{
  "best_metric": 0.1198035180568695,
  "best_model_checkpoint": "DinoVdeauTest-large-2024_09_24-batch-size32_freeze/checkpoint-26208",
  "epoch": 106.0,
  "eval_steps": 500,
  "global_step": 28938,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "eval_accuracy": 0.2169092169092169,
      "eval_f1_macro": 0.5520222156367808,
      "eval_f1_micro": 0.7485599799649386,
      "eval_loss": 0.17755262553691864,
      "eval_runtime": 583.8306,
      "eval_samples_per_second": 4.943,
      "eval_steps_per_second": 0.156,
      "learning_rate": 0.001,
      "step": 273
    },
    {
      "epoch": 1.8315018315018317,
      "grad_norm": 0.2782972455024719,
      "learning_rate": 0.001,
      "loss": 0.2736,
      "step": 500
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.24047124047124047,
      "eval_f1_macro": 0.5698350844395493,
      "eval_f1_micro": 0.7693301744561706,
      "eval_loss": 0.15282750129699707,
      "eval_runtime": 589.1224,
      "eval_samples_per_second": 4.899,
      "eval_steps_per_second": 0.154,
      "learning_rate": 0.001,
      "step": 546
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.2494802494802495,
      "eval_f1_macro": 0.6216984689952817,
      "eval_f1_micro": 0.7773639211038686,
      "eval_loss": 0.14834792912006378,
      "eval_runtime": 584.7179,
      "eval_samples_per_second": 4.936,
      "eval_steps_per_second": 0.156,
      "learning_rate": 0.001,
      "step": 819
    },
    {
      "epoch": 3.663003663003663,
      "grad_norm": 0.260547012090683,
      "learning_rate": 0.001,
      "loss": 0.1699,
      "step": 1000
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.2553707553707554,
      "eval_f1_macro": 0.6272242063327677,
      "eval_f1_micro": 0.777158535539741,
      "eval_loss": 0.1467081755399704,
      "eval_runtime": 581.5001,
      "eval_samples_per_second": 4.963,
      "eval_steps_per_second": 0.156,
      "learning_rate": 0.001,
      "step": 1092
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.24636174636174638,
      "eval_f1_macro": 0.6281234234815563,
      "eval_f1_micro": 0.7772219080349763,
      "eval_loss": 0.1453460305929184,
      "eval_runtime": 586.5696,
      "eval_samples_per_second": 4.92,
      "eval_steps_per_second": 0.155,
      "learning_rate": 0.001,
      "step": 1365
    },
    {
      "epoch": 5.4945054945054945,
      "grad_norm": 0.18768520653247833,
      "learning_rate": 0.001,
      "loss": 0.1622,
      "step": 1500
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.262993762993763,
      "eval_f1_macro": 0.6191104313765434,
      "eval_f1_micro": 0.7809704878364948,
      "eval_loss": 0.14370940625667572,
      "eval_runtime": 588.2849,
      "eval_samples_per_second": 4.906,
      "eval_steps_per_second": 0.155,
      "learning_rate": 0.001,
      "step": 1638
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.2598752598752599,
      "eval_f1_macro": 0.6172423187752968,
      "eval_f1_micro": 0.7811054928907905,
      "eval_loss": 0.14275749027729034,
      "eval_runtime": 590.2652,
      "eval_samples_per_second": 4.889,
      "eval_steps_per_second": 0.154,
      "learning_rate": 0.001,
      "step": 1911
    },
    {
      "epoch": 7.326007326007326,
      "grad_norm": 0.21249784529209137,
      "learning_rate": 0.001,
      "loss": 0.1593,
      "step": 2000
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.25571725571725573,
      "eval_f1_macro": 0.6252572102986063,
      "eval_f1_micro": 0.7799277605779154,
      "eval_loss": 0.1432911604642868,
      "eval_runtime": 589.0928,
      "eval_samples_per_second": 4.899,
      "eval_steps_per_second": 0.154,
      "learning_rate": 0.001,
      "step": 2184
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.25190575190575193,
      "eval_f1_macro": 0.6529741984810739,
      "eval_f1_micro": 0.7892645815722739,
      "eval_loss": 0.14202255010604858,
      "eval_runtime": 586.5195,
      "eval_samples_per_second": 4.921,
      "eval_steps_per_second": 0.155,
      "learning_rate": 0.001,
      "step": 2457
    },
    {
      "epoch": 9.157509157509157,
      "grad_norm": 0.19528812170028687,
      "learning_rate": 0.001,
      "loss": 0.1569,
      "step": 2500
    },
    {
      "epoch": 10.0,
      "eval_accuracy": 0.2553707553707554,
      "eval_f1_macro": 0.6021759252294137,
      "eval_f1_micro": 0.7441173839133935,
      "eval_loss": 0.24619852006435394,
      "eval_runtime": 594.8068,
      "eval_samples_per_second": 4.852,
      "eval_steps_per_second": 0.153,
      "learning_rate": 0.001,
      "step": 2730
    },
    {
      "epoch": 10.989010989010989,
      "grad_norm": 0.16383688151836395,
      "learning_rate": 0.001,
      "loss": 0.156,
      "step": 3000
    },
    {
      "epoch": 11.0,
      "eval_accuracy": 0.2702702702702703,
      "eval_f1_macro": 0.6483450509947135,
      "eval_f1_micro": 0.7882727742379886,
      "eval_loss": 0.14001792669296265,
      "eval_runtime": 585.5736,
      "eval_samples_per_second": 4.929,
      "eval_steps_per_second": 0.155,
      "learning_rate": 0.001,
      "step": 3003
    },
    {
      "epoch": 12.0,
      "eval_accuracy": 0.2598752598752599,
      "eval_f1_macro": 0.6590290067827226,
      "eval_f1_micro": 0.7905884326636748,
      "eval_loss": 0.13999028503894806,
      "eval_runtime": 594.5312,
      "eval_samples_per_second": 4.854,
      "eval_steps_per_second": 0.153,
      "learning_rate": 0.001,
      "step": 3276
    },
    {
      "epoch": 12.820512820512821,
      "grad_norm": 0.164137601852417,
      "learning_rate": 0.001,
      "loss": 0.1547,
      "step": 3500
    },
    {
      "epoch": 13.0,
      "eval_accuracy": 0.2643797643797644,
      "eval_f1_macro": 0.6466696619852873,
      "eval_f1_micro": 0.7875647668393783,
      "eval_loss": 0.13937941193580627,
      "eval_runtime": 616.229,
      "eval_samples_per_second": 4.683,
      "eval_steps_per_second": 0.148,
      "learning_rate": 0.001,
      "step": 3549
    },
    {
      "epoch": 14.0,
      "eval_accuracy": 0.26403326403326405,
      "eval_f1_macro": 0.6469191780902835,
