
    {
      "epoch": 1.47,
      "learning_rate": 1e-05,
      "loss": 0.4189,
      "step": 2720
    },
    {
      "epoch": 1.47,
      "learning_rate": 1e-05,
      "loss": 0.4099,
      "step": 2730
    },
    {
      "epoch": 1.48,
      "learning_rate": 1e-05,
      "loss": 0.4134,
      "step": 2740
    },
    {
      "epoch": 1.48,
      "learning_rate": 1e-05,
      "loss": 0.4084,
      "step": 2750
    },
    {
      "epoch": 1.49,
      "learning_rate": 1e-05,
      "loss": 0.4077,
      "step": 2760
    },
    {
      "epoch": 1.5,
      "learning_rate": 1e-05,
      "loss": 0.405,
      "step": 2770
    },
    {
      "epoch": 1.5,
      "learning_rate": 1e-05,
      "loss": 0.4121,
      "step": 2780
    },
    {
      "epoch": 1.51,
      "learning_rate": 1e-05,
      "loss": 0.4038,
      "step": 2790
    },
    {
      "epoch": 1.51,
      "learning_rate": 1e-05,
      "loss": 0.4087,
      "step": 2800
    },
    {
      "epoch": 1.51,
      "eval_val_accuracy": 0.6361333333333333,
      "eval_val_loss": 0.43955838680267334,
      "eval_val_runtime": 2248.9213,
      "eval_val_samples_per_second": 4.447,
      "eval_val_steps_per_second": 0.556,
      "step": 2800
    },
    {
      "epoch": 1.51,
      "eval_test_accuracy": 0.63665,
      "eval_test_loss": 0.4383476972579956,
      "eval_test_runtime": 2249.9839,
      "eval_test_samples_per_second": 4.444,
      "eval_test_steps_per_second": 0.556,
      "step": 2800
    },
    {
      "epoch": 1.52,
      "learning_rate": 1e-05,
      "loss": 0.412,
      "step": 2810
    },
    {
      "epoch": 1.52,
      "learning_rate": 1e-05,
      "loss": 0.4162,
      "step": 2820
    },
    {
      "epoch": 1.53,
      "learning_rate": 1e-05,
      "loss": 0.4083,
      "step": 2830
    },
    {
      "epoch": 1.53,
      "learning_rate": 1e-05,
      "loss": 0.4044,
      "step": 2840
    },
    {
      "epoch": 1.54,
      "learning_rate": 1e-05,
      "loss": 0.4106,
      "step": 2850
    },
    {
      "epoch": 1.54,
      "learning_rate": 1e-05,
      "loss": 0.4104,
      "step": 2860
    },
    {
      "epoch": 1.55,
      "learning_rate": 1e-05,
      "loss": 0.4119,
      "step": 2870
    },
    {
      "epoch": 1.56,
      "learning_rate": 1e-05,
      "loss": 0.4093,
      "step": 2880
    },
    {
      "epoch": 1.56,
      "learning_rate": 1e-05,
      "loss": 0.4112,
      "step": 2890
    },
    {
      "epoch": 1.57,
      "learning_rate": 1e-05,
      "loss": 0.3965,
      "step": 2900
    },
    {
      "epoch": 1.57,
      "learning_rate": 1e-05,
      "loss": 0.4056,
      "step": 2910
    },
    {
      "epoch": 1.58,
      "learning_rate": 1e-05,
      "loss": 0.4132,
      "step": 2920
    },
    {
      "epoch": 1.58,
      "learning_rate": 1e-05,
      "loss": 0.4036,
      "step": 2930
    },
    {
      "epoch": 1.59,
      "learning_rate": 1e-05,
      "loss": 0.4129,
      "step": 2940
    },
    {
      "epoch": 1.59,
      "learning_rate": 1e-05,
      "loss": 0.4023,
      "step": 2950
    },
    {
      "epoch": 1.6,
      "learning_rate": 1e-05,
      "loss": 0.4079,
      "step": 2960
    },
    {
      "epoch": 1.6,
      "learning_rate": 1e-05,
      "loss": 0.3995,
      "step": 2970
    },
    {
      "epoch": 1.61,
      "learning_rate": 1e-05,
      "loss": 0.4095,
      "step": 2980
    },
    {
      "epoch": 1.61,
      "learning_rate": 1e-05,
      "loss": 0.4093,
      "step": 2990
    },
    {
      "epoch": 1.62,
      "learning_rate": 1e-05,
      "loss": 0.4064,
      "step": 3000
    },
    {
      "epoch": 1.62,
      "eval_val_accuracy": 0.6355166666666667,
      "eval_val_loss": 0.4397803246974945,
      "eval_val_runtime": 2248.375,
      "eval_val_samples_per_second": 4.448,
      "eval_val_steps_per_second": 0.556,
      "step": 3000
    },
    {
      "epoch": 1.62,
      "eval_test_accuracy": 0.6365666666666667,
      "eval_test_loss": 0.4391345679759979,
      "eval_test_runtime": 2249.4769,
      "eval_test_samples_per_second": 4.445,
      "eval_test_steps_per_second": 0.556,
      "step": 3000
    },
    {
      "epoch": 1.63,
      "learning_rate": 1e-05,
      "loss": 0.4087,
      "step": 3010
    },
    {
      "epoch": 1.63,
      "learning_rate": 1e-05,
      "loss": 0.4032,
      "step": 3020
    },
    {
      "epoch": 1.64,
      "learning_rate": 1e-05,
      "loss": 0.4093,
      "step": 3030
    },
    {
      "epoch": 1.64,
      "learning_rate": 1e-05,
      "loss": 0.407,
      "step": 3040
    },
    {
      "epoch": 1.65,
      "learning_rate": 1e-05,
      "loss": 0.4022,
      "step": 3050
    },
    {
      "epoch": 1.65,
      "learning_rate": 1e-05,
      "loss": 0.4119,
      "step": 3060
    },
    {
      "epoch": 1.66,
      "learning_rate": 1e-05,
