{
  "best_metric": 0.19156721234321594,
  "best_model_checkpoint": "./ryan03312024_lr_2e-5_wd_001/checkpoint-3100",
  "epoch": 1.5001803101334295,
  "eval_steps": 100,
  "global_step": 4160,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.01,
      "grad_norm": 0.6417293548583984,
      "learning_rate": 1.9879807692307693e-05,
      "loss": 0.5741,
      "step": 25
    },
    {
      "epoch": 0.02,
      "grad_norm": 1.4069948196411133,
      "learning_rate": 1.975961538461539e-05,
      "loss": 0.4638,
      "step": 50
    },
    {
      "epoch": 0.03,
      "grad_norm": 1.5593534708023071,
      "learning_rate": 1.963942307692308e-05,
      "loss": 0.4114,
      "step": 75
    },
    {
      "epoch": 0.04,
      "grad_norm": 0.6675819158554077,
      "learning_rate": 1.9519230769230772e-05,
      "loss": 0.4436,
      "step": 100
    },
    {
      "epoch": 0.04,
      "eval_loss": 0.3697698712348938,
      "eval_na_accuracy": 0.7989690899848938,
      "eval_ordinal_accuracy": 0.3331620991230011,
      "eval_ordinal_mae": 0.8705630302429199,
      "eval_runtime": 346.2992,
      "eval_samples_per_second": 12.922,
      "eval_steps_per_second": 1.617,
      "step": 100
    },
    {
      "epoch": 0.05,
      "grad_norm": 0.6941640377044678,
      "learning_rate": 1.9399038461538464e-05,
      "loss": 0.3901,
      "step": 125
    },
    {
      "epoch": 0.05,
      "grad_norm": 0.7081687450408936,
      "learning_rate": 1.9278846153846155e-05,
      "loss": 0.366,
      "step": 150
    },
    {
      "epoch": 0.06,
      "grad_norm": 0.9151293635368347,
      "learning_rate": 1.9158653846153847e-05,
      "loss": 0.3773,
      "step": 175
    },
    {
      "epoch": 0.07,
      "grad_norm": 0.5204830169677734,
      "learning_rate": 1.903846153846154e-05,
      "loss": 0.3143,
      "step": 200
    },
    {
      "epoch": 0.07,
      "eval_loss": 0.3215162754058838,
      "eval_na_accuracy": 0.8092783689498901,
      "eval_ordinal_accuracy": 0.4017467200756073,
      "eval_ordinal_mae": 0.855476975440979,
      "eval_runtime": 209.7588,
      "eval_samples_per_second": 21.334,
      "eval_steps_per_second": 2.67,
      "step": 200
    },
    {
      "epoch": 0.08,
      "grad_norm": 1.357649803161621,
      "learning_rate": 1.8918269230769234e-05,
      "loss": 0.3559,
      "step": 225
    },
    {
      "epoch": 0.09,
      "grad_norm": 0.7635074257850647,
      "learning_rate": 1.8798076923076926e-05,
      "loss": 0.3616,
      "step": 250
    },
    {
      "epoch": 0.1,
      "grad_norm": 2.4480478763580322,
      "learning_rate": 1.8677884615384617e-05,
      "loss": 0.3209,
      "step": 275
    },
    {
      "epoch": 0.11,
      "grad_norm": 0.8215653300285339,
      "learning_rate": 1.855769230769231e-05,
      "loss": 0.3385,
      "step": 300
    },
    {
      "epoch": 0.11,
      "eval_loss": 0.2996984124183655,
      "eval_na_accuracy": 0.8591065406799316,
      "eval_ordinal_accuracy": 0.44849729537963867,
      "eval_ordinal_mae": 0.8302922248840332,
      "eval_runtime": 193.7095,
      "eval_samples_per_second": 23.102,
      "eval_steps_per_second": 2.891,
      "step": 300
    },
    {
      "epoch": 0.12,
      "grad_norm": 0.48835399746894836,
      "learning_rate": 1.84375e-05,
      "loss": 0.2976,
      "step": 325
    },
    {
      "epoch": 0.13,
      "grad_norm": 0.6756967306137085,
      "learning_rate": 1.8317307692307693e-05,
      "loss": 0.3043,
      "step": 350
    },
    {
      "epoch": 0.14,
      "grad_norm": 0.20951023697853088,
      "learning_rate": 1.8197115384615388e-05,
      "loss": 0.3046,
      "step": 375
    },
    {
      "epoch": 0.14,
      "grad_norm": 0.9027990102767944,
      "learning_rate": 1.807692307692308e-05,
      "loss": 0.3127,
      "step": 400
    },
    {
      "epoch": 0.14,
      "eval_loss": 0.28894639015197754,
      "eval_na_accuracy": 0.8745704293251038,
      "eval_ordinal_accuracy": 0.4880554974079132,
      "eval_ordinal_mae": 0.8012504577636719,
      "eval_runtime": 195.2283,
      "eval_samples_per_second": 22.922,
      "eval_steps_per_second": 2.868,
      "step": 400
    },
    {
      "epoch": 0.15,
      "grad_norm": 1.5711051225662231,
      "learning_rate": 1.795673076923077e-05,
      "loss": 0.3094,
      "step": 425
    },
    {
      "epoch": 0.16,
      "grad_norm": 2.6442267894744873,
      "learning_rate": 1.7836538461538463e-05,
      "loss": 0.2979,
      "step": 450
    },
    {
      "epoch": 0.17,
      "grad_norm": 1.7498008012771606,
      "learning_rate": 1.7716346153846155e-05,
      "loss": 0.3193,
      "step": 475
    },
    {
      "epoch": 0.18,
      "grad_norm": 1.4141407012939453,
      "learning_rate": 1.7596153846153846e-05,
      "loss": 0.3054,
      "step": 500
    },
    {
      "epoch": 0.18,
      "eval_loss": 0.28038087487220764,
      "eval_na_accuracy": 0.8780068755149841,
      "eval_ordinal_accuracy": 0.5324942469596863,
      "eval_ordinal_mae": 0.7619425058364868,
      "eval_runtime": 189.7487,
      "eval_samples_per_second": 23.584,
      "eval_steps_per_second": 2.951,
      "step": 500
    },
    {
      "epoch": 0.19,