      "eval_f1_micro": 0.7879462374796679,
      "eval_loss": 0.13986562192440033,
      "eval_runtime": 604.9775,
      "eval_samples_per_second": 4.77,
      "eval_steps_per_second": 0.15,
      "learning_rate": 0.001,
      "step": 3822
    },
    {
      "epoch": 14.652014652014651,
      "grad_norm": 0.17560669779777527,
      "learning_rate": 0.001,
      "loss": 0.1543,
      "step": 4000
    },
    {
      "epoch": 15.0,
      "eval_accuracy": 0.2512127512127512,
      "eval_f1_macro": 0.6413408300553888,
      "eval_f1_micro": 0.7881215116526777,
      "eval_loss": 0.13914281129837036,
      "eval_runtime": 609.152,
      "eval_samples_per_second": 4.738,
      "eval_steps_per_second": 0.149,
      "learning_rate": 0.001,
      "step": 4095
    },
    {
      "epoch": 16.0,
      "eval_accuracy": 0.26334026334026334,
      "eval_f1_macro": 0.6347762053394217,
      "eval_f1_micro": 0.7906518010291596,
      "eval_loss": 0.14137022197246552,
      "eval_runtime": 599.9977,
      "eval_samples_per_second": 4.81,
      "eval_steps_per_second": 0.152,
      "learning_rate": 0.001,
      "step": 4368
    },
    {
      "epoch": 16.483516483516482,
      "grad_norm": 0.16528823971748352,
      "learning_rate": 0.001,
      "loss": 0.1536,
      "step": 4500
    },
    {
      "epoch": 17.0,
      "eval_accuracy": 0.2616077616077616,
      "eval_f1_macro": 0.6445091939759507,
      "eval_f1_micro": 0.7903470350404312,
      "eval_loss": 0.14031356573104858,
      "eval_runtime": 579.4575,
      "eval_samples_per_second": 4.981,
      "eval_steps_per_second": 0.157,
      "learning_rate": 0.001,
      "step": 4641
    },
    {
      "epoch": 18.0,
      "eval_accuracy": 0.26022176022176025,
      "eval_f1_macro": 0.6463740590949281,
      "eval_f1_micro": 0.792909822326517,
      "eval_loss": 0.14039942622184753,
      "eval_runtime": 586.9618,
      "eval_samples_per_second": 4.917,
      "eval_steps_per_second": 0.155,
      "learning_rate": 0.001,
      "step": 4914
    },
    {
      "epoch": 18.315018315018314,
      "grad_norm": 0.17777948081493378,
      "learning_rate": 0.001,
      "loss": 0.1556,
      "step": 5000
    },
    {
      "epoch": 19.0,
      "eval_accuracy": 0.25848925848925847,
      "eval_f1_macro": 0.6526097604635809,
      "eval_f1_micro": 0.7935513900055008,
      "eval_loss": 0.14035724103450775,
      "eval_runtime": 574.4193,
      "eval_samples_per_second": 5.024,
      "eval_steps_per_second": 0.158,
      "learning_rate": 0.001,
      "step": 5187
    },
    {
      "epoch": 20.0,
      "eval_accuracy": 0.255024255024255,
      "eval_f1_macro": 0.6491113591903472,
      "eval_f1_micro": 0.7900462566386842,
      "eval_loss": 0.1390993297100067,
      "eval_runtime": 576.4077,
      "eval_samples_per_second": 5.007,
      "eval_steps_per_second": 0.158,
      "learning_rate": 0.001,
      "step": 5460
    },
    {
      "epoch": 20.146520146520146,
      "grad_norm": 0.14455388486385345,
      "learning_rate": 0.001,
      "loss": 0.1534,
      "step": 5500
    },
    {
      "epoch": 21.0,
      "eval_accuracy": 0.25814275814275817,
      "eval_f1_macro": 0.6507882842254223,
      "eval_f1_micro": 0.7916862276471341,
      "eval_loss": 0.13827534019947052,
      "eval_runtime": 598.284,
      "eval_samples_per_second": 4.824,
      "eval_steps_per_second": 0.152,
      "learning_rate": 0.001,
      "step": 5733
    },
    {
      "epoch": 21.978021978021978,
      "grad_norm": 0.1771659553050995,
      "learning_rate": 0.001,
      "loss": 0.1533,
      "step": 6000
    },
    {
      "epoch": 22.0,
      "eval_accuracy": 0.26853776853776856,
      "eval_f1_macro": 0.6489112697391156,
      "eval_f1_micro": 0.7935178133197935,
      "eval_loss": 0.13889800012111664,
      "eval_runtime": 594.9509,
      "eval_samples_per_second": 4.851,
      "eval_steps_per_second": 0.153,
      "learning_rate": 0.001,
      "step": 6006
    },
    {
      "epoch": 23.0,
      "eval_accuracy": 0.2560637560637561,
      "eval_f1_macro": 0.6549504216490796,
      "eval_f1_micro": 0.7842915050342152,
      "eval_loss": 0.1385144591331482,
      "eval_runtime": 583.6881,
      "eval_samples_per_second": 4.944,
      "eval_steps_per_second": 0.156,
      "learning_rate": 0.001,
      "step": 6279
    },
    {
      "epoch": 23.80952380952381,
      "grad_norm": 0.17443060874938965,
      "learning_rate": 0.001,
      "loss": 0.1531,
      "step": 6500
    },
    {
      "epoch": 24.0,
      "eval_accuracy": 0.2695772695772696,
      "eval_f1_macro": 0.6507209930045859,
      "eval_f1_micro": 0.7921174652241113,
      "eval_loss": 0.13674499094486237,
      "eval_runtime": 574.2321,
      "eval_samples_per_second": 5.026,
      "eval_steps_per_second": 0.158,
      "learning_rate": 0.001,
      "step": 6552
    },
    {
      "epoch": 25.0,
      "eval_accuracy": 0.2751212751212751,
      "eval_f1_macro": 0.6408412361180978,
      "eval_f1_micro": 0.7893792922444337,
      "eval_loss": 0.13791973888874054,
      "eval_runtime": 577.6745,
      "eval_samples_per_second": 4.996,
      "eval_steps_per_second": 0.158,
      "learning_rate": 0.001,
      "step": 6825
    },
    {
      "epoch": 25.641025641025642,
      "grad_norm": 0.16064241528511047,
      "learning_rate": 0.001,
      "loss": 0.1533,
      "step": 7000
    },
    {
      "epoch": 26.0,
      "eval_accuracy": 0.27096327096327094,
      "eval_f1_macro": 0.6468889221342444,
      "eval_f1_micro": 0.7942844956280657,
      "eval_loss": 0.1375003606081009,
      "eval_runtime": 576.2047,
      "eval_samples_per_second": 5.009,
      "eval_steps_per_second": 0.158,
      "learning_rate": 0.001,
      "step": 7098
    },
    {
      "epoch": 27.0,
      "eval_accuracy": 0.2643797643797644,
      "eval_f1_macro": 0.6516252219978993,
      "eval_f1_micro": 0.7921208479958802,
      "eval_loss": 0.13920167088508606,
      "eval_runtime": 574.3969,
      "eval_samples_per_second": 5.024,
      "eval_steps_per_second": 0.158,
      "learning_rate": 0.001,
      "step": 7371