      "loss": 0.4055,
      "step": 3070
    },
    {
      "epoch": 1.66,
      "learning_rate": 1e-05,
      "loss": 0.4018,
      "step": 3080
    },
    {
      "epoch": 1.67,
      "learning_rate": 1e-05,
      "loss": 0.4056,
      "step": 3090
    },
    {
      "epoch": 1.67,
      "learning_rate": 1e-05,
      "loss": 0.4002,
      "step": 3100
    },
    {
      "epoch": 1.68,
      "learning_rate": 1e-05,
      "loss": 0.4102,
      "step": 3110
    },
    {
      "epoch": 1.68,
      "learning_rate": 1e-05,
      "loss": 0.4121,
      "step": 3120
    },
    {
      "epoch": 1.69,
      "learning_rate": 1e-05,
      "loss": 0.4023,
      "step": 3130
    },
    {
      "epoch": 1.7,
      "learning_rate": 1e-05,
      "loss": 0.4111,
      "step": 3140
    },
    {
      "epoch": 1.7,
      "learning_rate": 1e-05,
      "loss": 0.4048,
      "step": 3150
    },
    {
      "epoch": 1.71,
      "learning_rate": 1e-05,
      "loss": 0.4009,
      "step": 3160
    },
    {
      "epoch": 1.71,
      "learning_rate": 1e-05,
      "loss": 0.3981,
      "step": 3170
    },
    {
      "epoch": 1.72,
      "learning_rate": 1e-05,
      "loss": 0.4048,
      "step": 3180
    },
    {
      "epoch": 1.72,
      "learning_rate": 1e-05,
      "loss": 0.4084,
      "step": 3190
    },
    {
      "epoch": 1.73,
      "learning_rate": 1e-05,
      "loss": 0.4022,
      "step": 3200
    },
    {
      "epoch": 1.73,
      "eval_val_accuracy": 0.6352666666666666,
      "eval_val_loss": 0.43992578983306885,
      "eval_val_runtime": 2247.4168,
      "eval_val_samples_per_second": 4.45,
      "eval_val_steps_per_second": 0.556,
      "step": 3200
    },
    {
      "epoch": 1.73,
      "eval_test_accuracy": 0.63605,
      "eval_test_loss": 0.43864262104034424,
      "eval_test_runtime": 2248.9018,
      "eval_test_samples_per_second": 4.447,
      "eval_test_steps_per_second": 0.556,
      "step": 3200
    },
    {
      "epoch": 1.73,
      "learning_rate": 1e-05,
      "loss": 0.4113,
      "step": 3210
    },
    {
      "epoch": 1.74,
      "learning_rate": 1e-05,
      "loss": 0.4111,
      "step": 3220
    },
    {
      "epoch": 1.74,
      "learning_rate": 1e-05,
      "loss": 0.3997,
      "step": 3230
    },
    {
      "epoch": 1.75,
      "learning_rate": 1e-05,
      "loss": 0.4084,
      "step": 3240
    },
    {
      "epoch": 1.75,
      "learning_rate": 1e-05,
      "loss": 0.4024,
      "step": 3250
    },
    {
      "epoch": 1.76,
      "learning_rate": 1e-05,
      "loss": 0.4062,
      "step": 3260
    },
    {
      "epoch": 1.77,
      "learning_rate": 1e-05,
      "loss": 0.3965,
      "step": 3270
    },
    {
      "epoch": 1.77,
      "learning_rate": 1e-05,
      "loss": 0.4047,
      "step": 3280
    },
    {
      "epoch": 1.78,
      "learning_rate": 1e-05,
      "loss": 0.4053,
      "step": 3290
    },
    {
      "epoch": 1.78,
      "learning_rate": 1e-05,
      "loss": 0.3962,
      "step": 3300
    },
    {
      "epoch": 1.79,
      "learning_rate": 1e-05,
      "loss": 0.4032,
      "step": 3310
    },
    {
      "epoch": 1.79,
      "learning_rate": 1e-05,
      "loss": 0.3995,
      "step": 3320
    },
    {
      "epoch": 1.8,
      "learning_rate": 1e-05,
      "loss": 0.3986,
      "step": 3330
    },
    {
      "epoch": 1.8,
      "learning_rate": 1e-05,
      "loss": 0.3976,
      "step": 3340
    },
    {
      "epoch": 1.81,
      "learning_rate": 1e-05,
      "loss": 0.4036,
      "step": 3350
    },
    {
      "epoch": 1.81,
      "learning_rate": 1e-05,
      "loss": 0.4038,
      "step": 3360
    },
    {
      "epoch": 1.82,
      "learning_rate": 1e-05,
      "loss": 0.4066,
      "step": 3370
    },
    {
      "epoch": 1.83,
      "learning_rate": 1e-05,
      "loss": 0.4072,
      "step": 3380
    },
    {
      "epoch": 1.83,
      "learning_rate": 1e-05,
      "loss": 0.3931,
      "step": 3390
    },
    {
      "epoch": 1.84,
      "learning_rate": 1e-05,
      "loss": 0.4023,
      "step": 3400
    },
    {
      "epoch": 1.84,
      "eval_val_accuracy": 0.6358166666666667,
      "eval_val_loss": 0.44163385033607483,
      "eval_val_runtime": 2247.1606,
      "eval_val_samples_per_second": 4.45,
      "eval_val_steps_per_second": 0.556,
      "step": 3400
    },
    {
      "epoch": 1.84,
      "eval_test_accuracy": 0.6360333333333333,
      "eval_test_loss": 0.43999171257019043,
      "eval_test_runtime": 2248.1819,
      "eval_test_samples_per_second": 4.448,
      "eval_test_steps_per_second": 0.556,
      "step": 3400
    },
    {
      "epoch": 1.84,
      "learning_rate": 1e-05,
      "loss": 0.4105,
      "step": 3410
    },
    {
      "epoch": 1.85,
      "learning_rate": 1e-05,
      "loss": 0.4004,
      "step": 3420
    },
    {
      "epoch": 1.85,