      "grad_norm": 1.3073471784591675,
      "learning_rate": 1.7475961538461538e-05,
      "loss": 0.302,
      "step": 525
    },
    {
      "epoch": 0.2,
      "grad_norm": 1.1622358560562134,
      "learning_rate": 1.7355769230769233e-05,
      "loss": 0.2844,
      "step": 550
    },
    {
      "epoch": 0.21,
      "grad_norm": 0.3688335716724396,
      "learning_rate": 1.7235576923076925e-05,
      "loss": 0.2745,
      "step": 575
    },
    {
      "epoch": 0.22,
      "grad_norm": 1.1387437582015991,
      "learning_rate": 1.7115384615384617e-05,
      "loss": 0.3051,
      "step": 600
    },
    {
      "epoch": 0.22,
      "eval_loss": 0.27521631121635437,
      "eval_na_accuracy": 0.9158075451850891,
      "eval_ordinal_accuracy": 0.5235037207603455,
      "eval_ordinal_mae": 0.7215057015419006,
      "eval_runtime": 192.1403,
      "eval_samples_per_second": 23.29,
      "eval_steps_per_second": 2.915,
      "step": 600
    },
    {
      "epoch": 0.23,
      "grad_norm": 0.613787055015564,
      "learning_rate": 1.699519230769231e-05,
      "loss": 0.3017,
      "step": 625
    },
    {
      "epoch": 0.23,
      "grad_norm": 0.595897376537323,
      "learning_rate": 1.6875e-05,
      "loss": 0.2806,
      "step": 650
    },
    {
      "epoch": 0.24,
      "grad_norm": 1.4324086904525757,
      "learning_rate": 1.6754807692307692e-05,
      "loss": 0.256,
      "step": 675
    },
    {
      "epoch": 0.25,
      "grad_norm": 3.7929930686950684,
      "learning_rate": 1.6634615384615387e-05,
      "loss": 0.2833,
      "step": 700
    },
    {
      "epoch": 0.25,
      "eval_loss": 0.26531103253364563,
      "eval_na_accuracy": 0.8969072103500366,
      "eval_ordinal_accuracy": 0.5486770868301392,
      "eval_ordinal_mae": 0.6806999444961548,
      "eval_runtime": 191.9358,
      "eval_samples_per_second": 23.315,
      "eval_steps_per_second": 2.918,
      "step": 700
    },
    {
      "epoch": 0.26,
      "grad_norm": 0.27854958176612854,
      "learning_rate": 1.651442307692308e-05,
      "loss": 0.2805,
      "step": 725
    },
    {
      "epoch": 0.27,
      "grad_norm": 1.8652201890945435,
      "learning_rate": 1.6399038461538462e-05,
      "loss": 0.2737,
      "step": 750
    },
    {
      "epoch": 0.28,
      "grad_norm": 2.1118507385253906,
      "learning_rate": 1.6278846153846154e-05,
      "loss": 0.2611,
      "step": 775
    },
    {
      "epoch": 0.29,
      "grad_norm": 2.6029460430145264,
      "learning_rate": 1.6158653846153845e-05,
      "loss": 0.2907,
      "step": 800
    },
    {
      "epoch": 0.29,
      "eval_loss": 0.255000501871109,
      "eval_na_accuracy": 0.8350515365600586,
      "eval_ordinal_accuracy": 0.5617775321006775,
      "eval_ordinal_mae": 0.6431577205657959,
      "eval_runtime": 190.5444,
      "eval_samples_per_second": 23.485,
      "eval_steps_per_second": 2.939,
      "step": 800
    },
    {
      "epoch": 0.3,
      "grad_norm": 1.63986337184906,
      "learning_rate": 1.603846153846154e-05,
      "loss": 0.2661,
      "step": 825
    },
    {
      "epoch": 0.31,
      "grad_norm": 0.8340407013893127,
      "learning_rate": 1.5918269230769232e-05,
      "loss": 0.2513,
      "step": 850
    },
    {
      "epoch": 0.32,
      "grad_norm": 1.321059226989746,
      "learning_rate": 1.5798076923076924e-05,
      "loss": 0.2676,
      "step": 875
    },
    {
      "epoch": 0.32,
      "grad_norm": 2.41912841796875,
      "learning_rate": 1.567788461538462e-05,
      "loss": 0.2468,
      "step": 900
    },
    {
      "epoch": 0.32,
      "eval_loss": 0.2521895170211792,
      "eval_na_accuracy": 0.8058419227600098,
      "eval_ordinal_accuracy": 0.5972257852554321,
      "eval_ordinal_mae": 0.6118690967559814,
      "eval_runtime": 197.239,
      "eval_samples_per_second": 22.688,
      "eval_steps_per_second": 2.839,
      "step": 900
    },
    {
      "epoch": 0.33,
      "grad_norm": 0.9854594469070435,
      "learning_rate": 1.555769230769231e-05,
      "loss": 0.2727,
      "step": 925
    },
    {
      "epoch": 0.34,
      "grad_norm": 5.904122352600098,
      "learning_rate": 1.54375e-05,
      "loss": 0.3027,
      "step": 950
    },
    {
      "epoch": 0.35,
      "grad_norm": 1.8180320262908936,
      "learning_rate": 1.5317307692307694e-05,
      "loss": 0.2465,
      "step": 975
    },
    {
      "epoch": 0.36,
      "grad_norm": 0.5673078894615173,
      "learning_rate": 1.5197115384615386e-05,
      "loss": 0.2199,
      "step": 1000
    },
    {
      "epoch": 0.36,
      "eval_loss": 0.24374203383922577,
      "eval_na_accuracy": 0.8127147555351257,
      "eval_ordinal_accuracy": 0.606216311454773,
      "eval_ordinal_mae": 0.6023499965667725,
      "eval_runtime": 192.8009,
      "eval_samples_per_second": 23.21,
      "eval_steps_per_second": 2.905,
      "step": 1000
    },
    {
      "epoch": 0.37,
      "grad_norm": 0.8171074390411377,
      "learning_rate": 1.5076923076923078e-05,
      "loss": 0.221,
      "step": 1025
    },
    {
      "epoch": 0.38,
      "grad_norm": 0.881005048751831,
      "learning_rate": 1.495673076923077e-05,
      "loss": 0.235,
      "step": 1050
