{
  "best_metric": 0.24052156507968903,
  "best_model_checkpoint": "./oral_disease_mobilevit/checkpoint-4600",
  "epoch": 5.0,
  "eval_steps": 100,
  "global_step": 5830,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.04288164665523156,
      "grad_norm": 2.5394277572631836,
      "learning_rate": 4.957118353344769e-05,
      "loss": 1.7703,
      "step": 50
    },
    {
      "epoch": 0.08576329331046312,
      "grad_norm": 2.5862715244293213,
      "learning_rate": 4.914236706689537e-05,
      "loss": 1.6871,
      "step": 100
    },
    {
      "epoch": 0.08576329331046312,
      "eval_accuracy": 0.577005577005577,
      "eval_loss": 1.6024609804153442,
      "eval_runtime": 14.1663,
      "eval_samples_per_second": 164.545,
      "eval_steps_per_second": 20.612,
      "step": 100
    },
    {
      "epoch": 0.12864493996569468,
      "grad_norm": 3.443279504776001,
      "learning_rate": 4.8713550600343055e-05,
      "loss": 1.5656,
      "step": 150
    },
    {
      "epoch": 0.17152658662092624,
      "grad_norm": 2.7117512226104736,
      "learning_rate": 4.828473413379074e-05,
      "loss": 1.4961,
      "step": 200
    },
    {
      "epoch": 0.17152658662092624,
      "eval_accuracy": 0.5971685971685972,
      "eval_loss": 1.3290797472000122,
      "eval_runtime": 14.4239,
      "eval_samples_per_second": 161.607,
      "eval_steps_per_second": 20.244,
      "step": 200
    },
    {
      "epoch": 0.2144082332761578,
      "grad_norm": 3.41520619392395,
      "learning_rate": 4.7855917667238424e-05,
      "loss": 1.3941,
      "step": 250
    },
    {
      "epoch": 0.25728987993138935,
      "grad_norm": 2.960935354232788,
      "learning_rate": 4.742710120068611e-05,
      "loss": 1.2984,
      "step": 300
    },
    {
      "epoch": 0.25728987993138935,
      "eval_accuracy": 0.6525096525096525,
      "eval_loss": 1.1308239698410034,
      "eval_runtime": 14.2271,
      "eval_samples_per_second": 163.842,
      "eval_steps_per_second": 20.524,
      "step": 300
    },
    {
      "epoch": 0.30017152658662094,
      "grad_norm": 3.7776384353637695,
      "learning_rate": 4.699828473413379e-05,
      "loss": 1.2457,
      "step": 350
    },
    {
      "epoch": 0.34305317324185247,
      "grad_norm": 6.4286088943481445,
      "learning_rate": 4.656946826758148e-05,
      "loss": 1.12,
      "step": 400
    },
    {
      "epoch": 0.34305317324185247,
      "eval_accuracy": 0.7232947232947233,
      "eval_loss": 0.9413465261459351,
      "eval_runtime": 14.2093,
      "eval_samples_per_second": 164.048,
      "eval_steps_per_second": 20.55,
      "step": 400
    },
    {
      "epoch": 0.38593481989708406,
      "grad_norm": 3.400545120239258,
      "learning_rate": 4.614065180102916e-05,
      "loss": 1.0816,
      "step": 450
    },
    {
      "epoch": 0.4288164665523156,
      "grad_norm": 4.862150192260742,
      "learning_rate": 4.5711835334476845e-05,
      "loss": 1.0873,
      "step": 500
    },
    {
      "epoch": 0.4288164665523156,
      "eval_accuracy": 0.7610467610467611,
      "eval_loss": 0.8114073872566223,
      "eval_runtime": 14.2607,
      "eval_samples_per_second": 163.456,
      "eval_steps_per_second": 20.476,
      "step": 500
    },
    {
      "epoch": 0.4716981132075472,
      "grad_norm": 3.4492220878601074,
      "learning_rate": 4.528301886792453e-05,
      "loss": 0.9814,
      "step": 550
    },
    {
      "epoch": 0.5145797598627787,
      "grad_norm": 3.7629287242889404,
      "learning_rate": 4.4854202401372214e-05,
      "loss": 1.0012,
      "step": 600
    },
    {
      "epoch": 0.5145797598627787,
      "eval_accuracy": 0.7919347919347919,
      "eval_loss": 0.6951708197593689,
      "eval_runtime": 13.8986,
      "eval_samples_per_second": 167.714,
      "eval_steps_per_second": 21.009,
      "step": 600
    },
    {
      "epoch": 0.5574614065180102,
      "grad_norm": 4.381726264953613,
      "learning_rate": 4.4425385934819905e-05,
      "loss": 0.9516,
      "step": 650
    },
    {
      "epoch": 0.6003430531732419,