      "learning_rate": 1e-05,
      "loss": 0.4037,
      "step": 3430
    },
    {
      "epoch": 1.86,
      "learning_rate": 1e-05,
      "loss": 0.4033,
      "step": 3440
    },
    {
      "epoch": 1.86,
      "learning_rate": 1e-05,
      "loss": 0.3995,
      "step": 3450
    },
    {
      "epoch": 1.87,
      "learning_rate": 1e-05,
      "loss": 0.3978,
      "step": 3460
    },
    {
      "epoch": 1.87,
      "learning_rate": 1e-05,
      "loss": 0.3984,
      "step": 3470
    },
    {
      "epoch": 1.88,
      "learning_rate": 1e-05,
      "loss": 0.4024,
      "step": 3480
    },
    {
      "epoch": 1.88,
      "learning_rate": 1e-05,
      "loss": 0.395,
      "step": 3490
    },
    {
      "epoch": 1.89,
      "learning_rate": 1e-05,
      "loss": 0.3954,
      "step": 3500
    },
    {
      "epoch": 1.9,
      "learning_rate": 1e-05,
      "loss": 0.4008,
      "step": 3510
    },
    {
      "epoch": 1.9,
      "learning_rate": 1e-05,
      "loss": 0.3973,
      "step": 3520
    },
    {
      "epoch": 1.91,
      "learning_rate": 1e-05,
      "loss": 0.4,
      "step": 3530
    },
    {
      "epoch": 1.91,
      "learning_rate": 1e-05,
      "loss": 0.3968,
      "step": 3540
    },
    {
      "epoch": 1.92,
      "learning_rate": 1e-05,
      "loss": 0.3995,
      "step": 3550
    },
    {
      "epoch": 1.92,
      "learning_rate": 1e-05,
      "loss": 0.3975,
      "step": 3560
    },
    {
      "epoch": 1.93,
      "learning_rate": 1e-05,
      "loss": 0.4002,
      "step": 3570
    },
    {
      "epoch": 1.93,
      "learning_rate": 1e-05,
      "loss": 0.3956,
      "step": 3580
    },
    {
      "epoch": 1.94,
      "learning_rate": 1e-05,
      "loss": 0.3897,
      "step": 3590
    },
    {
      "epoch": 1.94,
      "learning_rate": 1e-05,
      "loss": 0.3941,
      "step": 3600
    },
    {
      "epoch": 1.94,
      "eval_val_accuracy": 0.63455,
      "eval_val_loss": 0.44235870242118835,
      "eval_val_runtime": 2248.8702,
      "eval_val_samples_per_second": 4.447,
      "eval_val_steps_per_second": 0.556,
      "step": 3600
    },
    {
      "epoch": 1.94,
      "eval_test_accuracy": 0.6345,
      "eval_test_loss": 0.44232964515686035,
      "eval_test_runtime": 2250.1651,
      "eval_test_samples_per_second": 4.444,
      "eval_test_steps_per_second": 0.556,
      "step": 3600
    },
    {
      "epoch": 1.95,
      "learning_rate": 1e-05,
      "loss": 0.391,
      "step": 3610
    },
    {
      "epoch": 1.95,
      "learning_rate": 1e-05,
      "loss": 0.3977,
      "step": 3620
    },
    {
      "epoch": 1.96,
      "learning_rate": 1e-05,
      "loss": 0.3987,
      "step": 3630
    },
    {
      "epoch": 1.97,
      "learning_rate": 1e-05,
      "loss": 0.3922,
      "step": 3640
    },
    {
      "epoch": 1.97,
      "learning_rate": 1e-05,
      "loss": 0.4079,
      "step": 3650
    },
    {
      "epoch": 1.98,
      "learning_rate": 1e-05,
      "loss": 0.4411,
      "step": 3660
    },
    {
      "epoch": 1.98,
      "learning_rate": 1e-05,
      "loss": 0.4464,
      "step": 3670
    },
    {
      "epoch": 1.99,
      "learning_rate": 1e-05,
      "loss": 0.4466,
      "step": 3680
    },
    {
      "epoch": 1.99,
      "learning_rate": 1e-05,
      "loss": 0.4457,
      "step": 3690
    },
    {
      "epoch": 2.0,
      "learning_rate": 1e-05,
      "loss": 0.4457,
      "step": 3700
    },
    {
      "epoch": 2.0,
      "learning_rate": 1e-05,
      "loss": 0.4256,
      "step": 3710
    },
    {
      "epoch": 2.01,
      "learning_rate": 1e-05,
      "loss": 0.42,
      "step": 3720
    },
    {
      "epoch": 2.01,
      "learning_rate": 1e-05,
      "loss": 0.4279,
      "step": 3730
    },
    {
      "epoch": 2.02,
      "learning_rate": 1e-05,
      "loss": 0.4219,
      "step": 3740
    },
    {
      "epoch": 2.02,
      "learning_rate": 1e-05,
      "loss": 0.4155,
      "step": 3750
    },
    {
      "epoch": 2.03,
      "learning_rate": 1e-05,
      "loss": 0.4181,
      "step": 3760
    },
    {
      "epoch": 2.04,
      "learning_rate": 1e-05,
      "loss": 0.4159,
      "step": 3770
    },
    {
      "epoch": 2.04,
      "learning_rate": 1e-05,
      "loss": 0.4229,
      "step": 3780
    },
    {
      "epoch": 2.05,
      "learning_rate": 1e-05,
      "loss": 0.4156,
      "step": 3790
    },
    {
      "epoch": 2.05,
      "learning_rate": 1e-05,
      "loss": 0.4156,
      "step": 3800
    },
    {
      "epoch": 2.05,
      "eval_val_accuracy": 0.6359666666666667,
      "eval_val_loss": 0.43870726227760315,
      "eval_val_runtime": 2247.0507,
      "eval_val_samples_per_second": 4.45,
      "eval_val_steps_per_second": 0.556,
      "step": 3800
    },
    {
      "epoch": 2.05,
      "eval_test_accuracy": 0.6372333333333333,