    },
    {
      "epoch": 0.39,
      "grad_norm": 2.021958351135254,
      "learning_rate": 1.4836538461538463e-05,
      "loss": 0.2841,
      "step": 1075
    },
    {
      "epoch": 0.4,
      "grad_norm": 1.8785498142242432,
      "learning_rate": 1.4716346153846155e-05,
      "loss": 0.2219,
      "step": 1100
    },
    {
      "epoch": 0.4,
      "eval_loss": 0.23609140515327454,
      "eval_na_accuracy": 0.9037800431251526,
      "eval_ordinal_accuracy": 0.595941424369812,
      "eval_ordinal_mae": 0.5573533177375793,
      "eval_runtime": 193.7176,
      "eval_samples_per_second": 23.101,
      "eval_steps_per_second": 2.891,
      "step": 1100
    },
    {
      "epoch": 0.41,
      "grad_norm": 0.6871089935302734,
      "learning_rate": 1.4596153846153846e-05,
      "loss": 0.201,
      "step": 1125
    },
    {
      "epoch": 0.41,
      "grad_norm": 0.3621855080127716,
      "learning_rate": 1.447596153846154e-05,
      "loss": 0.2119,
      "step": 1150
    },
    {
      "epoch": 0.42,
      "grad_norm": 1.404956340789795,
      "learning_rate": 1.4355769230769232e-05,
      "loss": 0.2566,
      "step": 1175
    },
    {
      "epoch": 0.43,
      "grad_norm": 0.8949152231216431,
      "learning_rate": 1.4235576923076923e-05,
      "loss": 0.2071,
      "step": 1200
    },
    {
      "epoch": 0.43,
      "eval_loss": 0.23867186903953552,
      "eval_na_accuracy": 0.7714776396751404,
      "eval_ordinal_accuracy": 0.6175186038017273,
      "eval_ordinal_mae": 0.5438615679740906,
      "eval_runtime": 196.3391,
      "eval_samples_per_second": 22.792,
      "eval_steps_per_second": 2.852,
      "step": 1200
    },
    {
      "epoch": 0.44,
      "grad_norm": 1.7730196714401245,
      "learning_rate": 1.4115384615384617e-05,
      "loss": 0.2756,
      "step": 1225
    },
    {
      "epoch": 0.45,
      "grad_norm": 0.7472477555274963,
      "learning_rate": 1.3995192307692308e-05,
      "loss": 0.2318,
      "step": 1250
    },
    {
      "epoch": 0.46,
      "grad_norm": 2.942986249923706,
      "learning_rate": 1.3875e-05,
      "loss": 0.2568,
      "step": 1275
    },
    {
      "epoch": 0.47,
      "grad_norm": 0.5726996064186096,
      "learning_rate": 1.3754807692307695e-05,
      "loss": 0.2214,
      "step": 1300
    },
    {
      "epoch": 0.47,
      "eval_loss": 0.2340591698884964,
      "eval_na_accuracy": 0.7955326437950134,
      "eval_ordinal_accuracy": 0.6231697797775269,
      "eval_ordinal_mae": 0.5256503224372864,
      "eval_runtime": 193.9402,
      "eval_samples_per_second": 23.074,
      "eval_steps_per_second": 2.887,
      "step": 1300
    },
    {
      "epoch": 0.48,
      "grad_norm": 0.94366455078125,
      "learning_rate": 1.3634615384615385e-05,
      "loss": 0.2146,
      "step": 1325
    },
    {
      "epoch": 0.49,
      "grad_norm": 3.612720251083374,
      "learning_rate": 1.3514423076923077e-05,
      "loss": 0.1979,
      "step": 1350
    },
    {
      "epoch": 0.5,
      "grad_norm": 0.878446638584137,
      "learning_rate": 1.3394230769230769e-05,
      "loss": 0.2233,
      "step": 1375
    },
    {
      "epoch": 0.5,
      "grad_norm": 1.1531257629394531,
      "learning_rate": 1.3274038461538464e-05,
      "loss": 0.2627,
      "step": 1400
    },
    {
      "epoch": 0.5,
      "eval_loss": 0.2315448820590973,
      "eval_na_accuracy": 0.7989690899848938,
      "eval_ordinal_accuracy": 0.6123812198638916,
      "eval_ordinal_mae": 0.5152010917663574,
      "eval_runtime": 196.8075,
      "eval_samples_per_second": 22.738,
      "eval_steps_per_second": 2.845,
      "step": 1400
    },
    {
      "epoch": 0.51,
      "grad_norm": 2.2066402435302734,
      "learning_rate": 1.3153846153846156e-05,
      "loss": 0.2415,
      "step": 1425
    },
    {
      "epoch": 0.52,
      "grad_norm": 1.4936281442642212,
      "learning_rate": 1.3033653846153846e-05,
      "loss": 0.2536,
      "step": 1450
    },
    {
      "epoch": 0.53,
      "grad_norm": 2.112527847290039,
      "learning_rate": 1.291346153846154e-05,
      "loss": 0.2105,
      "step": 1475
    },
    {
      "epoch": 0.54,
      "grad_norm": 1.8110361099243164,
      "learning_rate": 1.2793269230769233e-05,
      "loss": 0.2067,
      "step": 1500
    },
    {
      "epoch": 0.54,
      "eval_loss": 0.22465108335018158,
      "eval_na_accuracy": 0.8109965920448303,
      "eval_ordinal_accuracy": 0.6396095752716064,
      "eval_ordinal_mae": 0.5025707483291626,
      "eval_runtime": 194.1538,
      "eval_samples_per_second": 23.049,
      "eval_steps_per_second": 2.884,
      "step": 1500
    },
    {
      "epoch": 0.55,
      "grad_norm": 0.8933520913124084,
      "learning_rate": 1.2673076923076924e-05,
      "loss": 0.1957,
      "step": 1525
    },
    {
      "epoch": 0.56,
      "grad_norm": 1.3568251132965088,
      "learning_rate": 1.2557692307692309e-05,
      "loss": 0.2286,
      "step": 1550
    },
    {
      "epoch": 0.57,
      "grad_norm": 2.600196123123169,
      "learning_rate": 1.24375e-05,
      "loss": 0.2292,
      "step": 1575
    },
    {
      "epoch": 0.58,
      "grad_norm": 1.2017379999160767,