      "grad_norm": 3.579646348953247,
      "learning_rate": 4.399656946826758e-05,
      "loss": 0.8379,
      "step": 700
    },
    {
      "epoch": 0.6003430531732419,
      "eval_accuracy": 0.8125268125268126,
      "eval_loss": 0.6312650442123413,
      "eval_runtime": 13.8871,
      "eval_samples_per_second": 167.853,
      "eval_steps_per_second": 21.027,
      "step": 700
    },
    {
      "epoch": 0.6432246998284734,
      "grad_norm": 5.985883712768555,
      "learning_rate": 4.356775300171527e-05,
      "loss": 0.8269,
      "step": 750
    },
    {
      "epoch": 0.6861063464837049,
      "grad_norm": 5.968649864196777,
      "learning_rate": 4.313893653516296e-05,
      "loss": 0.8627,
      "step": 800
    },
    {
      "epoch": 0.6861063464837049,
      "eval_accuracy": 0.8181038181038182,
      "eval_loss": 0.5455917716026306,
      "eval_runtime": 13.9813,
      "eval_samples_per_second": 166.723,
      "eval_steps_per_second": 20.885,
      "step": 800
    },
    {
      "epoch": 0.7289879931389366,
      "grad_norm": 5.903740882873535,
      "learning_rate": 4.2710120068610635e-05,
      "loss": 0.8055,
      "step": 850
    },
    {
      "epoch": 0.7718696397941681,
      "grad_norm": 6.328656196594238,
      "learning_rate": 4.228130360205832e-05,
      "loss": 0.8252,
      "step": 900
    },
    {
      "epoch": 0.7718696397941681,
      "eval_accuracy": 0.8314028314028314,
      "eval_loss": 0.5072463750839233,
      "eval_runtime": 13.8768,
      "eval_samples_per_second": 167.979,
      "eval_steps_per_second": 21.042,
      "step": 900
    },
    {
      "epoch": 0.8147512864493996,
      "grad_norm": 4.45211124420166,
      "learning_rate": 4.185248713550601e-05,
      "loss": 0.7173,
      "step": 950
    },
    {
      "epoch": 0.8576329331046312,
      "grad_norm": 5.749237537384033,
      "learning_rate": 4.142367066895369e-05,
      "loss": 0.7223,
      "step": 1000
    },
    {
      "epoch": 0.8576329331046312,
      "eval_accuracy": 0.8378378378378378,
      "eval_loss": 0.46284219622612,
      "eval_runtime": 14.2987,
      "eval_samples_per_second": 163.022,
      "eval_steps_per_second": 20.421,
      "step": 1000
    },
    {
      "epoch": 0.9005145797598628,
      "grad_norm": 16.25695037841797,
      "learning_rate": 4.099485420240137e-05,
      "loss": 0.7153,
      "step": 1050
    },
    {
      "epoch": 0.9433962264150944,
      "grad_norm": 7.836055755615234,
      "learning_rate": 4.0566037735849064e-05,
      "loss": 0.6617,
      "step": 1100
    },
    {
      "epoch": 0.9433962264150944,
      "eval_accuracy": 0.8468468468468469,
      "eval_loss": 0.43263953924179077,
      "eval_runtime": 13.9551,
      "eval_samples_per_second": 167.035,
      "eval_steps_per_second": 20.924,
      "step": 1100
    },
    {
      "epoch": 0.9862778730703259,
      "grad_norm": 8.199559211730957,
      "learning_rate": 4.013722126929674e-05,
      "loss": 0.7156,
      "step": 1150
    },
    {
      "epoch": 1.0291595197255574,
      "grad_norm": 8.776689529418945,
      "learning_rate": 3.9708404802744425e-05,
      "loss": 0.7784,
      "step": 1200
    },
    {
      "epoch": 1.0291595197255574,
      "eval_accuracy": 0.8511368511368511,
      "eval_loss": 0.42261621356010437,
      "eval_runtime": 14.1288,
      "eval_samples_per_second": 164.982,
      "eval_steps_per_second": 20.667,
      "step": 1200
    },
    {
      "epoch": 1.072041166380789,
      "grad_norm": 5.693437099456787,
      "learning_rate": 3.9279588336192116e-05,
      "loss": 0.6163,
      "step": 1250
    },
    {
      "epoch": 1.1149228130360207,
      "grad_norm": 16.413869857788086,
      "learning_rate": 3.8850771869639794e-05,
      "loss": 0.6086,
      "step": 1300
    },
    {
      "epoch": 1.1149228130360207,
      "eval_accuracy": 0.8592878592878593,
      "eval_loss": 0.3927444517612457,
      "eval_runtime": 13.8914,
      "eval_samples_per_second": 167.801,
      "eval_steps_per_second": 21.02,
      "step": 1300
    },
    {
      "epoch": 1.1578044596912522,
      "grad_norm": 4.260317325592041,