    },
    {
      "epoch": 27.47252747252747,
      "grad_norm": 0.15187925100326538,
      "learning_rate": 0.001,
      "loss": 0.1529,
      "step": 7500
    },
    {
      "epoch": 28.0,
      "eval_accuracy": 0.2636867636867637,
      "eval_f1_macro": 0.6424574000169906,
      "eval_f1_micro": 0.7917504599717599,
      "eval_loss": 0.13846370577812195,
      "eval_runtime": 583.6888,
      "eval_samples_per_second": 4.944,
      "eval_steps_per_second": 0.156,
      "learning_rate": 0.001,
      "step": 7644
    },
    {
      "epoch": 29.0,
      "eval_accuracy": 0.2626472626472626,
      "eval_f1_macro": 0.6496760555529415,
      "eval_f1_micro": 0.7883173722754258,
      "eval_loss": 0.14017289876937866,
      "eval_runtime": 569.4194,
      "eval_samples_per_second": 5.068,
      "eval_steps_per_second": 0.16,
      "learning_rate": 0.001,
      "step": 7917
    },
    {
      "epoch": 29.304029304029303,
      "grad_norm": 0.1429462879896164,
      "learning_rate": 0.001,
      "loss": 0.153,
      "step": 8000
    },
    {
      "epoch": 30.0,
      "eval_accuracy": 0.2668052668052668,
      "eval_f1_macro": 0.6553418932728605,
      "eval_f1_micro": 0.788684508761224,
      "eval_loss": 0.13773277401924133,
      "eval_runtime": 575.6917,
      "eval_samples_per_second": 5.013,
      "eval_steps_per_second": 0.158,
      "learning_rate": 0.001,
      "step": 8190
    },
    {
      "epoch": 31.0,
      "eval_accuracy": 0.27893277893277896,
      "eval_f1_macro": 0.6718196714045327,
      "eval_f1_micro": 0.8017825006427286,
      "eval_loss": 0.1312580555677414,
      "eval_runtime": 580.0807,
      "eval_samples_per_second": 4.975,
      "eval_steps_per_second": 0.157,
      "learning_rate": 0.0001,
      "step": 8463
    },
    {
      "epoch": 31.135531135531135,
      "grad_norm": 0.1542915403842926,
      "learning_rate": 0.0001,
      "loss": 0.1486,
      "step": 8500
    },
    {
      "epoch": 32.0,
      "eval_accuracy": 0.2806652806652807,
      "eval_f1_macro": 0.6772473958552189,
      "eval_f1_micro": 0.8060657118786858,
      "eval_loss": 0.1318267583847046,
      "eval_runtime": 567.7754,
      "eval_samples_per_second": 5.083,
      "eval_steps_per_second": 0.16,
      "learning_rate": 0.0001,
      "step": 8736
    },
    {
      "epoch": 32.967032967032964,
      "grad_norm": 0.1691681146621704,
      "learning_rate": 0.0001,
      "loss": 0.1415,
      "step": 9000
    },
    {
      "epoch": 33.0,
      "eval_accuracy": 0.27754677754677753,
      "eval_f1_macro": 0.6791810095569879,
      "eval_f1_micro": 0.8050148746281343,
      "eval_loss": 0.1309264600276947,
      "eval_runtime": 573.2692,
      "eval_samples_per_second": 5.034,
      "eval_steps_per_second": 0.159,
      "learning_rate": 0.0001,
      "step": 9009
    },
    {
      "epoch": 34.0,
      "eval_accuracy": 0.28205128205128205,
      "eval_f1_macro": 0.6774879025186358,
      "eval_f1_micro": 0.804873906767428,
      "eval_loss": 0.1296369731426239,
      "eval_runtime": 573.8273,
      "eval_samples_per_second": 5.029,
      "eval_steps_per_second": 0.159,
      "learning_rate": 0.0001,
      "step": 9282
    },
    {
      "epoch": 34.798534798534796,
      "grad_norm": 0.16409221291542053,
      "learning_rate": 0.0001,
      "loss": 0.1395,
      "step": 9500
    },
    {
      "epoch": 35.0,
      "eval_accuracy": 0.28932778932778935,
      "eval_f1_macro": 0.6864816840760736,
      "eval_f1_micro": 0.8084945013272659,
      "eval_loss": 0.12816031277179718,
      "eval_runtime": 577.8146,
      "eval_samples_per_second": 4.995,
      "eval_steps_per_second": 0.157,
      "learning_rate": 0.0001,
      "step": 9555
    },
    {
      "epoch": 36.0,
      "eval_accuracy": 0.2830907830907831,
      "eval_f1_macro": 0.6827519815644174,
      "eval_f1_micro": 0.8055151151836945,
      "eval_loss": 0.12886497378349304,
      "eval_runtime": 576.0485,
      "eval_samples_per_second": 5.01,
      "eval_steps_per_second": 0.158,
      "learning_rate": 0.0001,
      "step": 9828
    },
    {
      "epoch": 36.63003663003663,
      "grad_norm": 0.1629299372434616,
      "learning_rate": 0.0001,
      "loss": 0.1387,
      "step": 10000
    },
    {
      "epoch": 37.0,
      "eval_accuracy": 0.2830907830907831,
      "eval_f1_macro": 0.677476009100308,
      "eval_f1_micro": 0.8054549328787556,
      "eval_loss": 0.12768980860710144,
      "eval_runtime": 578.6622,
      "eval_samples_per_second": 4.987,
      "eval_steps_per_second": 0.157,
      "learning_rate": 0.0001,
      "step": 10101
    },
    {
      "epoch": 38.0,
      "eval_accuracy": 0.288981288981289,
      "eval_f1_macro": 0.6882635597933275,
      "eval_f1_micro": 0.8084112149532711,
      "eval_loss": 0.1274958997964859,
      "eval_runtime": 583.3399,
      "eval_samples_per_second": 4.947,
      "eval_steps_per_second": 0.156,
      "learning_rate": 0.0001,
      "step": 10374
    },
    {
      "epoch": 38.46153846153846,
      "grad_norm": 0.20412367582321167,
      "learning_rate": 0.0001,
      "loss": 0.1354,
      "step": 10500
    },
    {
      "epoch": 39.0,
      "eval_accuracy": 0.28794178794178793,
      "eval_f1_macro": 0.6853651532137502,
      "eval_f1_micro": 0.8099270562589084,
      "eval_loss": 0.12658803164958954,
      "eval_runtime": 588.9538,
      "eval_samples_per_second": 4.9,
      "eval_steps_per_second": 0.155,
      "learning_rate": 0.0001,
      "step": 10647
    },
    {
      "epoch": 40.0,
      "eval_accuracy": 0.28863478863478864,
      "eval_f1_macro": 0.698146081234717,
      "eval_f1_micro": 0.8117343111925994,
      "eval_loss": 0.12820227444171906,
      "eval_runtime": 583.5775,
      "eval_samples_per_second": 4.945,
      "eval_steps_per_second": 0.156,
      "learning_rate": 0.0001,
      "step": 10920
    },
    {
      "epoch": 40.29304029304029,
      "grad_norm": 0.1920085847377777,
      "learning_rate": 0.0001,
      "loss": 0.1355,
      "step": 11000
    },
    {
      "epoch": 41.0,
      "eval_accuracy": 0.2882882882882883,