      "learning_rate": 1.2317307692307694e-05,
      "loss": 0.2086,
      "step": 1600
    },
    {
      "epoch": 0.58,
      "eval_loss": 0.21920213103294373,
      "eval_na_accuracy": 0.8041236996650696,
      "eval_ordinal_accuracy": 0.6588749289512634,
      "eval_ordinal_mae": 0.49550649523735046,
      "eval_runtime": 189.1959,
      "eval_samples_per_second": 23.653,
      "eval_steps_per_second": 2.96,
      "step": 1600
    },
    {
      "epoch": 0.59,
      "grad_norm": 2.4263405799865723,
      "learning_rate": 1.2197115384615386e-05,
      "loss": 0.2384,
      "step": 1625
    },
    {
      "epoch": 0.6,
      "grad_norm": 2.1560556888580322,
      "learning_rate": 1.2076923076923078e-05,
      "loss": 0.2557,
      "step": 1650
    },
    {
      "epoch": 0.6,
      "grad_norm": 0.8151688575744629,
      "learning_rate": 1.1956730769230771e-05,
      "loss": 0.2623,
      "step": 1675
    },
    {
      "epoch": 0.61,
      "grad_norm": 2.8995933532714844,
      "learning_rate": 1.1836538461538463e-05,
      "loss": 0.1993,
      "step": 1700
    },
    {
      "epoch": 0.61,
      "eval_loss": 0.21818678081035614,
      "eval_na_accuracy": 0.8127147555351257,
      "eval_ordinal_accuracy": 0.6521962285041809,
      "eval_ordinal_mae": 0.47375088930130005,
      "eval_runtime": 196.7859,
      "eval_samples_per_second": 22.74,
      "eval_steps_per_second": 2.846,
      "step": 1700
    },
    {
      "epoch": 0.62,
      "grad_norm": 0.6673493385314941,
      "learning_rate": 1.1716346153846155e-05,
      "loss": 0.2627,
      "step": 1725
    },
    {
      "epoch": 0.63,
      "grad_norm": 4.143210411071777,
      "learning_rate": 1.1596153846153848e-05,
      "loss": 0.2294,
      "step": 1750
    },
    {
      "epoch": 0.64,
      "grad_norm": 0.9360339641571045,
      "learning_rate": 1.147596153846154e-05,
      "loss": 0.1854,
      "step": 1775
    },
    {
      "epoch": 0.65,
      "grad_norm": 4.869482040405273,
      "learning_rate": 1.1355769230769231e-05,
      "loss": 0.1962,
      "step": 1800
    },
    {
      "epoch": 0.65,
      "eval_loss": 0.22110989689826965,
      "eval_na_accuracy": 0.9140893220901489,
      "eval_ordinal_accuracy": 0.6231697797775269,
      "eval_ordinal_mae": 0.4857858419418335,
      "eval_runtime": 187.6432,
      "eval_samples_per_second": 23.848,
      "eval_steps_per_second": 2.984,
      "step": 1800
    },
    {
      "epoch": 0.66,
      "grad_norm": 5.728977203369141,
      "learning_rate": 1.1235576923076923e-05,
      "loss": 0.2294,
      "step": 1825
    },
    {
      "epoch": 0.67,
      "grad_norm": 0.5772213339805603,
      "learning_rate": 1.1115384615384617e-05,
      "loss": 0.1981,
      "step": 1850
    },
    {
      "epoch": 0.68,
      "grad_norm": 2.899949550628662,
      "learning_rate": 1.0995192307692308e-05,
      "loss": 0.1885,
      "step": 1875
    },
    {
      "epoch": 0.69,
      "grad_norm": 1.060448408126831,
      "learning_rate": 1.0875e-05,
      "loss": 0.1882,
      "step": 1900
    },
    {
      "epoch": 0.69,
      "eval_loss": 0.20448338985443115,
      "eval_na_accuracy": 0.8625429272651672,
      "eval_ordinal_accuracy": 0.6632417440414429,
      "eval_ordinal_mae": 0.4668627381324768,
      "eval_runtime": 187.7227,
      "eval_samples_per_second": 23.838,
      "eval_steps_per_second": 2.983,
      "step": 1900
    },
    {
      "epoch": 0.69,
      "grad_norm": 8.596738815307617,
      "learning_rate": 1.0754807692307693e-05,
      "loss": 0.2388,
      "step": 1925
    },
    {
      "epoch": 0.7,
      "grad_norm": 2.922056198120117,
      "learning_rate": 1.0634615384615385e-05,
      "loss": 0.2172,
      "step": 1950
    },
    {
      "epoch": 0.71,
      "grad_norm": 2.6090290546417236,
      "learning_rate": 1.0514423076923077e-05,
      "loss": 0.2432,
      "step": 1975
    },
    {
      "epoch": 0.72,
      "grad_norm": 0.9129126667976379,
      "learning_rate": 1.039423076923077e-05,
      "loss": 0.1895,
      "step": 2000
    },
    {
      "epoch": 0.72,
      "eval_loss": 0.2081986963748932,
      "eval_na_accuracy": 0.8608247637748718,
      "eval_ordinal_accuracy": 0.6316465735435486,
      "eval_ordinal_mae": 0.46963009238243103,
      "eval_runtime": 177.2649,
      "eval_samples_per_second": 25.245,
      "eval_steps_per_second": 3.159,
      "step": 2000
    },
    {
      "epoch": 0.73,
      "grad_norm": 14.650406837463379,
      "learning_rate": 1.0274038461538462e-05,
      "loss": 0.2168,
      "step": 2025
    },
    {
      "epoch": 0.74,
      "grad_norm": 2.405910015106201,
      "learning_rate": 1.0153846153846154e-05,
      "loss": 0.224,
      "step": 2050
    },
    {
      "epoch": 0.75,
      "grad_norm": 1.4129964113235474,
      "learning_rate": 1.0033653846153847e-05,
      "loss": 0.1908,
      "step": 2075
    },
    {
      "epoch": 0.76,
      "grad_norm": 2.485114812850952,
      "learning_rate": 9.913461538461539e-06,
      "loss": 0.1979,
      "step": 2100
    },
    {
      "epoch": 0.76,
      "eval_loss": 0.22696280479431152,
      "eval_na_accuracy": 0.900343656539917,
      "eval_ordinal_accuracy": 0.6372976899147034,