      "learning_rate": 3.842195540308748e-05,
      "loss": 0.6237,
      "step": 1350
    },
    {
      "epoch": 1.2006861063464838,
      "grad_norm": 8.143071174621582,
      "learning_rate": 3.799313893653517e-05,
      "loss": 0.6196,
      "step": 1400
    },
    {
      "epoch": 1.2006861063464838,
      "eval_accuracy": 0.8618618618618619,
      "eval_loss": 0.3923815190792084,
      "eval_runtime": 13.8559,
      "eval_samples_per_second": 168.231,
      "eval_steps_per_second": 21.074,
      "step": 1400
    },
    {
      "epoch": 1.2435677530017153,
      "grad_norm": 8.181282043457031,
      "learning_rate": 3.756432246998285e-05,
      "loss": 0.5606,
      "step": 1450
    },
    {
      "epoch": 1.2864493996569468,
      "grad_norm": 14.405867576599121,
      "learning_rate": 3.713550600343053e-05,
      "loss": 0.6561,
      "step": 1500
    },
    {
      "epoch": 1.2864493996569468,
      "eval_accuracy": 0.8601458601458601,
      "eval_loss": 0.36444056034088135,
      "eval_runtime": 13.8772,
      "eval_samples_per_second": 167.974,
      "eval_steps_per_second": 21.042,
      "step": 1500
    },
    {
      "epoch": 1.3293310463121784,
      "grad_norm": 14.855607986450195,
      "learning_rate": 3.670668953687822e-05,
      "loss": 0.5548,
      "step": 1550
    },
    {
      "epoch": 1.3722126929674099,
      "grad_norm": 4.9101409912109375,
      "learning_rate": 3.62778730703259e-05,
      "loss": 0.5426,
      "step": 1600
    },
    {
      "epoch": 1.3722126929674099,
      "eval_accuracy": 0.8738738738738738,
      "eval_loss": 0.3387666344642639,
      "eval_runtime": 14.6787,
      "eval_samples_per_second": 158.801,
      "eval_steps_per_second": 19.893,
      "step": 1600
    },
    {
      "epoch": 1.4150943396226414,
      "grad_norm": 3.1763756275177,
      "learning_rate": 3.5849056603773584e-05,
      "loss": 0.5917,
      "step": 1650
    },
    {
      "epoch": 1.457975986277873,
      "grad_norm": 2.9457991123199463,
      "learning_rate": 3.5420240137221275e-05,
      "loss": 0.5858,
      "step": 1700
    },
    {
      "epoch": 1.457975986277873,
      "eval_accuracy": 0.8652938652938653,
      "eval_loss": 0.3525221347808838,
      "eval_runtime": 14.1719,
      "eval_samples_per_second": 164.48,
      "eval_steps_per_second": 20.604,
      "step": 1700
    },
    {
      "epoch": 1.5008576329331045,
      "grad_norm": 7.572704792022705,
      "learning_rate": 3.499142367066895e-05,
      "loss": 0.6098,
      "step": 1750
    },
    {
      "epoch": 1.5437392795883362,
      "grad_norm": 1.232725739479065,
      "learning_rate": 3.456260720411664e-05,
      "loss": 0.574,
      "step": 1800
    },
    {
      "epoch": 1.5437392795883362,
      "eval_accuracy": 0.8764478764478765,
      "eval_loss": 0.32060739398002625,
      "eval_runtime": 14.3858,
      "eval_samples_per_second": 162.035,
      "eval_steps_per_second": 20.298,
      "step": 1800
    },
    {
      "epoch": 1.5866209262435678,
      "grad_norm": 10.918987274169922,
      "learning_rate": 3.413379073756433e-05,
      "loss": 0.5824,
      "step": 1850
    },
    {
      "epoch": 1.6295025728987993,
      "grad_norm": 10.205852508544922,
      "learning_rate": 3.3704974271012005e-05,
      "loss": 0.5015,
      "step": 1900
    },
    {
      "epoch": 1.6295025728987993,
      "eval_accuracy": 0.882024882024882,
      "eval_loss": 0.32507938146591187,
      "eval_runtime": 14.2196,
      "eval_samples_per_second": 163.929,
      "eval_steps_per_second": 20.535,
      "step": 1900
    },
    {
      "epoch": 1.6723842195540308,
      "grad_norm": 7.081895351409912,
      "learning_rate": 3.3276157804459696e-05,
      "loss": 0.4681,
      "step": 1950
    },
    {
      "epoch": 1.7152658662092626,
      "grad_norm": 9.101844787597656,
      "learning_rate": 3.284734133790738e-05,
      "loss": 0.493,
      "step": 2000
    },
    {
      "epoch": 1.7152658662092626,
      "eval_accuracy": 0.8854568854568855,
      "eval_loss": 0.308988094329834,
      "eval_runtime": 14.3569,