      "eval_f1_macro": 0.6850722645729148,
      "eval_f1_micro": 0.8081688455951466,
      "eval_loss": 0.12668322026729584,
      "eval_runtime": 580.9481,
      "eval_samples_per_second": 4.968,
      "eval_steps_per_second": 0.157,
      "learning_rate": 0.0001,
      "step": 11193
    },
    {
      "epoch": 42.0,
      "eval_accuracy": 0.2907137907137907,
      "eval_f1_macro": 0.6941693850580287,
      "eval_f1_micro": 0.8112331081081081,
      "eval_loss": 0.12620903551578522,
      "eval_runtime": 587.956,
      "eval_samples_per_second": 4.909,
      "eval_steps_per_second": 0.155,
      "learning_rate": 0.0001,
      "step": 11466
    },
    {
      "epoch": 42.124542124542124,
      "grad_norm": 0.20102928578853607,
      "learning_rate": 0.0001,
      "loss": 0.1347,
      "step": 11500
    },
    {
      "epoch": 43.0,
      "eval_accuracy": 0.2910602910602911,
      "eval_f1_macro": 0.69077693559172,
      "eval_f1_micro": 0.8107360033550011,
      "eval_loss": 0.12590545415878296,
      "eval_runtime": 582.6431,
      "eval_samples_per_second": 4.953,
      "eval_steps_per_second": 0.156,
      "learning_rate": 0.0001,
      "step": 11739
    },
    {
      "epoch": 43.956043956043956,
      "grad_norm": 0.20798929035663605,
      "learning_rate": 0.0001,
      "loss": 0.1337,
      "step": 12000
    },
    {
      "epoch": 44.0,
      "eval_accuracy": 0.29313929313929316,
      "eval_f1_macro": 0.6925170228864211,
      "eval_f1_micro": 0.8114530416894076,
      "eval_loss": 0.12642185389995575,
      "eval_runtime": 580.9769,
      "eval_samples_per_second": 4.967,
      "eval_steps_per_second": 0.157,
      "learning_rate": 0.0001,
      "step": 12012
    },
    {
      "epoch": 45.0,
      "eval_accuracy": 0.2966042966042966,
      "eval_f1_macro": 0.6974852591826903,
      "eval_f1_micro": 0.8110017663386323,
      "eval_loss": 0.1257808804512024,
      "eval_runtime": 587.2129,
      "eval_samples_per_second": 4.915,
      "eval_steps_per_second": 0.155,
      "learning_rate": 0.0001,
      "step": 12285
    },
    {
      "epoch": 45.78754578754579,
      "grad_norm": 0.21279653906822205,
      "learning_rate": 0.0001,
      "loss": 0.1329,
      "step": 12500
    },
    {
      "epoch": 46.0,
      "eval_accuracy": 0.2986832986832987,
      "eval_f1_macro": 0.694112624204446,
      "eval_f1_micro": 0.8109046268467172,
      "eval_loss": 0.12542255222797394,
      "eval_runtime": 584.2715,
      "eval_samples_per_second": 4.939,
      "eval_steps_per_second": 0.156,
      "learning_rate": 0.0001,
      "step": 12558
    },
    {
      "epoch": 47.0,
      "eval_accuracy": 0.2920997920997921,
      "eval_f1_macro": 0.6937033329875891,
      "eval_f1_micro": 0.8098117995347853,
      "eval_loss": 0.1256514936685562,
      "eval_runtime": 582.8653,
      "eval_samples_per_second": 4.951,
      "eval_steps_per_second": 0.156,
      "learning_rate": 0.0001,
      "step": 12831
    },
    {
      "epoch": 47.61904761904762,
      "grad_norm": 0.19104164838790894,
      "learning_rate": 0.0001,
      "loss": 0.1331,
      "step": 13000
    },
    {
      "epoch": 48.0,
      "eval_accuracy": 0.29140679140679143,
      "eval_f1_macro": 0.6904792900172609,
      "eval_f1_micro": 0.8106740862526874,
      "eval_loss": 0.1253652125597,
      "eval_runtime": 590.5724,
      "eval_samples_per_second": 4.887,
      "eval_steps_per_second": 0.154,
      "learning_rate": 0.0001,
      "step": 13104
    },
    {
      "epoch": 49.0,
      "eval_accuracy": 0.2945252945252945,
      "eval_f1_macro": 0.6973960572840976,
      "eval_f1_micro": 0.8136526746223817,
      "eval_loss": 0.1252448409795761,
      "eval_runtime": 596.869,
      "eval_samples_per_second": 4.835,
      "eval_steps_per_second": 0.152,
      "learning_rate": 0.0001,
      "step": 13377
    },
    {
      "epoch": 49.45054945054945,
      "grad_norm": 0.25627467036247253,
      "learning_rate": 0.0001,
      "loss": 0.1309,
      "step": 13500
    },
    {
      "epoch": 50.0,
      "eval_accuracy": 0.2983367983367983,
      "eval_f1_macro": 0.7025806374316313,
      "eval_f1_micro": 0.814972104255142,
      "eval_loss": 0.12482810020446777,
      "eval_runtime": 598.3632,
      "eval_samples_per_second": 4.823,
      "eval_steps_per_second": 0.152,
      "learning_rate": 0.0001,
      "step": 13650
    },
    {
      "epoch": 51.0,
      "eval_accuracy": 0.2959112959112959,
      "eval_f1_macro": 0.7066639116616246,
      "eval_f1_micro": 0.815773054365459,
      "eval_loss": 0.12458823621273041,
      "eval_runtime": 578.8177,
      "eval_samples_per_second": 4.986,
      "eval_steps_per_second": 0.157,
      "learning_rate": 0.0001,
      "step": 13923
    },
    {
      "epoch": 51.282051282051285,
      "grad_norm": 0.2170032560825348,
      "learning_rate": 0.0001,
      "loss": 0.1304,
      "step": 14000
    },
    {
      "epoch": 52.0,
      "eval_accuracy": 0.29521829521829523,
      "eval_f1_macro": 0.7008785276837475,
      "eval_f1_micro": 0.8120651620566774,
      "eval_loss": 0.12461517751216888,
      "eval_runtime": 581.8661,
      "eval_samples_per_second": 4.96,
      "eval_steps_per_second": 0.156,
      "learning_rate": 0.0001,
      "step": 14196
    },
    {
      "epoch": 53.0,
      "eval_accuracy": 0.29902979902979904,
      "eval_f1_macro": 0.697381421827107,
      "eval_f1_micro": 0.8142762173840784,
      "eval_loss": 0.12417320907115936,
      "eval_runtime": 587.3587,
      "eval_samples_per_second": 4.914,
      "eval_steps_per_second": 0.155,
      "learning_rate": 0.0001,
      "step": 14469
    },
    {
      "epoch": 53.11355311355312,
      "grad_norm": 0.18614046275615692,
      "learning_rate": 0.0001,
      "loss": 0.1309,
      "step": 14500
    },
    {
      "epoch": 54.0,
      "eval_accuracy": 0.2966042966042966,
      "eval_f1_macro": 0.7000554609843794,
      "eval_f1_micro": 0.8135221716405328,
      "eval_loss": 0.12411519885063171,
      "eval_runtime": 590.4943,
      "eval_samples_per_second": 4.887,
      "eval_steps_per_second": 0.154,
      "learning_rate": 0.0001,