      "eval_test_loss": 0.4378372132778168,
      "eval_test_runtime": 2248.256,
      "eval_test_samples_per_second": 4.448,
      "eval_test_steps_per_second": 0.556,
      "step": 3800
    },
    {
      "epoch": 2.06,
      "learning_rate": 1e-05,
      "loss": 0.4159,
      "step": 3810
    },
    {
      "epoch": 2.06,
      "learning_rate": 1e-05,
      "loss": 0.4142,
      "step": 3820
    },
    {
      "epoch": 2.07,
      "learning_rate": 1e-05,
      "loss": 0.4213,
      "step": 3830
    },
    {
      "epoch": 2.07,
      "learning_rate": 1e-05,
      "loss": 0.4147,
      "step": 3840
    },
    {
      "epoch": 2.08,
      "learning_rate": 1e-05,
      "loss": 0.4111,
      "step": 3850
    },
    {
      "epoch": 2.08,
      "learning_rate": 1e-05,
      "loss": 0.411,
      "step": 3860
    },
    {
      "epoch": 2.09,
      "learning_rate": 1e-05,
      "loss": 0.4088,
      "step": 3870
    },
    {
      "epoch": 2.1,
      "learning_rate": 1e-05,
      "loss": 0.411,
      "step": 3880
    },
    {
      "epoch": 2.1,
      "learning_rate": 1e-05,
      "loss": 0.4172,
      "step": 3890
    },
    {
      "epoch": 2.11,
      "learning_rate": 1e-05,
      "loss": 0.4154,
      "step": 3900
    },
    {
      "epoch": 2.11,
      "learning_rate": 1e-05,
      "loss": 0.4074,
      "step": 3910
    },
    {
      "epoch": 2.12,
      "learning_rate": 1e-05,
      "loss": 0.4072,
      "step": 3920
    },
    {
      "epoch": 2.12,
      "learning_rate": 1e-05,
      "loss": 0.4157,
      "step": 3930
    },
    {
      "epoch": 2.13,
      "learning_rate": 1e-05,
      "loss": 0.4066,
      "step": 3940
    },
    {
      "epoch": 2.13,
      "learning_rate": 1e-05,
      "loss": 0.4126,
      "step": 3950
    },
    {
      "epoch": 2.14,
      "learning_rate": 1e-05,
      "loss": 0.4182,
      "step": 3960
    },
    {
      "epoch": 2.14,
      "learning_rate": 1e-05,
      "loss": 0.4099,
      "step": 3970
    },
    {
      "epoch": 2.15,
      "learning_rate": 1e-05,
      "loss": 0.4104,
      "step": 3980
    },
    {
      "epoch": 2.15,
      "learning_rate": 1e-05,
      "loss": 0.413,
      "step": 3990
    },
    {
      "epoch": 2.16,
      "learning_rate": 1e-05,
      "loss": 0.412,
      "step": 4000
    },
    {
      "epoch": 2.16,
      "eval_val_accuracy": 0.63665,
      "eval_val_loss": 0.4387384355068207,
      "eval_val_runtime": 2248.7708,
      "eval_val_samples_per_second": 4.447,
      "eval_val_steps_per_second": 0.556,
      "step": 4000
    },
    {
      "epoch": 2.16,
      "eval_test_accuracy": 0.63735,
      "eval_test_loss": 0.4378214478492737,
      "eval_test_runtime": 2249.9861,
      "eval_test_samples_per_second": 4.444,
      "eval_test_steps_per_second": 0.556,
      "step": 4000
    },
    {
      "epoch": 2.17,
      "learning_rate": 1e-05,
      "loss": 0.4156,
      "step": 4010
    },
    {
      "epoch": 2.17,
      "learning_rate": 1e-05,
      "loss": 0.4152,
      "step": 4020
    },
    {
      "epoch": 2.18,
      "learning_rate": 1e-05,
      "loss": 0.4076,
      "step": 4030
    },
    {
      "epoch": 2.18,
      "learning_rate": 1e-05,
      "loss": 0.4146,
      "step": 4040
    },
    {
      "epoch": 2.19,
      "learning_rate": 1e-05,
      "loss": 0.4185,
      "step": 4050
    },
    {
      "epoch": 2.19,
      "learning_rate": 1e-05,
      "loss": 0.4192,
      "step": 4060
    },
    {
      "epoch": 2.2,
      "learning_rate": 1e-05,
      "loss": 0.4129,
      "step": 4070
    },
    {
      "epoch": 2.2,
      "learning_rate": 1e-05,
      "loss": 0.4127,
      "step": 4080
    },
    {
      "epoch": 2.21,
      "learning_rate": 1e-05,
      "loss": 0.4143,
      "step": 4090
    },
    {
      "epoch": 2.21,
      "learning_rate": 1e-05,
      "loss": 0.4114,
      "step": 4100
    },
    {
      "epoch": 2.22,
      "learning_rate": 1e-05,
      "loss": 0.4067,
      "step": 4110
    },
    {
      "epoch": 2.22,
      "learning_rate": 1e-05,
      "loss": 0.4104,
      "step": 4120
    },
    {
      "epoch": 2.23,
      "learning_rate": 1e-05,
      "loss": 0.4111,
      "step": 4130
    },
    {
      "epoch": 2.24,
      "learning_rate": 1e-05,
      "loss": 0.4092,
      "step": 4140
    },
    {
      "epoch": 2.24,
      "learning_rate": 1e-05,
      "loss": 0.4186,
      "step": 4150
    },
    {
      "epoch": 2.25,
      "learning_rate": 1e-05,
      "loss": 0.4052,
      "step": 4160
    },
    {
      "epoch": 2.25,
      "learning_rate": 1e-05,
      "loss": 0.4148,
      "step": 4170
    },
    {
      "epoch": 2.26,
      "learning_rate": 1e-05,
      "loss": 0.4093,
      "step": 4180
    },
    {
      "epoch": 2.26,
      "learning_rate": 1e-05,
      "loss": 0.4081,