      "eval_samples_per_second": 162.36,
      "eval_steps_per_second": 20.339,
      "step": 2000
    },
    {
      "epoch": 1.758147512864494,
      "grad_norm": 6.105254650115967,
      "learning_rate": 3.241852487135506e-05,
      "loss": 0.533,
      "step": 2050
    },
    {
      "epoch": 1.8010291595197256,
      "grad_norm": 14.518338203430176,
      "learning_rate": 3.198970840480275e-05,
      "loss": 0.5381,
      "step": 2100
    },
    {
      "epoch": 1.8010291595197256,
      "eval_accuracy": 0.8773058773058773,
      "eval_loss": 0.3106726109981537,
      "eval_runtime": 14.2106,
      "eval_samples_per_second": 164.033,
      "eval_steps_per_second": 20.548,
      "step": 2100
    },
    {
      "epoch": 1.8439108061749572,
      "grad_norm": 3.1229450702667236,
      "learning_rate": 3.1560891938250434e-05,
      "loss": 0.5641,
      "step": 2150
    },
    {
      "epoch": 1.8867924528301887,
      "grad_norm": 14.663033485412598,
      "learning_rate": 3.113207547169811e-05,
      "loss": 0.4979,
      "step": 2200
    },
    {
      "epoch": 1.8867924528301887,
      "eval_accuracy": 0.8901758901758902,
      "eval_loss": 0.2851434051990509,
      "eval_runtime": 14.2521,
      "eval_samples_per_second": 163.554,
      "eval_steps_per_second": 20.488,
      "step": 2200
    },
    {
      "epoch": 1.9296740994854202,
      "grad_norm": 3.774958848953247,
      "learning_rate": 3.07032590051458e-05,
      "loss": 0.5117,
      "step": 2250
    },
    {
      "epoch": 1.9725557461406518,
      "grad_norm": 25.906217575073242,
      "learning_rate": 3.0274442538593483e-05,
      "loss": 0.5145,
      "step": 2300
    },
    {
      "epoch": 1.9725557461406518,
      "eval_accuracy": 0.8927498927498928,
      "eval_loss": 0.2909970283508301,
      "eval_runtime": 14.2775,
      "eval_samples_per_second": 163.264,
      "eval_steps_per_second": 20.452,
      "step": 2300
    },
    {
      "epoch": 2.0154373927958833,
      "grad_norm": 8.114184379577637,
      "learning_rate": 2.9845626072041167e-05,
      "loss": 0.4499,
      "step": 2350
    },
    {
      "epoch": 2.058319039451115,
      "grad_norm": 9.68442153930664,
      "learning_rate": 2.9416809605488855e-05,
      "loss": 0.4846,
      "step": 2400
    },
    {
      "epoch": 2.058319039451115,
      "eval_accuracy": 0.8833118833118833,
      "eval_loss": 0.3212469518184662,
      "eval_runtime": 14.6038,
      "eval_samples_per_second": 159.616,
      "eval_steps_per_second": 19.995,
      "step": 2400
    },
    {
      "epoch": 2.1012006861063464,
      "grad_norm": 3.7848308086395264,
      "learning_rate": 2.8987993138936536e-05,
      "loss": 0.4533,
      "step": 2450
    },
    {
      "epoch": 2.144082332761578,
      "grad_norm": 8.865078926086426,
      "learning_rate": 2.855917667238422e-05,
      "loss": 0.4545,
      "step": 2500
    },
    {
      "epoch": 2.144082332761578,
      "eval_accuracy": 0.8828828828828829,
      "eval_loss": 0.3031620681285858,
      "eval_runtime": 14.3961,
      "eval_samples_per_second": 161.919,
      "eval_steps_per_second": 20.283,
      "step": 2500
    },
    {
      "epoch": 2.1869639794168094,
      "grad_norm": 7.019817352294922,
      "learning_rate": 2.8130360205831908e-05,
      "loss": 0.4766,
      "step": 2550
    },
    {
      "epoch": 2.2298456260720414,
      "grad_norm": 4.404627799987793,
      "learning_rate": 2.770154373927959e-05,
      "loss": 0.4345,
      "step": 2600
    },
    {
      "epoch": 2.2298456260720414,
      "eval_accuracy": 0.8923208923208923,
      "eval_loss": 0.2739025354385376,
      "eval_runtime": 14.9267,
      "eval_samples_per_second": 156.163,
      "eval_steps_per_second": 19.562,
      "step": 2600
    },
    {
      "epoch": 2.2727272727272725,
      "grad_norm": 4.765737056732178,
      "learning_rate": 2.7272727272727273e-05,
      "loss": 0.424,
      "step": 2650
    },
    {
      "epoch": 2.3156089193825045,
      "grad_norm": 31.984413146972656,
      "learning_rate": 2.684391080617496e-05,
      "loss": 0.4697,