      "step": 14742
    },
    {
      "epoch": 54.94505494505494,
      "grad_norm": 0.296142578125,
      "learning_rate": 0.0001,
      "loss": 0.1289,
      "step": 15000
    },
    {
      "epoch": 55.0,
      "eval_accuracy": 0.29521829521829523,
      "eval_f1_macro": 0.6996806217177095,
      "eval_f1_micro": 0.8130868668154448,
      "eval_loss": 0.12417341023683548,
      "eval_runtime": 580.598,
      "eval_samples_per_second": 4.971,
      "eval_steps_per_second": 0.157,
      "learning_rate": 0.0001,
      "step": 15015
    },
    {
      "epoch": 56.0,
      "eval_accuracy": 0.30214830214830213,
      "eval_f1_macro": 0.7064033143378504,
      "eval_f1_micro": 0.8178580303155802,
      "eval_loss": 0.1234845370054245,
      "eval_runtime": 578.4648,
      "eval_samples_per_second": 4.989,
      "eval_steps_per_second": 0.157,
      "learning_rate": 0.0001,
      "step": 15288
    },
    {
      "epoch": 56.776556776556774,
      "grad_norm": 0.23026500642299652,
      "learning_rate": 0.0001,
      "loss": 0.1286,
      "step": 15500
    },
    {
      "epoch": 57.0,
      "eval_accuracy": 0.2993762993762994,
      "eval_f1_macro": 0.6963082116005005,
      "eval_f1_micro": 0.8150138550675959,
      "eval_loss": 0.12350637465715408,
      "eval_runtime": 578.6503,
      "eval_samples_per_second": 4.987,
      "eval_steps_per_second": 0.157,
      "learning_rate": 0.0001,
      "step": 15561
    },
    {
      "epoch": 58.0,
      "eval_accuracy": 0.2983367983367983,
      "eval_f1_macro": 0.7011763866264856,
      "eval_f1_micro": 0.814454598809424,
      "eval_loss": 0.12311282008886337,
      "eval_runtime": 575.7868,
      "eval_samples_per_second": 5.012,
      "eval_steps_per_second": 0.158,
      "learning_rate": 0.0001,
      "step": 15834
    },
    {
      "epoch": 58.608058608058606,
      "grad_norm": 0.2271023392677307,
      "learning_rate": 0.0001,
      "loss": 0.1282,
      "step": 16000
    },
    {
      "epoch": 59.0,
      "eval_accuracy": 0.30006930006930005,
      "eval_f1_macro": 0.7021882450681799,
      "eval_f1_micro": 0.8153207895067678,
      "eval_loss": 0.12340909987688065,
      "eval_runtime": 574.6752,
      "eval_samples_per_second": 5.022,
      "eval_steps_per_second": 0.158,
      "learning_rate": 0.0001,
      "step": 16107
    },
    {
      "epoch": 60.0,
      "eval_accuracy": 0.2972972972972973,
      "eval_f1_macro": 0.6978379224258093,
      "eval_f1_micro": 0.8121788610981358,
      "eval_loss": 0.12386388331651688,
      "eval_runtime": 580.8249,
      "eval_samples_per_second": 4.969,
      "eval_steps_per_second": 0.157,
      "learning_rate": 0.0001,
      "step": 16380
    },
    {
      "epoch": 60.43956043956044,
      "grad_norm": 0.26319652795791626,
      "learning_rate": 0.0001,
      "loss": 0.1282,
      "step": 16500
    },
    {
      "epoch": 61.0,
      "eval_accuracy": 0.30145530145530147,
      "eval_f1_macro": 0.7114090358686158,
      "eval_f1_micro": 0.8157532819337362,
      "eval_loss": 0.12355918437242508,
      "eval_runtime": 580.0589,
      "eval_samples_per_second": 4.975,
      "eval_steps_per_second": 0.157,
      "learning_rate": 0.0001,
      "step": 16653
    },
    {
      "epoch": 62.0,
      "eval_accuracy": 0.3031878031878032,
      "eval_f1_macro": 0.7119609165997448,
      "eval_f1_micro": 0.8168304358780549,
      "eval_loss": 0.12266429513692856,
      "eval_runtime": 571.7568,
      "eval_samples_per_second": 5.048,
      "eval_steps_per_second": 0.159,
      "learning_rate": 0.0001,
      "step": 16926
    },
    {
      "epoch": 62.27106227106227,
      "grad_norm": 0.2723998725414276,
      "learning_rate": 0.0001,
      "loss": 0.1265,
      "step": 17000
    },
    {
      "epoch": 63.0,
      "eval_accuracy": 0.2948717948717949,
      "eval_f1_macro": 0.7077249537938423,
      "eval_f1_micro": 0.8136965505608501,
      "eval_loss": 0.12305888533592224,
      "eval_runtime": 579.5124,
      "eval_samples_per_second": 4.98,
      "eval_steps_per_second": 0.157,
      "learning_rate": 0.0001,
      "step": 17199
    },
    {
      "epoch": 64.0,
      "eval_accuracy": 0.30561330561330563,
      "eval_f1_macro": 0.708440477929098,
      "eval_f1_micro": 0.8172187094342783,
      "eval_loss": 0.12276986986398697,
      "eval_runtime": 580.0855,
      "eval_samples_per_second": 4.975,
      "eval_steps_per_second": 0.157,
      "learning_rate": 0.0001,
      "step": 17472
    },
    {
      "epoch": 64.1025641025641,
      "grad_norm": 0.2615499794483185,
      "learning_rate": 0.0001,
      "loss": 0.1273,
      "step": 17500
    },
    {
      "epoch": 65.0,
      "eval_accuracy": 0.3076923076923077,
      "eval_f1_macro": 0.7103116175502073,
      "eval_f1_micro": 0.8182651445712857,
      "eval_loss": 0.12323789298534393,
      "eval_runtime": 578.4314,
      "eval_samples_per_second": 4.989,
      "eval_steps_per_second": 0.157,
      "learning_rate": 0.0001,
      "step": 17745
    },
    {
      "epoch": 65.93406593406593,
      "grad_norm": 0.278421550989151,
      "learning_rate": 0.0001,
      "loss": 0.1258,
      "step": 18000
    },
    {
      "epoch": 66.0,
      "eval_accuracy": 0.30353430353430355,
      "eval_f1_macro": 0.706549053618703,
      "eval_f1_micro": 0.8178908425822006,
      "eval_loss": 0.12264254689216614,
      "eval_runtime": 582.1432,
      "eval_samples_per_second": 4.958,
      "eval_steps_per_second": 0.156,
      "learning_rate": 0.0001,
      "step": 18018
    },
    {
      "epoch": 67.0,
      "eval_accuracy": 0.3052668052668053,
      "eval_f1_macro": 0.7104724530407897,
      "eval_f1_micro": 0.8184709429598117,
      "eval_loss": 0.12282951176166534,
      "eval_runtime": 572.3604,
      "eval_samples_per_second": 5.042,
      "eval_steps_per_second": 0.159,
      "learning_rate": 0.0001,
      "step": 18291
    },
    {
      "epoch": 67.76556776556777,
      "grad_norm": 0.33604416251182556,
      "learning_rate": 0.0001,
      "loss": 0.125,
      "step": 18500
    },
    {
      "epoch": 68.0,