      "step": 4190
    },
    {
      "epoch": 2.27,
      "learning_rate": 1e-05,
      "loss": 0.4133,
      "step": 4200
    },
    {
      "epoch": 2.27,
      "eval_val_accuracy": 0.63635,
      "eval_val_loss": 0.4388815462589264,
      "eval_val_runtime": 2248.6122,
      "eval_val_samples_per_second": 4.447,
      "eval_val_steps_per_second": 0.556,
      "step": 4200
    },
    {
      "epoch": 2.27,
      "eval_test_accuracy": 0.6375166666666667,
      "eval_test_loss": 0.4379238188266754,
      "eval_test_runtime": 2249.9685,
      "eval_test_samples_per_second": 4.445,
      "eval_test_steps_per_second": 0.556,
      "step": 4200
    },
    {
      "epoch": 2.27,
      "learning_rate": 1e-05,
      "loss": 0.4127,
      "step": 4210
    },
    {
      "epoch": 2.28,
      "learning_rate": 1e-05,
      "loss": 0.41,
      "step": 4220
    },
    {
      "epoch": 2.28,
      "learning_rate": 1e-05,
      "loss": 0.4115,
      "step": 4230
    },
    {
      "epoch": 2.29,
      "learning_rate": 1e-05,
      "loss": 0.4087,
      "step": 4240
    },
    {
      "epoch": 2.29,
      "learning_rate": 1e-05,
      "loss": 0.41,
      "step": 4250
    },
    {
      "epoch": 2.3,
      "learning_rate": 1e-05,
      "loss": 0.4068,
      "step": 4260
    },
    {
      "epoch": 2.31,
      "learning_rate": 1e-05,
      "loss": 0.4066,
      "step": 4270
    },
    {
      "epoch": 2.31,
      "learning_rate": 1e-05,
      "loss": 0.4145,
      "step": 4280
    },
    {
      "epoch": 2.32,
      "learning_rate": 1e-05,
      "loss": 0.4175,
      "step": 4290
    },
    {
      "epoch": 2.32,
      "learning_rate": 1e-05,
      "loss": 0.4112,
      "step": 4300
    },
    {
      "epoch": 2.33,
      "learning_rate": 1e-05,
      "loss": 0.4129,
      "step": 4310
    },
    {
      "epoch": 2.33,
      "learning_rate": 1e-05,
      "loss": 0.4129,
      "step": 4320
    },
    {
      "epoch": 2.34,
      "learning_rate": 1e-05,
      "loss": 0.4045,
      "step": 4330
    },
    {
      "epoch": 2.34,
      "learning_rate": 1e-05,
      "loss": 0.4112,
      "step": 4340
    },
    {
      "epoch": 2.35,
      "learning_rate": 1e-05,
      "loss": 0.4038,
      "step": 4350
    },
    {
      "epoch": 2.35,
      "learning_rate": 1e-05,
      "loss": 0.4069,
      "step": 4360
    },
    {
      "epoch": 2.36,
      "learning_rate": 1e-05,
      "loss": 0.4085,
      "step": 4370
    },
    {
      "epoch": 2.37,
      "learning_rate": 1e-05,
      "loss": 0.4097,
      "step": 4380
    },
    {
      "epoch": 2.37,
      "learning_rate": 1e-05,
      "loss": 0.4036,
      "step": 4390
    },
    {
      "epoch": 2.38,
      "learning_rate": 1e-05,
      "loss": 0.4075,
      "step": 4400
    },
    {
      "epoch": 2.38,
      "eval_val_accuracy": 0.63655,
      "eval_val_loss": 0.4389641582965851,
      "eval_val_runtime": 2248.6386,
      "eval_val_samples_per_second": 4.447,
      "eval_val_steps_per_second": 0.556,
      "step": 4400
    },
    {
      "epoch": 2.38,
      "eval_test_accuracy": 0.6379,
      "eval_test_loss": 0.4379016160964966,
      "eval_test_runtime": 2249.1414,
      "eval_test_samples_per_second": 4.446,
      "eval_test_steps_per_second": 0.556,
      "step": 4400
    },
    {
      "epoch": 2.38,
      "learning_rate": 1e-05,
      "loss": 0.4062,
      "step": 4410
    },
    {
      "epoch": 2.39,
      "learning_rate": 1e-05,
      "loss": 0.4106,
      "step": 4420
    },
    {
      "epoch": 2.39,
      "learning_rate": 1e-05,
      "loss": 0.408,
      "step": 4430
    },
    {
      "epoch": 2.4,
      "learning_rate": 1e-05,
      "loss": 0.4068,
      "step": 4440
    },
    {
      "epoch": 2.4,
      "learning_rate": 1e-05,
      "loss": 0.4058,
      "step": 4450
    },
    {
      "epoch": 2.41,
      "learning_rate": 1e-05,
      "loss": 0.4092,
      "step": 4460
    },
    {
      "epoch": 2.41,
      "learning_rate": 1e-05,
      "loss": 0.4168,
      "step": 4470
    },
    {
      "epoch": 2.42,
      "learning_rate": 1e-05,
      "loss": 0.406,
      "step": 4480
    },
    {
      "epoch": 2.42,
      "learning_rate": 1e-05,
      "loss": 0.412,
      "step": 4490
    },
    {
      "epoch": 2.43,
      "learning_rate": 1e-05,
      "loss": 0.4095,
      "step": 4500
    },
    {
      "epoch": 2.44,
      "learning_rate": 1e-05,
      "loss": 0.411,
      "step": 4510
    },
    {
      "epoch": 2.44,
      "learning_rate": 1e-05,
      "loss": 0.4105,
      "step": 4520
    },
    {
      "epoch": 2.45,
      "learning_rate": 1e-05,
      "loss": 0.4105,
      "step": 4530
    },
    {
      "epoch": 2.45,
      "learning_rate": 1e-05,
      "loss": 0.413,
      "step": 4540
    },
    {
      "epoch": 2.46,
      "learning_rate": 1e-05,
      "loss": 0.4134,