      "step": 2700
    },
    {
      "epoch": 2.3156089193825045,
      "eval_accuracy": 0.8931788931788932,
      "eval_loss": 0.2758677005767822,
      "eval_runtime": 14.2259,
      "eval_samples_per_second": 163.856,
      "eval_steps_per_second": 20.526,
      "step": 2700
    },
    {
      "epoch": 2.358490566037736,
      "grad_norm": 5.218204498291016,
      "learning_rate": 2.641509433962264e-05,
      "loss": 0.4927,
      "step": 2750
    },
    {
      "epoch": 2.4013722126929675,
      "grad_norm": 19.66315269470215,
      "learning_rate": 2.5986277873070326e-05,
      "loss": 0.4313,
      "step": 2800
    },
    {
      "epoch": 2.4013722126929675,
      "eval_accuracy": 0.9051909051909052,
      "eval_loss": 0.25858309864997864,
      "eval_runtime": 14.6303,
      "eval_samples_per_second": 159.327,
      "eval_steps_per_second": 19.959,
      "step": 2800
    },
    {
      "epoch": 2.444253859348199,
      "grad_norm": 1.5584404468536377,
      "learning_rate": 2.5557461406518014e-05,
      "loss": 0.4136,
      "step": 2850
    },
    {
      "epoch": 2.4871355060034306,
      "grad_norm": 6.387534141540527,
      "learning_rate": 2.5128644939965695e-05,
      "loss": 0.4215,
      "step": 2900
    },
    {
      "epoch": 2.4871355060034306,
      "eval_accuracy": 0.9004719004719005,
      "eval_loss": 0.2652466893196106,
      "eval_runtime": 14.1925,
      "eval_samples_per_second": 164.242,
      "eval_steps_per_second": 20.574,
      "step": 2900
    },
    {
      "epoch": 2.530017152658662,
      "grad_norm": 24.956953048706055,
      "learning_rate": 2.4699828473413382e-05,
      "loss": 0.4174,
      "step": 2950
    },
    {
      "epoch": 2.5728987993138936,
      "grad_norm": 4.991686820983887,
      "learning_rate": 2.4271012006861067e-05,
      "loss": 0.4504,
      "step": 3000
    },
    {
      "epoch": 2.5728987993138936,
      "eval_accuracy": 0.8927498927498928,
      "eval_loss": 0.2829649746417999,
      "eval_runtime": 13.721,
      "eval_samples_per_second": 169.885,
      "eval_steps_per_second": 21.281,
      "step": 3000
    },
    {
      "epoch": 2.615780445969125,
      "grad_norm": 8.409246444702148,
      "learning_rate": 2.3842195540308747e-05,
      "loss": 0.433,
      "step": 3050
    },
    {
      "epoch": 2.6586620926243567,
      "grad_norm": 20.707319259643555,
      "learning_rate": 2.3413379073756435e-05,
      "loss": 0.4218,
      "step": 3100
    },
    {
      "epoch": 2.6586620926243567,
      "eval_accuracy": 0.8936078936078936,
      "eval_loss": 0.2900983989238739,
      "eval_runtime": 13.1413,
      "eval_samples_per_second": 177.38,
      "eval_steps_per_second": 22.22,
      "step": 3100
    },
    {
      "epoch": 2.7015437392795882,
      "grad_norm": 6.371370792388916,
      "learning_rate": 2.298456260720412e-05,
      "loss": 0.454,
      "step": 3150
    },
    {
      "epoch": 2.7444253859348198,
      "grad_norm": 2.8135769367218018,
      "learning_rate": 2.25557461406518e-05,
      "loss": 0.4889,
      "step": 3200
    },
    {
      "epoch": 2.7444253859348198,
      "eval_accuracy": 0.8910338910338911,
      "eval_loss": 0.2830560803413391,
      "eval_runtime": 13.2735,
      "eval_samples_per_second": 175.613,
      "eval_steps_per_second": 21.999,
      "step": 3200
    },
    {
      "epoch": 2.7873070325900513,
      "grad_norm": 4.3516058921813965,
      "learning_rate": 2.2126929674099488e-05,
      "loss": 0.5019,
      "step": 3250
    },
    {
      "epoch": 2.830188679245283,
      "grad_norm": 10.607779502868652,
      "learning_rate": 2.1698113207547172e-05,
      "loss": 0.47,
      "step": 3300
    },
    {
      "epoch": 2.830188679245283,
      "eval_accuracy": 0.8927498927498928,
      "eval_loss": 0.2812536060810089,
      "eval_runtime": 13.5257,
      "eval_samples_per_second": 172.338,
      "eval_steps_per_second": 21.589,
      "step": 3300
    },
    {
      "epoch": 2.873070325900515,
      "grad_norm": 3.962139844894409,
      "learning_rate": 2.1269296740994853e-05,
      "loss": 0.3699,