      "eval_accuracy": 0.3042273042273042,
      "eval_f1_macro": 0.7127667888432785,
      "eval_f1_micro": 0.818125541661508,
      "eval_loss": 0.12277437746524811,
      "eval_runtime": 566.9946,
      "eval_samples_per_second": 5.09,
      "eval_steps_per_second": 0.16,
      "learning_rate": 0.0001,
      "step": 18564
    },
    {
      "epoch": 69.0,
      "eval_accuracy": 0.3052668052668053,
      "eval_f1_macro": 0.703803449192515,
      "eval_f1_micro": 0.8137196924896511,
      "eval_loss": 0.12282923609018326,
      "eval_runtime": 578.0179,
      "eval_samples_per_second": 4.993,
      "eval_steps_per_second": 0.157,
      "learning_rate": 0.0001,
      "step": 18837
    },
    {
      "epoch": 69.59706959706959,
      "grad_norm": 0.237099289894104,
      "learning_rate": 0.0001,
      "loss": 0.125,
      "step": 19000
    },
    {
      "epoch": 70.0,
      "eval_accuracy": 0.30180180180180183,
      "eval_f1_macro": 0.7079779831387449,
      "eval_f1_micro": 0.8154847434705434,
      "eval_loss": 0.1231524795293808,
      "eval_runtime": 578.443,
      "eval_samples_per_second": 4.989,
      "eval_steps_per_second": 0.157,
      "learning_rate": 0.0001,
      "step": 19110
    },
    {
      "epoch": 71.0,
      "eval_accuracy": 0.29902979902979904,
      "eval_f1_macro": 0.7111485120209401,
      "eval_f1_micro": 0.8155765340525961,
      "eval_loss": 0.12310803681612015,
      "eval_runtime": 577.0393,
      "eval_samples_per_second": 5.001,
      "eval_steps_per_second": 0.158,
      "learning_rate": 0.0001,
      "step": 19383
    },
    {
      "epoch": 71.42857142857143,
      "grad_norm": 0.31944581866264343,
      "learning_rate": 0.0001,
      "loss": 0.1245,
      "step": 19500
    },
    {
      "epoch": 72.0,
      "eval_accuracy": 0.30076230076230076,
      "eval_f1_macro": 0.7149554589558469,
      "eval_f1_micro": 0.8161826422695988,
      "eval_loss": 0.12233748286962509,
      "eval_runtime": 578.4014,
      "eval_samples_per_second": 4.99,
      "eval_steps_per_second": 0.157,
      "learning_rate": 0.0001,
      "step": 19656
    },
    {
      "epoch": 73.0,
      "eval_accuracy": 0.3049203049203049,
      "eval_f1_macro": 0.7041677932151664,
      "eval_f1_micro": 0.8173891171292027,
      "eval_loss": 0.12232980877161026,
      "eval_runtime": 572.1622,
      "eval_samples_per_second": 5.044,
      "eval_steps_per_second": 0.159,
      "learning_rate": 0.0001,
      "step": 19929
    },
    {
      "epoch": 73.26007326007326,
      "grad_norm": 0.2972647249698639,
      "learning_rate": 0.0001,
      "loss": 0.1248,
      "step": 20000
    },
    {
      "epoch": 74.0,
      "eval_accuracy": 0.29625779625779625,
      "eval_f1_macro": 0.7009378784168091,
      "eval_f1_micro": 0.8124655033329088,
      "eval_loss": 0.12370481342077255,
      "eval_runtime": 589.0452,
      "eval_samples_per_second": 4.899,
      "eval_steps_per_second": 0.154,
      "learning_rate": 0.0001,
      "step": 20202
    },
    {
      "epoch": 75.0,
      "eval_accuracy": 0.30457380457380456,
      "eval_f1_macro": 0.7045397920320168,
      "eval_f1_micro": 0.8151614807319335,
      "eval_loss": 0.12251079827547073,
      "eval_runtime": 585.6991,
      "eval_samples_per_second": 4.927,
      "eval_steps_per_second": 0.155,
      "learning_rate": 0.0001,
      "step": 20475
    },
    {
      "epoch": 75.0915750915751,
      "grad_norm": 0.34343692660331726,
      "learning_rate": 0.0001,
      "loss": 0.1249,
      "step": 20500
    },
    {
      "epoch": 76.0,
      "eval_accuracy": 0.30076230076230076,
      "eval_f1_macro": 0.7099467053038022,
      "eval_f1_micro": 0.8159596808063839,
      "eval_loss": 0.12471619248390198,
      "eval_runtime": 580.0788,
      "eval_samples_per_second": 4.975,
      "eval_steps_per_second": 0.157,
      "learning_rate": 0.0001,
      "step": 20748
    },
    {
      "epoch": 76.92307692307692,
      "grad_norm": 0.34839001297950745,
      "learning_rate": 0.0001,
      "loss": 0.1238,
      "step": 21000
    },
    {
      "epoch": 77.0,
      "eval_accuracy": 0.29902979902979904,
      "eval_f1_macro": 0.7139220834888879,
      "eval_f1_micro": 0.8179482930062977,
      "eval_loss": 0.1225149855017662,
      "eval_runtime": 592.1949,
      "eval_samples_per_second": 4.873,
      "eval_steps_per_second": 0.154,
      "learning_rate": 0.0001,
      "step": 21021
    },
    {
      "epoch": 78.0,
      "eval_accuracy": 0.30457380457380456,
      "eval_f1_macro": 0.7060530834173799,
      "eval_f1_micro": 0.8188081009296149,
      "eval_loss": 0.12218895554542542,
      "eval_runtime": 580.5057,
      "eval_samples_per_second": 4.972,
      "eval_steps_per_second": 0.157,
      "learning_rate": 0.0001,
      "step": 21294
    },
    {
      "epoch": 78.75457875457876,
      "grad_norm": 0.3252258002758026,
      "learning_rate": 0.0001,
      "loss": 0.1233,
      "step": 21500
    },
    {
      "epoch": 79.0,
      "eval_accuracy": 0.30180180180180183,
      "eval_f1_macro": 0.7100736740896743,
      "eval_f1_micro": 0.8152018201735907,
      "eval_loss": 0.12460680305957794,
      "eval_runtime": 580.1369,
      "eval_samples_per_second": 4.975,
      "eval_steps_per_second": 0.157,
      "learning_rate": 0.0001,
      "step": 21567
    },
    {
      "epoch": 80.0,
      "eval_accuracy": 0.3038808038808039,
      "eval_f1_macro": 0.710254660316209,
      "eval_f1_micro": 0.8179503235232728,
      "eval_loss": 0.12210072576999664,
      "eval_runtime": 588.8207,
      "eval_samples_per_second": 4.901,
      "eval_steps_per_second": 0.155,
      "learning_rate": 0.0001,
      "step": 21840
    },
    {
      "epoch": 80.58608058608058,
      "grad_norm": 0.31095778942108154,
      "learning_rate": 0.0001,
      "loss": 0.1225,
      "step": 22000
    },
    {
      "epoch": 81.0,
      "eval_accuracy": 0.30180180180180183,
      "eval_f1_macro": 0.7156736962695093,
      "eval_f1_micro": 0.8184843191082273,
      "eval_loss": 0.12116113305091858,
      "eval_runtime": 583.0708,
      "eval_samples_per_second": 4.95,