      "step": 4550
    },
    {
      "epoch": 2.46,
      "learning_rate": 1e-05,
      "loss": 0.4145,
      "step": 4560
    },
    {
      "epoch": 2.47,
      "learning_rate": 1e-05,
      "loss": 0.418,
      "step": 4570
    },
    {
      "epoch": 2.47,
      "learning_rate": 1e-05,
      "loss": 0.4157,
      "step": 4580
    },
    {
      "epoch": 2.48,
      "learning_rate": 1e-05,
      "loss": 0.4149,
      "step": 4590
    },
    {
      "epoch": 2.48,
      "learning_rate": 1e-05,
      "loss": 0.4128,
      "step": 4600
    },
    {
      "epoch": 2.48,
      "eval_val_accuracy": 0.6363666666666667,
      "eval_val_loss": 0.43885067105293274,
      "eval_val_runtime": 2248.1145,
      "eval_val_samples_per_second": 4.448,
      "eval_val_steps_per_second": 0.556,
      "step": 4600
    },
    {
      "epoch": 2.48,
      "eval_test_accuracy": 0.6380166666666667,
      "eval_test_loss": 0.43778306245803833,
      "eval_test_runtime": 2249.4486,
      "eval_test_samples_per_second": 4.446,
      "eval_test_steps_per_second": 0.556,
      "step": 4600
    },
    {
      "epoch": 2.49,
      "learning_rate": 1e-05,
      "loss": 0.4095,
      "step": 4610
    },
    {
      "epoch": 2.49,
      "learning_rate": 1e-05,
      "loss": 0.4111,
      "step": 4620
    },
    {
      "epoch": 2.5,
      "learning_rate": 1e-05,
      "loss": 0.4155,
      "step": 4630
    },
    {
      "epoch": 2.51,
      "learning_rate": 1e-05,
      "loss": 0.4047,
      "step": 4640
    },
    {
      "epoch": 2.51,
      "learning_rate": 1e-05,
      "loss": 0.4178,
      "step": 4650
    },
    {
      "epoch": 2.52,
      "learning_rate": 1e-05,
      "loss": 0.4129,
      "step": 4660
    },
    {
      "epoch": 2.52,
      "learning_rate": 1e-05,
      "loss": 0.4192,
      "step": 4670
    },
    {
      "epoch": 2.53,
      "learning_rate": 1e-05,
      "loss": 0.4127,
      "step": 4680
    },
    {
      "epoch": 2.53,
      "learning_rate": 1e-05,
      "loss": 0.4083,
      "step": 4690
    },
    {
      "epoch": 2.54,
      "learning_rate": 1e-05,
      "loss": 0.4159,
      "step": 4700
    },
    {
      "epoch": 2.54,
      "learning_rate": 1e-05,
      "loss": 0.4144,
      "step": 4710
    },
    {
      "epoch": 2.55,
      "learning_rate": 1e-05,
      "loss": 0.4147,
      "step": 4720
    },
    {
      "epoch": 2.55,
      "learning_rate": 1e-05,
      "loss": 0.418,
      "step": 4730
    },
    {
      "epoch": 2.56,
      "learning_rate": 1e-05,
      "loss": 0.4113,
      "step": 4740
    },
    {
      "epoch": 2.56,
      "learning_rate": 1e-05,
      "loss": 0.4021,
      "step": 4750
    },
    {
      "epoch": 2.57,
      "learning_rate": 1e-05,
      "loss": 0.4114,
      "step": 4760
    },
    {
      "epoch": 2.58,
      "learning_rate": 1e-05,
      "loss": 0.4122,
      "step": 4770
    },
    {
      "epoch": 2.58,
      "learning_rate": 1e-05,
      "loss": 0.4128,
      "step": 4780
    },
    {
      "epoch": 2.59,
      "learning_rate": 1e-05,
      "loss": 0.4131,
      "step": 4790
    },
    {
      "epoch": 2.59,
      "learning_rate": 1e-05,
      "loss": 0.4092,
      "step": 4800
    },
    {
      "epoch": 2.59,
      "eval_val_accuracy": 0.6356166666666667,
      "eval_val_loss": 0.4385939836502075,
      "eval_val_runtime": 2245.6742,
      "eval_val_samples_per_second": 4.453,
      "eval_val_steps_per_second": 0.557,
      "step": 4800
    },
    {
      "epoch": 2.59,
      "eval_test_accuracy": 0.6375666666666667,
      "eval_test_loss": 0.437671035528183,
      "eval_test_runtime": 2246.5821,
      "eval_test_samples_per_second": 4.451,
      "eval_test_steps_per_second": 0.556,
      "step": 4800
    },
    {
      "epoch": 2.6,
      "learning_rate": 1e-05,
      "loss": 0.4109,
      "step": 4810
    },
    {
      "epoch": 2.6,
      "learning_rate": 1e-05,
      "loss": 0.407,
      "step": 4820
    },
    {
      "epoch": 2.61,
      "learning_rate": 1e-05,
      "loss": 0.4106,
      "step": 4830
    },
    {
      "epoch": 2.61,
      "learning_rate": 1e-05,
      "loss": 0.4144,
      "step": 4840
    },
    {
      "epoch": 2.62,
      "learning_rate": 1e-05,
      "loss": 0.4129,
      "step": 4850
    },
    {
      "epoch": 2.62,
      "learning_rate": 1e-05,
      "loss": 0.4124,
      "step": 4860
    },
    {
      "epoch": 2.63,
      "learning_rate": 1e-05,
      "loss": 0.4142,
      "step": 4870
    },
    {
      "epoch": 2.64,
      "learning_rate": 1e-05,
      "loss": 0.4158,
      "step": 4880
    },
    {
      "epoch": 2.64,
      "learning_rate": 1e-05,
      "loss": 0.4097,
      "step": 4890
    },
    {
      "epoch": 2.65,
      "learning_rate": 1e-05,
      "loss": 0.4075,
      "step": 4900
    },
    {
      "epoch": 2.65,