      "step": 3350
    },
    {
      "epoch": 2.915951972555746,
      "grad_norm": 10.8749361038208,
      "learning_rate": 2.084048027444254e-05,
      "loss": 0.4758,
      "step": 3400
    },
    {
      "epoch": 2.915951972555746,
      "eval_accuracy": 0.8944658944658944,
      "eval_loss": 0.2788408696651459,
      "eval_runtime": 13.5439,
      "eval_samples_per_second": 172.107,
      "eval_steps_per_second": 21.56,
      "step": 3400
    },
    {
      "epoch": 2.958833619210978,
      "grad_norm": 20.483320236206055,
      "learning_rate": 2.0411663807890225e-05,
      "loss": 0.4581,
      "step": 3450
    },
    {
      "epoch": 3.0017152658662094,
      "grad_norm": 12.251073837280273,
      "learning_rate": 1.998284734133791e-05,
      "loss": 0.4482,
      "step": 3500
    },
    {
      "epoch": 3.0017152658662094,
      "eval_accuracy": 0.8953238953238953,
      "eval_loss": 0.26482564210891724,
      "eval_runtime": 13.5212,
      "eval_samples_per_second": 172.396,
      "eval_steps_per_second": 21.596,
      "step": 3500
    },
    {
      "epoch": 3.044596912521441,
      "grad_norm": 6.609012603759766,
      "learning_rate": 1.9554030874785594e-05,
      "loss": 0.4557,
      "step": 3550
    },
    {
      "epoch": 3.0874785591766725,
      "grad_norm": 31.05156898498535,
      "learning_rate": 1.9125214408233278e-05,
      "loss": 0.3735,
      "step": 3600
    },
    {
      "epoch": 3.0874785591766725,
      "eval_accuracy": 0.8987558987558988,
      "eval_loss": 0.25668570399284363,
      "eval_runtime": 13.5267,
      "eval_samples_per_second": 172.325,
      "eval_steps_per_second": 21.587,
      "step": 3600
    },
    {
      "epoch": 3.130360205831904,
      "grad_norm": 33.095829010009766,
      "learning_rate": 1.8696397941680962e-05,
      "loss": 0.3992,
      "step": 3650
    },
    {
      "epoch": 3.1732418524871355,
      "grad_norm": 10.586121559143066,
      "learning_rate": 1.8267581475128647e-05,
      "loss": 0.4338,
      "step": 3700
    },
    {
      "epoch": 3.1732418524871355,
      "eval_accuracy": 0.9086229086229086,
      "eval_loss": 0.2435009777545929,
      "eval_runtime": 13.3274,
      "eval_samples_per_second": 174.903,
      "eval_steps_per_second": 21.91,
      "step": 3700
    },
    {
      "epoch": 3.216123499142367,
      "grad_norm": 2.1484551429748535,
      "learning_rate": 1.783876500857633e-05,
      "loss": 0.4019,
      "step": 3750
    },
    {
      "epoch": 3.2590051457975986,
      "grad_norm": 1.3401719331741333,
      "learning_rate": 1.7409948542024015e-05,
      "loss": 0.3423,
      "step": 3800
    },
    {
      "epoch": 3.2590051457975986,
      "eval_accuracy": 0.9021879021879022,
      "eval_loss": 0.25972625613212585,
      "eval_runtime": 14.1124,
      "eval_samples_per_second": 165.174,
      "eval_steps_per_second": 20.691,
      "step": 3800
    },
    {
      "epoch": 3.30188679245283,
      "grad_norm": 2.6398582458496094,
      "learning_rate": 1.69811320754717e-05,
      "loss": 0.3813,
      "step": 3850
    },
    {
      "epoch": 3.3447684391080617,
      "grad_norm": 8.979487419128418,
      "learning_rate": 1.6552315608919384e-05,
      "loss": 0.3689,
      "step": 3900
    },
    {
      "epoch": 3.3447684391080617,
      "eval_accuracy": 0.9077649077649078,
      "eval_loss": 0.261087566614151,
      "eval_runtime": 13.587,
      "eval_samples_per_second": 171.561,
      "eval_steps_per_second": 21.491,
      "step": 3900
    },
    {
      "epoch": 3.387650085763293,
      "grad_norm": 4.436942100524902,
      "learning_rate": 1.6123499142367068e-05,
      "loss": 0.4099,
      "step": 3950
    },
    {
      "epoch": 3.4305317324185247,
      "grad_norm": 4.749749660491943,
      "learning_rate": 1.5694682675814752e-05,
      "loss": 0.347,
      "step": 4000
    },
    {
      "epoch": 3.4305317324185247,
      "eval_accuracy": 0.9103389103389103,
      "eval_loss": 0.24434834718704224,
      "eval_runtime": 13.493,
      "eval_samples_per_second": 172.757,
      "eval_steps_per_second": 21.641,