      "eval_steps_per_second": 0.156,
      "learning_rate": 0.0001,
      "step": 22113
    },
    {
      "epoch": 82.0,
      "eval_accuracy": 0.30803880803880807,
      "eval_f1_macro": 0.7088973142030256,
      "eval_f1_micro": 0.8151834668916069,
      "eval_loss": 0.12159755080938339,
      "eval_runtime": 583.0652,
      "eval_samples_per_second": 4.95,
      "eval_steps_per_second": 0.156,
      "learning_rate": 0.0001,
      "step": 22386
    },
    {
      "epoch": 82.41758241758242,
      "grad_norm": 0.30886727571487427,
      "learning_rate": 0.0001,
      "loss": 0.1216,
      "step": 22500
    },
    {
      "epoch": 83.0,
      "eval_accuracy": 0.30803880803880807,
      "eval_f1_macro": 0.7090314532601811,
      "eval_f1_micro": 0.8164913756836348,
      "eval_loss": 0.12142453342676163,
      "eval_runtime": 587.9863,
      "eval_samples_per_second": 4.908,
      "eval_steps_per_second": 0.155,
      "learning_rate": 0.0001,
      "step": 22659
    },
    {
      "epoch": 84.0,
      "eval_accuracy": 0.306999306999307,
      "eval_f1_macro": 0.7100088286352488,
      "eval_f1_micro": 0.8168789808917197,
      "eval_loss": 0.12157817929983139,
      "eval_runtime": 574.5086,
      "eval_samples_per_second": 5.023,
      "eval_steps_per_second": 0.158,
      "learning_rate": 0.0001,
      "step": 22932
    },
    {
      "epoch": 84.24908424908425,
      "grad_norm": 0.33512353897094727,
      "learning_rate": 0.0001,
      "loss": 0.1232,
      "step": 23000
    },
    {
      "epoch": 85.0,
      "eval_accuracy": 0.3052668052668053,
      "eval_f1_macro": 0.7109093140233731,
      "eval_f1_micro": 0.8188048474717927,
      "eval_loss": 0.12155645340681076,
      "eval_runtime": 588.285,
      "eval_samples_per_second": 4.906,
      "eval_steps_per_second": 0.155,
      "learning_rate": 0.0001,
      "step": 23205
    },
    {
      "epoch": 86.0,
      "eval_accuracy": 0.306999306999307,
      "eval_f1_macro": 0.7176309861117636,
      "eval_f1_micro": 0.8190986316274009,
      "eval_loss": 0.12185127288103104,
      "eval_runtime": 596.3097,
      "eval_samples_per_second": 4.84,
      "eval_steps_per_second": 0.153,
      "learning_rate": 0.0001,
      "step": 23478
    },
    {
      "epoch": 86.08058608058609,
      "grad_norm": 0.3552737832069397,
      "learning_rate": 0.0001,
      "loss": 0.1221,
      "step": 23500
    },
    {
      "epoch": 87.0,
      "eval_accuracy": 0.3063063063063063,
      "eval_f1_macro": 0.7080074645247395,
      "eval_f1_micro": 0.8176601181250785,
      "eval_loss": 0.12196006625890732,
      "eval_runtime": 582.057,
      "eval_samples_per_second": 4.958,
      "eval_steps_per_second": 0.156,
      "learning_rate": 0.0001,
      "step": 23751
    },
    {
      "epoch": 87.91208791208791,
      "grad_norm": 0.31451234221458435,
      "learning_rate": 1e-05,
      "loss": 0.1208,
      "step": 24000
    },
    {
      "epoch": 88.0,
      "eval_accuracy": 0.3049203049203049,
      "eval_f1_macro": 0.7158455151348946,
      "eval_f1_micro": 0.8210709982967056,
      "eval_loss": 0.1209772601723671,
      "eval_runtime": 583.2281,
      "eval_samples_per_second": 4.948,
      "eval_steps_per_second": 0.156,
      "learning_rate": 1e-05,
      "step": 24024
    },
    {
      "epoch": 89.0,
      "eval_accuracy": 0.30734580734580735,
      "eval_f1_macro": 0.7312085502704033,
      "eval_f1_micro": 0.82409381663113,
      "eval_loss": 0.12103869765996933,
      "eval_runtime": 571.5547,
      "eval_samples_per_second": 5.049,
      "eval_steps_per_second": 0.159,
      "learning_rate": 1e-05,
      "step": 24297
    },
    {
      "epoch": 89.74358974358974,
      "grad_norm": 0.3047947883605957,
      "learning_rate": 1e-05,
      "loss": 0.1189,
      "step": 24500
    },
    {
      "epoch": 90.0,
      "eval_accuracy": 0.306999306999307,
      "eval_f1_macro": 0.7232165780756572,
      "eval_f1_micro": 0.823466447097571,
      "eval_loss": 0.12060839682817459,
      "eval_runtime": 577.3382,
      "eval_samples_per_second": 4.999,
      "eval_steps_per_second": 0.158,
      "learning_rate": 1e-05,
      "step": 24570
    },
    {
      "epoch": 91.0,
      "eval_accuracy": 0.3087318087318087,
      "eval_f1_macro": 0.714800773362679,
      "eval_f1_micro": 0.8190853196327803,
      "eval_loss": 0.12036494165658951,
      "eval_runtime": 575.4555,
      "eval_samples_per_second": 5.015,
      "eval_steps_per_second": 0.158,
      "learning_rate": 1e-05,
      "step": 24843
    },
    {
      "epoch": 91.57509157509158,
      "grad_norm": 0.28365448117256165,
      "learning_rate": 1e-05,
      "loss": 0.1181,
      "step": 25000
    },
    {
      "epoch": 92.0,
      "eval_accuracy": 0.3087318087318087,
      "eval_f1_macro": 0.7131694962358415,
      "eval_f1_micro": 0.8194187804468336,
      "eval_loss": 0.12028194963932037,
      "eval_runtime": 576.0948,
      "eval_samples_per_second": 5.01,
      "eval_steps_per_second": 0.158,
      "learning_rate": 1e-05,
      "step": 25116
    },
    {
      "epoch": 93.0,
      "eval_accuracy": 0.30838530838530837,
      "eval_f1_macro": 0.7183857334665726,
      "eval_f1_micro": 0.8214921910601102,
      "eval_loss": 0.12036142498254776,
      "eval_runtime": 576.2755,
      "eval_samples_per_second": 5.008,
      "eval_steps_per_second": 0.158,
      "learning_rate": 1e-05,
      "step": 25389
    },
    {
      "epoch": 93.4065934065934,
      "grad_norm": 0.3265780806541443,
      "learning_rate": 1e-05,
      "loss": 0.1183,
      "step": 25500
    },
    {
      "epoch": 94.0,
      "eval_accuracy": 0.30665280665280664,
      "eval_f1_macro": 0.7195089038891936,
      "eval_f1_micro": 0.8207247828991316,
      "eval_loss": 0.12012535333633423,
      "eval_runtime": 574.7766,
      "eval_samples_per_second": 5.021,
      "eval_steps_per_second": 0.158,
      "learning_rate": 1e-05,
      "step": 25662
    },
    {
      "epoch": 95.0,
      "eval_accuracy": 0.30838530838530837,
      "eval_f1_macro": 0.7157953029268052,
      "eval_f1_micro": 0.8197421299397187,