      "learning_rate": 1e-05,
      "loss": 0.4153,
      "step": 4910
    },
    {
      "epoch": 2.66,
      "learning_rate": 1e-05,
      "loss": 0.4127,
      "step": 4920
    },
    {
      "epoch": 2.66,
      "learning_rate": 1e-05,
      "loss": 0.4071,
      "step": 4930
    },
    {
      "epoch": 2.67,
      "learning_rate": 1e-05,
      "loss": 0.4111,
      "step": 4940
    },
    {
      "epoch": 2.67,
      "learning_rate": 1e-05,
      "loss": 0.4077,
      "step": 4950
    },
    {
      "epoch": 2.68,
      "learning_rate": 1e-05,
      "loss": 0.4092,
      "step": 4960
    },
    {
      "epoch": 2.68,
      "learning_rate": 1e-05,
      "loss": 0.419,
      "step": 4970
    },
    {
      "epoch": 2.69,
      "learning_rate": 1e-05,
      "loss": 0.4113,
      "step": 4980
    },
    {
      "epoch": 2.69,
      "learning_rate": 1e-05,
      "loss": 0.4146,
      "step": 4990
    },
    {
      "epoch": 2.7,
      "learning_rate": 1e-05,
      "loss": 0.412,
      "step": 5000
    },
    {
      "epoch": 2.7,
      "eval_val_accuracy": 0.6359,
      "eval_val_loss": 0.43852442502975464,
      "eval_val_runtime": 2247.8472,
      "eval_val_samples_per_second": 4.449,
      "eval_val_steps_per_second": 0.556,
      "step": 5000
    },
    {
      "epoch": 2.7,
      "eval_test_accuracy": 0.6374166666666667,
      "eval_test_loss": 0.43765121698379517,
      "eval_test_runtime": 2249.8735,
      "eval_test_samples_per_second": 4.445,
      "eval_test_steps_per_second": 0.556,
      "step": 5000
    },
    {
      "epoch": 2.71,
      "learning_rate": 1e-05,
      "loss": 0.4063,
      "step": 5010
    },
    {
      "epoch": 2.71,
      "learning_rate": 1e-05,
      "loss": 0.4041,
      "step": 5020
    },
    {
      "epoch": 2.72,
      "learning_rate": 1e-05,
      "loss": 0.4079,
      "step": 5030
    },
    {
      "epoch": 2.72,
      "learning_rate": 1e-05,
      "loss": 0.4147,
      "step": 5040
    },
    {
      "epoch": 2.73,
      "learning_rate": 1e-05,
      "loss": 0.4083,
      "step": 5050
    },
    {
      "epoch": 2.73,
      "learning_rate": 1e-05,
      "loss": 0.4186,
      "step": 5060
    },
    {
      "epoch": 2.74,
      "learning_rate": 1e-05,
      "loss": 0.416,
      "step": 5070
    },
    {
      "epoch": 2.74,
      "learning_rate": 1e-05,
      "loss": 0.4101,
      "step": 5080
    },
    {
      "epoch": 2.75,
      "learning_rate": 1e-05,
      "loss": 0.4126,
      "step": 5090
    },
    {
      "epoch": 2.75,
      "learning_rate": 1e-05,
      "loss": 0.4081,
      "step": 5100
    },
    {
      "epoch": 2.76,
      "learning_rate": 1e-05,
      "loss": 0.4139,
      "step": 5110
    },
    {
      "epoch": 2.76,
      "learning_rate": 1e-05,
      "loss": 0.4055,
      "step": 5120
    },
    {
      "epoch": 2.77,
      "learning_rate": 1e-05,
      "loss": 0.4061,
      "step": 5130
    },
    {
      "epoch": 2.78,
      "learning_rate": 1e-05,
      "loss": 0.4147,
      "step": 5140
    },
    {
      "epoch": 2.78,
      "learning_rate": 1e-05,
      "loss": 0.4015,
      "step": 5150
    },
    {
      "epoch": 2.79,
      "learning_rate": 1e-05,
      "loss": 0.4088,
      "step": 5160
    },
    {
      "epoch": 2.79,
      "learning_rate": 1e-05,
      "loss": 0.4093,
      "step": 5170
    },
    {
      "epoch": 2.8,
      "learning_rate": 1e-05,
      "loss": 0.408,
      "step": 5180
    },
    {
      "epoch": 2.8,
      "learning_rate": 1e-05,
      "loss": 0.4031,
      "step": 5190
    },
    {
      "epoch": 2.81,
      "learning_rate": 1e-05,
      "loss": 0.4051,
      "step": 5200
    },
    {
      "epoch": 2.81,
      "eval_val_accuracy": 0.63635,
      "eval_val_loss": 0.4389340877532959,
      "eval_val_runtime": 2248.6302,
      "eval_val_samples_per_second": 4.447,
      "eval_val_steps_per_second": 0.556,
      "step": 5200
    },
    {
      "epoch": 2.81,
      "eval_test_accuracy": 0.63715,
      "eval_test_loss": 0.4379483759403229,
      "eval_test_runtime": 2249.8742,
      "eval_test_samples_per_second": 4.445,
      "eval_test_steps_per_second": 0.556,
      "step": 5200
    },
    {
      "epoch": 2.81,
      "learning_rate": 1e-05,
      "loss": 0.4098,
      "step": 5210
    },
    {
      "epoch": 2.82,
      "learning_rate": 1e-05,
      "loss": 0.4153,
      "step": 5220
    },
    {
      "epoch": 2.82,
      "learning_rate": 1e-05,
      "loss": 0.4157,
      "step": 5230
    },
    {
      "epoch": 2.83,
      "learning_rate": 1e-05,
      "loss": 0.4028,
      "step": 5240
    },
    {
      "epoch": 2.83,
      "learning_rate": 1e-05,
      "loss": 0.4057,
      "step": 5250
    },
    {
      "epoch": 2.84,
      "learning_rate": 1e-05,
      "loss": 0.418,
      "step": 5260
    },