      "eval_ordinal_mae": 0.4791434407234192,
      "eval_runtime": 187.006,
      "eval_samples_per_second": 23.93,
      "eval_steps_per_second": 2.995,
      "step": 2100
    },
    {
      "epoch": 0.77,
      "grad_norm": 1.7261921167373657,
      "learning_rate": 9.79326923076923e-06,
      "loss": 0.2169,
      "step": 2125
    },
    {
      "epoch": 0.78,
      "grad_norm": 5.685389995574951,
      "learning_rate": 9.673076923076924e-06,
      "loss": 0.2446,
      "step": 2150
    },
    {
      "epoch": 0.78,
      "grad_norm": 1.2241212129592896,
      "learning_rate": 9.552884615384616e-06,
      "loss": 0.2202,
      "step": 2175
    },
    {
      "epoch": 0.79,
      "grad_norm": 3.1511054039001465,
      "learning_rate": 9.432692307692308e-06,
      "loss": 0.2643,
      "step": 2200
    },
    {
      "epoch": 0.79,
      "eval_loss": 0.20690996944904327,
      "eval_na_accuracy": 0.8556700944900513,
      "eval_ordinal_accuracy": 0.6414076685905457,
      "eval_ordinal_mae": 0.46626007556915283,
      "eval_runtime": 194.378,
      "eval_samples_per_second": 23.022,
      "eval_steps_per_second": 2.881,
      "step": 2200
    },
    {
      "epoch": 0.8,
      "grad_norm": 2.382234811782837,
      "learning_rate": 9.312500000000001e-06,
      "loss": 0.1782,
      "step": 2225
    },
    {
      "epoch": 0.81,
      "grad_norm": 3.646544933319092,
      "learning_rate": 9.192307692307693e-06,
      "loss": 0.1901,
      "step": 2250
    },
    {
      "epoch": 0.82,
      "grad_norm": 3.5835981369018555,
      "learning_rate": 9.072115384615385e-06,
      "loss": 0.179,
      "step": 2275
    },
    {
      "epoch": 0.83,
      "grad_norm": 0.6391886472702026,
      "learning_rate": 8.951923076923078e-06,
      "loss": 0.2279,
      "step": 2300
    },
    {
      "epoch": 0.83,
      "eval_loss": 0.2029835283756256,
      "eval_na_accuracy": 0.869415819644928,
      "eval_ordinal_accuracy": 0.654251217842102,
      "eval_ordinal_mae": 0.4581436216831207,
      "eval_runtime": 190.8034,
      "eval_samples_per_second": 23.453,
      "eval_steps_per_second": 2.935,
      "step": 2300
    },
    {
      "epoch": 0.84,
      "grad_norm": 1.8583753108978271,
      "learning_rate": 8.83173076923077e-06,
      "loss": 0.2403,
      "step": 2325
    },
    {
      "epoch": 0.85,
      "grad_norm": 7.337312698364258,
      "learning_rate": 8.711538461538463e-06,
      "loss": 0.2262,
      "step": 2350
    },
    {
      "epoch": 0.86,
      "grad_norm": 4.292835712432861,
      "learning_rate": 8.591346153846155e-06,
      "loss": 0.2321,
      "step": 2375
    },
    {
      "epoch": 0.87,
      "grad_norm": 3.3267788887023926,
      "learning_rate": 8.471153846153847e-06,
      "loss": 0.1965,
      "step": 2400
    },
    {
      "epoch": 0.87,
      "eval_loss": 0.21094879508018494,
      "eval_na_accuracy": 0.800687313079834,
      "eval_ordinal_accuracy": 0.681993305683136,
      "eval_ordinal_mae": 0.44459667801856995,
      "eval_runtime": 184.1529,
      "eval_samples_per_second": 24.3,
      "eval_steps_per_second": 3.041,
      "step": 2400
    },
    {
      "epoch": 0.87,
      "grad_norm": 2.07079815864563,
      "learning_rate": 8.35096153846154e-06,
      "loss": 0.2122,
      "step": 2425
    },
    {
      "epoch": 0.88,
      "grad_norm": 3.6883444786071777,
      "learning_rate": 8.230769230769232e-06,
      "loss": 0.1876,
      "step": 2450
    },
    {
      "epoch": 0.89,
      "grad_norm": 0.5956806540489197,
      "learning_rate": 8.110576923076923e-06,
      "loss": 0.1804,
      "step": 2475
    },
    {
      "epoch": 0.9,
      "grad_norm": 0.6182098984718323,
      "learning_rate": 7.990384615384617e-06,
      "loss": 0.1637,
      "step": 2500
    },
    {
      "epoch": 0.9,
      "eval_loss": 0.20052286982536316,
      "eval_na_accuracy": 0.8556700944900513,
      "eval_ordinal_accuracy": 0.6763421297073364,
      "eval_ordinal_mae": 0.4438597857952118,
      "eval_runtime": 183.167,
      "eval_samples_per_second": 24.431,
      "eval_steps_per_second": 3.057,
      "step": 2500
    },
    {
      "epoch": 0.91,
      "grad_norm": 2.0849900245666504,
      "learning_rate": 7.875e-06,
      "loss": 0.2287,
      "step": 2525
    },
    {
      "epoch": 0.92,
      "grad_norm": 2.9747681617736816,
      "learning_rate": 7.754807692307693e-06,
      "loss": 0.202,
      "step": 2550
    },
    {
      "epoch": 0.93,
      "grad_norm": 0.7342644333839417,
      "learning_rate": 7.634615384615385e-06,
      "loss": 0.1887,
      "step": 2575
    },
    {
      "epoch": 0.94,
      "grad_norm": 2.5988609790802,
      "learning_rate": 7.514423076923078e-06,
      "loss": 0.1705,
      "step": 2600
    },
    {
      "epoch": 0.94,
      "eval_loss": 0.19641266763210297,
      "eval_na_accuracy": 0.8539518713951111,
      "eval_ordinal_accuracy": 0.6748009324073792,
      "eval_ordinal_mae": 0.43212634325027466,
      "eval_runtime": 189.9978,
      "eval_samples_per_second": 23.553,
      "eval_steps_per_second": 2.947,
      "step": 2600
    },
    {
      "epoch": 0.95,
      "grad_norm": 3.347687244415283,