      "step": 4000
    },
    {
      "epoch": 3.4734133790737562,
      "grad_norm": 62.90760040283203,
      "learning_rate": 1.5265866209262437e-05,
      "loss": 0.3681,
      "step": 4050
    },
    {
      "epoch": 3.516295025728988,
      "grad_norm": 29.556215286254883,
      "learning_rate": 1.4837049742710121e-05,
      "loss": 0.2854,
      "step": 4100
    },
    {
      "epoch": 3.516295025728988,
      "eval_accuracy": 0.9051909051909052,
      "eval_loss": 0.25928205251693726,
      "eval_runtime": 13.3683,
      "eval_samples_per_second": 174.368,
      "eval_steps_per_second": 21.843,
      "step": 4100
    },
    {
      "epoch": 3.5591766723842193,
      "grad_norm": 7.2456207275390625,
      "learning_rate": 1.4408233276157807e-05,
      "loss": 0.4479,
      "step": 4150
    },
    {
      "epoch": 3.6020583190394513,
      "grad_norm": 9.010571479797363,
      "learning_rate": 1.397941680960549e-05,
      "loss": 0.3513,
      "step": 4200
    },
    {
      "epoch": 3.6020583190394513,
      "eval_accuracy": 0.9043329043329044,
      "eval_loss": 0.2614918053150177,
      "eval_runtime": 13.6533,
      "eval_samples_per_second": 170.728,
      "eval_steps_per_second": 21.387,
      "step": 4200
    },
    {
      "epoch": 3.644939965694683,
      "grad_norm": 8.282753944396973,
      "learning_rate": 1.3550600343053174e-05,
      "loss": 0.3045,
      "step": 4250
    },
    {
      "epoch": 3.6878216123499143,
      "grad_norm": 8.212636947631836,
      "learning_rate": 1.312178387650086e-05,
      "loss": 0.4081,
      "step": 4300
    },
    {
      "epoch": 3.6878216123499143,
      "eval_accuracy": 0.9073359073359073,
      "eval_loss": 0.25328412652015686,
      "eval_runtime": 13.7186,
      "eval_samples_per_second": 169.915,
      "eval_steps_per_second": 21.285,
      "step": 4300
    },
    {
      "epoch": 3.730703259005146,
      "grad_norm": 1.7500250339508057,
      "learning_rate": 1.2692967409948542e-05,
      "loss": 0.3552,
      "step": 4350
    },
    {
      "epoch": 3.7735849056603774,
      "grad_norm": 38.88422393798828,
      "learning_rate": 1.2264150943396227e-05,
      "loss": 0.3346,
      "step": 4400
    },
    {
      "epoch": 3.7735849056603774,
      "eval_accuracy": 0.9021879021879022,
      "eval_loss": 0.27931204438209534,
      "eval_runtime": 15.0366,
      "eval_samples_per_second": 155.022,
      "eval_steps_per_second": 19.419,
      "step": 4400
    },
    {
      "epoch": 3.816466552315609,
      "grad_norm": 18.03175163269043,
      "learning_rate": 1.1835334476843911e-05,
      "loss": 0.3502,
      "step": 4450
    },
    {
      "epoch": 3.8593481989708405,
      "grad_norm": 6.491177082061768,
      "learning_rate": 1.1406518010291597e-05,
      "loss": 0.3495,
      "step": 4500
    },
    {
      "epoch": 3.8593481989708405,
      "eval_accuracy": 0.9056199056199056,
      "eval_loss": 0.2461409866809845,
      "eval_runtime": 14.6288,
      "eval_samples_per_second": 159.344,
      "eval_steps_per_second": 19.961,
      "step": 4500
    },
    {
      "epoch": 3.902229845626072,
      "grad_norm": 22.98536491394043,
      "learning_rate": 1.097770154373928e-05,
      "loss": 0.3662,
      "step": 4550
    },
    {
      "epoch": 3.9451114922813035,
      "grad_norm": 24.458087921142578,
      "learning_rate": 1.0548885077186965e-05,
      "loss": 0.3979,
      "step": 4600
    },
    {
      "epoch": 3.9451114922813035,
      "eval_accuracy": 0.9103389103389103,
      "eval_loss": 0.24052156507968903,
      "eval_runtime": 14.1187,
      "eval_samples_per_second": 165.1,
      "eval_steps_per_second": 20.682,
      "step": 4600
    },
    {
      "epoch": 3.987993138936535,
      "grad_norm": 22.207815170288086,
      "learning_rate": 1.012006861063465e-05,
      "loss": 0.379,
      "step": 4650
    },
    {
      "epoch": 4.030874785591767,
      "grad_norm": 19.702865600585938,
      "learning_rate": 9.691252144082332e-06,
      "loss": 0.3441,
      "step": 4700
    },
    {
      "epoch": 4.030874785591767,