    {
      "epoch": 2.85,
      "learning_rate": 1e-05,
      "loss": 0.4052,
      "step": 5270
    },
    {
      "epoch": 2.85,
      "learning_rate": 1e-05,
      "loss": 0.41,
      "step": 5280
    },
    {
      "epoch": 2.86,
      "learning_rate": 1e-05,
      "loss": 0.4114,
      "step": 5290
    },
    {
      "epoch": 2.86,
      "learning_rate": 1e-05,
      "loss": 0.4082,
      "step": 5300
    },
    {
      "epoch": 2.87,
      "learning_rate": 1e-05,
      "loss": 0.4049,
      "step": 5310
    },
    {
      "epoch": 2.87,
      "learning_rate": 1e-05,
      "loss": 0.4064,
      "step": 5320
    },
    {
      "epoch": 2.88,
      "learning_rate": 1e-05,
      "loss": 0.4104,
      "step": 5330
    },
    {
      "epoch": 2.88,
      "learning_rate": 1e-05,
      "loss": 0.3998,
      "step": 5340
    },
    {
      "epoch": 2.89,
      "learning_rate": 1e-05,
      "loss": 0.4025,
      "step": 5350
    },
    {
      "epoch": 2.89,
      "learning_rate": 1e-05,
      "loss": 0.4143,
      "step": 5360
    },
    {
      "epoch": 2.9,
      "learning_rate": 1e-05,
      "loss": 0.4021,
      "step": 5370
    },
    {
      "epoch": 2.91,
      "learning_rate": 1e-05,
      "loss": 0.4088,
      "step": 5380
    },
    {
      "epoch": 2.91,
      "learning_rate": 1e-05,
      "loss": 0.4078,
      "step": 5390
    },
    {
      "epoch": 2.92,
      "learning_rate": 1e-05,
      "loss": 0.4075,
      "step": 5400
    },
    {
      "epoch": 2.92,
      "eval_val_accuracy": 0.6363333333333333,
      "eval_val_loss": 0.4385863244533539,
      "eval_val_runtime": 2248.4881,
      "eval_val_samples_per_second": 4.447,
      "eval_val_steps_per_second": 0.556,
      "step": 5400
    },
    {
      "epoch": 2.92,
      "eval_test_accuracy": 0.6374166666666667,
      "eval_test_loss": 0.438106507062912,
      "eval_test_runtime": 2249.7964,
      "eval_test_samples_per_second": 4.445,
      "eval_test_steps_per_second": 0.556,
      "step": 5400
    },
    {
      "epoch": 2.92,
      "learning_rate": 1e-05,
      "loss": 0.4062,
      "step": 5410
    },
    {
      "epoch": 2.93,
      "learning_rate": 1e-05,
      "loss": 0.4064,
      "step": 5420
    },
    {
      "epoch": 2.93,
      "learning_rate": 1e-05,
      "loss": 0.4042,
      "step": 5430
    },
    {
      "epoch": 2.94,
      "learning_rate": 1e-05,
      "loss": 0.4046,
      "step": 5440
    },
    {
      "epoch": 2.94,
      "learning_rate": 1e-05,
      "loss": 0.4029,
      "step": 5450
    },
    {
      "epoch": 2.95,
      "learning_rate": 1e-05,
      "loss": 0.4012,
      "step": 5460
    },
    {
      "epoch": 2.95,
      "learning_rate": 1e-05,
      "loss": 0.4081,
      "step": 5470
    },
    {
      "epoch": 2.96,
      "learning_rate": 1e-05,
      "loss": 0.4052,
      "step": 5480
    },
    {
      "epoch": 2.96,
      "learning_rate": 1e-05,
      "loss": 0.397,
      "step": 5490
    },
    {
      "epoch": 2.97,
      "learning_rate": 1e-05,
      "loss": 0.4089,
      "step": 5500
    },
    {
      "epoch": 2.98,
      "learning_rate": 1e-05,
      "loss": 0.4272,
      "step": 5510
    },
    {
      "epoch": 2.98,
      "learning_rate": 1e-05,
      "loss": 0.4351,
      "step": 5520
    },
    {
      "epoch": 2.99,
      "learning_rate": 1e-05,
      "loss": 0.4342,
      "step": 5530
    },
    {
      "epoch": 2.99,
      "learning_rate": 1e-05,
      "loss": 0.4391,
      "step": 5540
    },
    {
      "epoch": 3.0,
      "learning_rate": 1e-05,
      "loss": 0.4306,
      "step": 5550
    },
    {
      "epoch": 3.0,
      "learning_rate": 1e-05,
      "loss": 0.4294,
      "step": 5560
    },
    {
      "epoch": 3.01,
      "learning_rate": 1e-05,
      "loss": 0.4197,
      "step": 5570
    },
    {
      "epoch": 3.01,
      "learning_rate": 1e-05,
      "loss": 0.4256,
      "step": 5580
    },
    {
      "epoch": 3.02,
      "learning_rate": 1e-05,
      "loss": 0.4225,
      "step": 5590
    },
    {
      "epoch": 3.02,
      "learning_rate": 1e-05,
      "loss": 0.4139,
      "step": 5600
    },
    {
      "epoch": 3.02,
      "eval_val_accuracy": 0.6360166666666667,
      "eval_val_loss": 0.4384986460208893,
      "eval_val_runtime": 2248.2925,
      "eval_val_samples_per_second": 4.448,
      "eval_val_steps_per_second": 0.556,
      "step": 5600
    },
    {
      "epoch": 3.02,
      "eval_test_accuracy": 0.6375666666666667,
      "eval_test_loss": 0.43760040402412415,
      "eval_test_runtime": 2249.9023,
      "eval_test_samples_per_second": 4.445,
      "eval_test_steps_per_second": 0.556,
      "step": 5600
    }
  ],
  "logging_steps": 10,
  "max_steps": 9255,
  "num_train_epochs": 5,
  "save_steps": 200,
  "total_flos": 0.0,
  "trial_name": null,
  "trial_params": null
}