      "learning_rate": 7.39423076923077e-06,
      "loss": 0.1947,
      "step": 2625
    },
    {
      "epoch": 0.96,
      "grad_norm": 1.516992449760437,
      "learning_rate": 7.274038461538462e-06,
      "loss": 0.2137,
      "step": 2650
    },
    {
      "epoch": 0.96,
      "grad_norm": 1.79114830493927,
      "learning_rate": 7.153846153846155e-06,
      "loss": 0.2398,
      "step": 2675
    },
    {
      "epoch": 0.97,
      "grad_norm": 2.8799991607666016,
      "learning_rate": 7.033653846153847e-06,
      "loss": 0.2412,
      "step": 2700
    },
    {
      "epoch": 0.97,
      "eval_loss": 0.19578155875205994,
      "eval_na_accuracy": 0.8780068755149841,
      "eval_ordinal_accuracy": 0.6730028390884399,
      "eval_ordinal_mae": 0.4344838559627533,
      "eval_runtime": 198.8509,
      "eval_samples_per_second": 22.504,
      "eval_steps_per_second": 2.816,
      "step": 2700
    },
    {
      "epoch": 0.98,
      "grad_norm": 1.0458216667175293,
      "learning_rate": 6.913461538461539e-06,
      "loss": 0.1697,
      "step": 2725
    },
    {
      "epoch": 0.99,
      "grad_norm": 2.2299771308898926,
      "learning_rate": 6.7932692307692315e-06,
      "loss": 0.1623,
      "step": 2750
    },
    {
      "epoch": 1.0,
      "grad_norm": 1.825832724571228,
      "learning_rate": 6.673076923076923e-06,
      "loss": 0.2052,
      "step": 2775
    },
    {
      "epoch": 1.01,
      "grad_norm": 0.6158725619316101,
      "learning_rate": 6.552884615384616e-06,
      "loss": 0.1438,
      "step": 2800
    },
    {
      "epoch": 1.01,
      "eval_loss": 0.19719891250133514,
      "eval_na_accuracy": 0.8470790386199951,
      "eval_ordinal_accuracy": 0.6783971190452576,
      "eval_ordinal_mae": 0.43012040853500366,
      "eval_runtime": 190.7061,
      "eval_samples_per_second": 23.465,
      "eval_steps_per_second": 2.936,
      "step": 2800
    },
    {
      "epoch": 1.02,
      "grad_norm": 1.041905164718628,
      "learning_rate": 6.432692307692308e-06,
      "loss": 0.1584,
      "step": 2825
    },
    {
      "epoch": 1.03,
      "grad_norm": 0.6891164779663086,
      "learning_rate": 6.3125e-06,
      "loss": 0.1488,
      "step": 2850
    },
    {
      "epoch": 1.04,
      "grad_norm": 4.814165115356445,
      "learning_rate": 6.192307692307693e-06,
      "loss": 0.1565,
      "step": 2875
    },
    {
      "epoch": 1.05,
      "grad_norm": 1.0739597082138062,
      "learning_rate": 6.0721153846153844e-06,
      "loss": 0.123,
      "step": 2900
    },
    {
      "epoch": 1.05,
      "eval_loss": 0.19954617321491241,
      "eval_na_accuracy": 0.8419243693351746,
      "eval_ordinal_accuracy": 0.6753146648406982,
      "eval_ordinal_mae": 0.42309799790382385,
      "eval_runtime": 185.744,
      "eval_samples_per_second": 24.092,
      "eval_steps_per_second": 3.015,
      "step": 2900
    },
    {
      "epoch": 1.05,
      "grad_norm": 3.492755651473999,
      "learning_rate": 5.951923076923077e-06,
      "loss": 0.1547,
      "step": 2925
    },
    {
      "epoch": 1.06,
      "grad_norm": 0.9093284010887146,
      "learning_rate": 5.8317307692307704e-06,
      "loss": 0.1258,
      "step": 2950
    },
    {
      "epoch": 1.07,
      "grad_norm": 0.7456061244010925,
      "learning_rate": 5.711538461538461e-06,
      "loss": 0.1648,
      "step": 2975
    },
    {
      "epoch": 1.08,
      "grad_norm": 0.8997055888175964,
      "learning_rate": 5.591346153846155e-06,
      "loss": 0.1411,
      "step": 3000
    },
    {
      "epoch": 1.08,
      "eval_loss": 0.19463004171848297,
      "eval_na_accuracy": 0.8453608155250549,
      "eval_ordinal_accuracy": 0.6817364692687988,
      "eval_ordinal_mae": 0.42202073335647583,
      "eval_runtime": 176.571,
      "eval_samples_per_second": 25.344,
      "eval_steps_per_second": 3.172,
      "step": 3000
    },
    {
      "epoch": 1.09,
      "grad_norm": 0.5158917903900146,
      "learning_rate": 5.471153846153847e-06,
      "loss": 0.168,
      "step": 3025
    },
    {
      "epoch": 1.1,
      "grad_norm": 1.5927814245224,
      "learning_rate": 5.350961538461539e-06,
      "loss": 0.1113,
      "step": 3050
    },
    {
      "epoch": 1.11,
      "grad_norm": 0.7580274343490601,
      "learning_rate": 5.230769230769232e-06,
      "loss": 0.1489,
      "step": 3075
    },
    {
      "epoch": 1.12,
      "grad_norm": 0.6561002135276794,
      "learning_rate": 5.110576923076923e-06,
      "loss": 0.1443,
      "step": 3100
    },
    {
      "epoch": 1.12,
      "eval_loss": 0.19156721234321594,
      "eval_na_accuracy": 0.8591065406799316,
      "eval_ordinal_accuracy": 0.682763934135437,
      "eval_ordinal_mae": 0.42213648557662964,
      "eval_runtime": 180.8137,
      "eval_samples_per_second": 24.749,
      "eval_steps_per_second": 3.097,
      "step": 3100
    },
    {
      "epoch": 1.13,
      "grad_norm": 0.7529481053352356,
      "learning_rate": 4.990384615384616e-06,
      "loss": 0.1546,
      "step": 3125
    },
    {
      "epoch": 1.14,
      "grad_norm": 3.369683027267456,
      "learning_rate": 4.870192307692308e-06,
      "loss": 0.136,
      "step": 3150