      "eval_loss": 0.12010551244020462,
      "eval_runtime": 579.3324,
      "eval_samples_per_second": 4.982,
      "eval_steps_per_second": 0.157,
      "learning_rate": 1e-05,
      "step": 25935
    },
    {
      "epoch": 95.23809523809524,
      "grad_norm": 0.3151456415653229,
      "learning_rate": 1e-05,
      "loss": 0.117,
      "step": 26000
    },
    {
      "epoch": 96.0,
      "eval_accuracy": 0.3052668052668053,
      "eval_f1_macro": 0.7193329750108997,
      "eval_f1_micro": 0.8217099503939306,
      "eval_loss": 0.1198035180568695,
      "eval_runtime": 605.2902,
      "eval_samples_per_second": 4.768,
      "eval_steps_per_second": 0.15,
      "learning_rate": 1e-05,
      "step": 26208
    },
    {
      "epoch": 97.0,
      "eval_accuracy": 0.3063063063063063,
      "eval_f1_macro": 0.7210913273742954,
      "eval_f1_micro": 0.8204592028773368,
      "eval_loss": 0.12007978558540344,
      "eval_runtime": 637.4782,
      "eval_samples_per_second": 4.527,
      "eval_steps_per_second": 0.143,
      "learning_rate": 1e-05,
      "step": 26481
    },
    {
      "epoch": 97.06959706959707,
      "grad_norm": 0.32921385765075684,
      "learning_rate": 1e-05,
      "loss": 0.1176,
      "step": 26500
    },
    {
      "epoch": 98.0,
      "eval_accuracy": 0.30803880803880807,
      "eval_f1_macro": 0.7250050738866957,
      "eval_f1_micro": 0.8226615276223631,
      "eval_loss": 0.12013950198888779,
      "eval_runtime": 575.4419,
      "eval_samples_per_second": 5.015,
      "eval_steps_per_second": 0.158,
      "learning_rate": 1e-05,
      "step": 26754
    },
    {
      "epoch": 98.9010989010989,
      "grad_norm": 0.37390851974487305,
      "learning_rate": 1e-05,
      "loss": 0.1176,
      "step": 27000
    },
    {
      "epoch": 99.0,
      "eval_accuracy": 0.30734580734580735,
      "eval_f1_macro": 0.7226446473273916,
      "eval_f1_micro": 0.8206050968740846,
      "eval_loss": 0.12003140896558762,
      "eval_runtime": 619.3031,
      "eval_samples_per_second": 4.66,
      "eval_steps_per_second": 0.147,
      "learning_rate": 1e-05,
      "step": 27027
    },
    {
      "epoch": 100.0,
      "eval_accuracy": 0.30803880803880807,
      "eval_f1_macro": 0.7191058080317732,
      "eval_f1_micro": 0.8216021451315569,
      "eval_loss": 0.1200241968035698,
      "eval_runtime": 615.3025,
      "eval_samples_per_second": 4.69,
      "eval_steps_per_second": 0.148,
      "learning_rate": 1e-05,
      "step": 27300
    },
    {
      "epoch": 100.73260073260073,
      "grad_norm": 0.39757904410362244,
      "learning_rate": 1e-05,
      "loss": 0.117,
      "step": 27500
    },
    {
      "epoch": 101.0,
      "eval_accuracy": 0.3097713097713098,
      "eval_f1_macro": 0.724185821594473,
      "eval_f1_micro": 0.8228338260398884,
      "eval_loss": 0.11989927291870117,
      "eval_runtime": 574.4571,
      "eval_samples_per_second": 5.024,
      "eval_steps_per_second": 0.158,
      "learning_rate": 1e-05,
      "step": 27573
    },
    {
      "epoch": 102.0,
      "eval_accuracy": 0.30734580734580735,
      "eval_f1_macro": 0.7216480982679994,
      "eval_f1_micro": 0.8192071374463429,
      "eval_loss": 0.12043397128582001,
      "eval_runtime": 627.7673,
      "eval_samples_per_second": 4.597,
      "eval_steps_per_second": 0.145,
      "learning_rate": 1e-05,
      "step": 27846
    },
    {
      "epoch": 102.56410256410257,
      "grad_norm": 0.3315370976924896,
      "learning_rate": 1.0000000000000002e-06,
      "loss": 0.1159,
      "step": 28000
    },
    {
      "epoch": 103.0,
      "eval_accuracy": 0.30665280665280664,
      "eval_f1_macro": 0.7231653777895258,
      "eval_f1_micro": 0.8221990402670561,
      "eval_loss": 0.1199527159333229,
      "eval_runtime": 574.6395,
      "eval_samples_per_second": 5.022,
      "eval_steps_per_second": 0.158,
      "learning_rate": 1.0000000000000002e-06,
      "step": 28119
    },
    {
      "epoch": 104.0,
      "eval_accuracy": 0.31011781011781014,
      "eval_f1_macro": 0.7234886652066129,
      "eval_f1_micro": 0.8217751676454663,
      "eval_loss": 0.12043838202953339,
      "eval_runtime": 571.7276,
      "eval_samples_per_second": 5.048,
      "eval_steps_per_second": 0.159,
      "learning_rate": 1.0000000000000002e-06,
      "step": 28392
    },
    {
      "epoch": 104.3956043956044,
      "grad_norm": 0.3368052840232849,
      "learning_rate": 1.0000000000000002e-06,
      "loss": 0.1151,
      "step": 28500
    },
    {
      "epoch": 105.0,
      "eval_accuracy": 0.3076923076923077,
      "eval_f1_macro": 0.7205886261735153,
      "eval_f1_micro": 0.8219212232009308,
      "eval_loss": 0.11985885351896286,
      "eval_runtime": 623.1101,
      "eval_samples_per_second": 4.632,
      "eval_steps_per_second": 0.146,
      "learning_rate": 1.0000000000000002e-06,
      "step": 28665
    },
    {
      "epoch": 106.0,
      "eval_accuracy": 0.31046431046431044,
      "eval_f1_macro": 0.7270103073654017,
      "eval_f1_micro": 0.8228195739014648,
      "eval_loss": 0.11984959244728088,
      "eval_runtime": 582.6782,
      "eval_samples_per_second": 4.953,
      "eval_steps_per_second": 0.156,
      "learning_rate": 1.0000000000000002e-06,
      "step": 28938
    },
    {
      "epoch": 106.0,
      "learning_rate": 1.0000000000000002e-06,
      "step": 28938,
      "total_flos": 1.3699618367216817e+21,
      "train_loss": 0.13672988786670365,
      "train_runtime": 251282.626,
      "train_samples_per_second": 5.203,
      "train_steps_per_second": 0.163
    }
  ],
  "logging_steps": 500,
  "max_steps": 40950,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 150,
  "save_steps": 500,
  "stateful_callbacks": {
    "EarlyStoppingCallback": {
      "args": {
        "early_stopping_patience": 10,
        "early_stopping_threshold": 0.0
      },
      "attributes": {
        "early_stopping_patience_counter": 0
      }
    },
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 1.3699618367216817e+21,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": null
}