    },
    {
      "epoch": 1.14,
      "grad_norm": 0.9865265488624573,
      "learning_rate": 4.75e-06,
      "loss": 0.1498,
      "step": 3175
    },
    {
      "epoch": 1.15,
      "grad_norm": 1.8516024351119995,
      "learning_rate": 4.629807692307693e-06,
      "loss": 0.208,
      "step": 3200
    },
    {
      "epoch": 1.15,
      "eval_loss": 0.19419582188129425,
      "eval_na_accuracy": 0.8676975965499878,
      "eval_ordinal_accuracy": 0.6740303039550781,
      "eval_ordinal_mae": 0.4163132607936859,
      "eval_runtime": 188.6406,
      "eval_samples_per_second": 23.722,
      "eval_steps_per_second": 2.969,
      "step": 3200
    },
    {
      "epoch": 1.16,
      "grad_norm": 1.1575658321380615,
      "learning_rate": 4.509615384615385e-06,
      "loss": 0.1316,
      "step": 3225
    },
    {
      "epoch": 1.17,
      "grad_norm": 0.9484291076660156,
      "learning_rate": 4.389423076923077e-06,
      "loss": 0.1204,
      "step": 3250
    },
    {
      "epoch": 1.18,
      "grad_norm": 2.565762519836426,
      "learning_rate": 4.26923076923077e-06,
      "loss": 0.1262,
      "step": 3275
    },
    {
      "epoch": 1.19,
      "grad_norm": 2.2757420539855957,
      "learning_rate": 4.149038461538462e-06,
      "loss": 0.1343,
      "step": 3300
    },
    {
      "epoch": 1.19,
      "eval_loss": 0.19619733095169067,
      "eval_na_accuracy": 0.8470790386199951,
      "eval_ordinal_accuracy": 0.6889288425445557,
      "eval_ordinal_mae": 0.4182307720184326,
      "eval_runtime": 187.8854,
      "eval_samples_per_second": 23.818,
      "eval_steps_per_second": 2.981,
      "step": 3300
    },
    {
      "epoch": 1.2,
      "grad_norm": 3.402385711669922,
      "learning_rate": 4.028846153846154e-06,
      "loss": 0.1397,
      "step": 3325
    },
    {
      "epoch": 1.21,
      "grad_norm": 0.881535530090332,
      "learning_rate": 3.908653846153847e-06,
      "loss": 0.1176,
      "step": 3350
    },
    {
      "epoch": 1.22,
      "grad_norm": 0.948428213596344,
      "learning_rate": 3.7884615384615388e-06,
      "loss": 0.1767,
      "step": 3375
    },
    {
      "epoch": 1.23,
      "grad_norm": 1.2142385244369507,
      "learning_rate": 3.668269230769231e-06,
      "loss": 0.1347,
      "step": 3400
    },
    {
      "epoch": 1.23,
      "eval_loss": 0.1938188225030899,
      "eval_na_accuracy": 0.8659793734550476,
      "eval_ordinal_accuracy": 0.6899563074111938,
      "eval_ordinal_mae": 0.4161255955696106,
      "eval_runtime": 185.9367,
      "eval_samples_per_second": 24.067,
      "eval_steps_per_second": 3.012,
      "step": 3400
    },
    {
      "epoch": 1.24,
      "grad_norm": 3.2418200969696045,
      "learning_rate": 3.5480769230769235e-06,
      "loss": 0.1585,
      "step": 3425
    },
    {
      "epoch": 1.24,
      "grad_norm": 1.845564842224121,
      "learning_rate": 3.4278846153846157e-06,
      "loss": 0.1499,
      "step": 3450
    },
    {
      "epoch": 1.25,
      "grad_norm": 1.7326797246932983,
      "learning_rate": 3.307692307692308e-06,
      "loss": 0.1286,
      "step": 3475
    },
    {
      "epoch": 1.26,
      "grad_norm": 4.901269912719727,
      "learning_rate": 3.1875e-06,
      "loss": 0.1076,
      "step": 3500
    },
    {
      "epoch": 1.26,
      "eval_loss": 0.19698283076286316,
      "eval_na_accuracy": 0.8470790386199951,
      "eval_ordinal_accuracy": 0.6943231225013733,
      "eval_ordinal_mae": 0.41806870698928833,
      "eval_runtime": 187.7833,
      "eval_samples_per_second": 23.831,
      "eval_steps_per_second": 2.982,
      "step": 3500
    },
    {
      "epoch": 1.27,
      "grad_norm": 1.3833277225494385,
      "learning_rate": 3.0673076923076926e-06,
      "loss": 0.1246,
      "step": 3525
    },
    {
      "epoch": 1.28,
      "grad_norm": 0.9357690215110779,
      "learning_rate": 2.947115384615385e-06,
      "loss": 0.1113,
      "step": 3550
    },
    {
      "epoch": 1.29,
      "grad_norm": 7.975602149963379,
      "learning_rate": 2.8269230769230773e-06,
      "loss": 0.1872,
      "step": 3575
    },
    {
      "epoch": 1.3,
      "grad_norm": 0.5384923815727234,
      "learning_rate": 2.7067307692307694e-06,
      "loss": 0.1248,
      "step": 3600
    },
    {
      "epoch": 1.3,
      "eval_loss": 0.19512778520584106,
      "eval_na_accuracy": 0.8470790386199951,
      "eval_ordinal_accuracy": 0.6958643794059753,
      "eval_ordinal_mae": 0.4150661826133728,
      "eval_runtime": 190.3088,
      "eval_samples_per_second": 23.514,
      "eval_steps_per_second": 2.943,
      "step": 3600
    },
    {
      "epoch": 1.31,
      "grad_norm": 0.7750712633132935,
      "learning_rate": 2.586538461538462e-06,
      "loss": 0.1249,
      "step": 3625
    },
    {
      "epoch": 1.32,
      "grad_norm": 0.7317385077476501,
      "learning_rate": 2.466346153846154e-06,
      "loss": 0.124,
      "step": 3650
    },
    {
      "epoch": 1.33,
      "grad_norm": 0.79685378074646,
      "learning_rate": 2.3461538461538463e-06,
      "loss": 0.1524,
      "step": 3675
    },
    {
      "epoch": 1.33,
      "grad_norm": 0.9845700263977051,
      "learning_rate": 2.2259615384615385e-06,
      "loss": 0.1455,
      "step": 3700
    },
    {
      "epoch": 1.33,
      "eval_loss": 0.1951962560415268,
      "eval_na_accuracy": 0.8814433217048645,
      "eval_ordinal_accuracy": 0.6850757598876953,
      "eval_ordinal_mae": 0.4146950840950012,
      "eval_runtime": 188.1776,
      "eval_samples_per_second": 23.781,
      "eval_steps_per_second": 2.976,
      "step": 3700
    },
    {
      "epoch": 1.34,
      "grad_norm": 2.780651330947876,
      "learning_rate": 2.105769230769231e-06,
      "loss": 0.174,
      "step": 3725
    },
    {
      "epoch": 1.35,
      "grad_norm": 1.0597649812698364,
      "learning_rate": 1.9855769230769232e-06,
      "loss": 0.1211,
      "step": 3750
    },
    {
      "epoch": 1.36,
      "grad_norm": 2.3485703468322754,
      "learning_rate": 1.8653846153846156e-06,
      "loss": 0.1537,
      "step": 3775
    },
    {
      "epoch": 1.37,
      "grad_norm": 6.283831596374512,
      "learning_rate": 1.7451923076923077e-06,
      "loss": 0.131,
      "step": 3800
    },
    {
      "epoch": 1.37,
      "eval_loss": 0.19528667628765106,
      "eval_na_accuracy": 0.8453608155250549,
      "eval_ordinal_accuracy": 0.6948369145393372,
      "eval_ordinal_mae": 0.4172358810901642,
      "eval_runtime": 191.9622,
      "eval_samples_per_second": 23.312,
      "eval_steps_per_second": 2.917,
      "step": 3800
    },
    {
      "epoch": 1.38,
      "grad_norm": 0.6940748691558838,
      "learning_rate": 1.6250000000000001e-06,
      "loss": 0.1321,
      "step": 3825
    },
    {
      "epoch": 1.39,
      "grad_norm": 2.970365285873413,
      "learning_rate": 1.5048076923076923e-06,
      "loss": 0.1578,
      "step": 3850
    },
    {
      "epoch": 1.4,
      "grad_norm": 3.247617483139038,
      "learning_rate": 1.3846153846153848e-06,
      "loss": 0.1484,
      "step": 3875
    },
    {
      "epoch": 1.41,
      "grad_norm": 0.7711169719696045,
      "learning_rate": 1.264423076923077e-06,
      "loss": 0.1307,
      "step": 3900
    },
    {
      "epoch": 1.41,
      "eval_loss": 0.19318200647830963,
      "eval_na_accuracy": 0.8642611503601074,
      "eval_ordinal_accuracy": 0.692781925201416,
      "eval_ordinal_mae": 0.41271111369132996,
      "eval_runtime": 192.6809,
      "eval_samples_per_second": 23.225,
      "eval_steps_per_second": 2.906,
      "step": 3900
    },
    {
      "epoch": 1.42,
      "grad_norm": 0.6948511004447937,
      "learning_rate": 1.1442307692307694e-06,
      "loss": 0.1566,
      "step": 3925
    },
    {
      "epoch": 1.42,
      "grad_norm": 0.9776670336723328,
      "learning_rate": 1.0240384615384615e-06,
      "loss": 0.153,
      "step": 3950
    },
    {
      "epoch": 1.43,
      "grad_norm": 1.5596826076507568,
      "learning_rate": 9.038461538461539e-07,
      "loss": 0.1316,
      "step": 3975
    },
    {
      "epoch": 1.44,
      "grad_norm": 0.5882940888404846,
      "learning_rate": 7.836538461538463e-07,
      "loss": 0.1198,
      "step": 4000
    },
    {
      "epoch": 1.44,
      "eval_loss": 0.19471855461597443,
      "eval_na_accuracy": 0.8573883175849915,
      "eval_ordinal_accuracy": 0.6940662860870361,
      "eval_ordinal_mae": 0.4110487997531891,
      "eval_runtime": 192.5912,
      "eval_samples_per_second": 23.236,
      "eval_steps_per_second": 2.908,
      "step": 4000
    },
    {
      "epoch": 1.45,
      "grad_norm": 1.4058364629745483,
      "learning_rate": 6.634615384615385e-07,
      "loss": 0.1154,
      "step": 4025
    },
    {
      "epoch": 1.46,
      "grad_norm": 2.0241715908050537,
      "learning_rate": 5.432692307692308e-07,
      "loss": 0.1359,
      "step": 4050
    },
    {
      "epoch": 1.47,
      "grad_norm": 2.556328773498535,
      "learning_rate": 4.2307692307692315e-07,
      "loss": 0.1374,
      "step": 4075
    },
    {
      "epoch": 1.48,
      "grad_norm": 0.89422607421875,
      "learning_rate": 3.028846153846154e-07,
      "loss": 0.1363,
      "step": 4100
    },
    {
      "epoch": 1.48,
      "eval_loss": 0.19520333409309387,
      "eval_na_accuracy": 0.8573883175849915,
      "eval_ordinal_accuracy": 0.6886719465255737,
      "eval_ordinal_mae": 0.4086832106113434,
      "eval_runtime": 191.6181,
      "eval_samples_per_second": 23.354,
      "eval_steps_per_second": 2.922,
      "step": 4100
    },
    {
      "epoch": 1.49,
      "grad_norm": 1.4230666160583496,
      "learning_rate": 1.8269230769230772e-07,
      "loss": 0.1273,
      "step": 4125
    },
    {
      "epoch": 1.5,
      "grad_norm": 2.7363412380218506,
      "learning_rate": 6.250000000000001e-08,
      "loss": 0.1068,
      "step": 4150
    },
    {
      "epoch": 1.5,
      "step": 4160,
      "total_flos": 5.158051742063002e+18,
      "train_loss": 0.21532289660893955,
      "train_runtime": 13643.9574,
      "train_samples_per_second": 4.878,
      "train_steps_per_second": 0.305
    }
  ],
  "logging_steps": 25,
  "max_steps": 4160,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 2,
  "save_steps": 100,
  "total_flos": 5.158051742063002e+18,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": null
}