      "eval_accuracy": 0.9017589017589017,
      "eval_loss": 0.27815285325050354,
      "eval_runtime": 14.2114,
      "eval_samples_per_second": 164.024,
      "eval_steps_per_second": 20.547,
      "step": 4700
    },
    {
      "epoch": 4.073756432246999,
      "grad_norm": 9.863823890686035,
      "learning_rate": 9.262435677530018e-06,
      "loss": 0.4128,
      "step": 4750
    },
    {
      "epoch": 4.11663807890223,
      "grad_norm": 13.863909721374512,
      "learning_rate": 8.833619210977703e-06,
      "loss": 0.3357,
      "step": 4800
    },
    {
      "epoch": 4.11663807890223,
      "eval_accuracy": 0.9094809094809094,
      "eval_loss": 0.26747584342956543,
      "eval_runtime": 13.9108,
      "eval_samples_per_second": 167.568,
      "eval_steps_per_second": 20.991,
      "step": 4800
    },
    {
      "epoch": 4.159519725557462,
      "grad_norm": 45.47288131713867,
      "learning_rate": 8.404802744425385e-06,
      "loss": 0.3301,
      "step": 4850
    },
    {
      "epoch": 4.202401372212693,
      "grad_norm": 4.5591864585876465,
      "learning_rate": 7.975986277873071e-06,
      "loss": 0.2867,
      "step": 4900
    },
    {
      "epoch": 4.202401372212693,
      "eval_accuracy": 0.9060489060489061,
      "eval_loss": 0.27105966210365295,
      "eval_runtime": 15.175,
      "eval_samples_per_second": 153.608,
      "eval_steps_per_second": 19.242,
      "step": 4900
    },
    {
      "epoch": 4.245283018867925,
      "grad_norm": 0.21429812908172607,
      "learning_rate": 7.547169811320755e-06,
      "loss": 0.3518,
      "step": 4950
    },
    {
      "epoch": 4.288164665523156,
      "grad_norm": 4.873584747314453,
      "learning_rate": 7.11835334476844e-06,
      "loss": 0.3084,
      "step": 5000
    },
    {
      "epoch": 4.288164665523156,
      "eval_accuracy": 0.9116259116259117,
      "eval_loss": 0.24336792528629303,
      "eval_runtime": 13.447,
      "eval_samples_per_second": 173.347,
      "eval_steps_per_second": 21.715,
      "step": 5000
    },
    {
      "epoch": 4.331046312178388,
      "grad_norm": 3.682224988937378,
      "learning_rate": 6.689536878216124e-06,
      "loss": 0.3335,
      "step": 5050
    },
    {
      "epoch": 4.373927958833619,
      "grad_norm": 8.78370475769043,
      "learning_rate": 6.2607204116638075e-06,
      "loss": 0.4865,
      "step": 5100
    },
    {
      "epoch": 4.373927958833619,
      "eval_accuracy": 0.9154869154869155,
      "eval_loss": 0.2427978515625,
      "eval_runtime": 13.4143,
      "eval_samples_per_second": 173.77,
      "eval_steps_per_second": 21.768,
      "step": 5100
    },
    {
      "epoch": 4.416809605488851,
      "grad_norm": 5.311465263366699,
      "learning_rate": 5.831903945111493e-06,
      "loss": 0.3465,
      "step": 5150
    },
    {
      "epoch": 4.459691252144083,
      "grad_norm": 9.669562339782715,
      "learning_rate": 5.403087478559177e-06,
      "loss": 0.3557,
      "step": 5200
    },
    {
      "epoch": 4.459691252144083,
      "eval_accuracy": 0.9141999141999142,
      "eval_loss": 0.2416524589061737,
      "eval_runtime": 13.3034,
      "eval_samples_per_second": 175.218,
      "eval_steps_per_second": 21.949,
      "step": 5200
    },
    {
      "epoch": 4.502572898799314,
      "grad_norm": 8.7540283203125,
      "learning_rate": 4.974271012006861e-06,
      "loss": 0.3254,
      "step": 5250
    },
    {
      "epoch": 4.545454545454545,
      "grad_norm": 0.46742549538612366,
      "learning_rate": 4.5454545454545455e-06,
      "loss": 0.3384,
      "step": 5300
    },
    {
      "epoch": 4.545454545454545,
      "eval_accuracy": 0.9133419133419134,
      "eval_loss": 0.24515940248966217,
      "eval_runtime": 13.6233,
      "eval_samples_per_second": 171.103,
      "eval_steps_per_second": 21.434,
      "step": 5300
    },
    {
      "epoch": 4.588336192109777,
      "grad_norm": 23.387781143188477,
      "learning_rate": 4.116638078902231e-06,
      "loss": 0.3508,
      "step": 5350
    },
    {
      "epoch": 4.631217838765009,
      "grad_norm": 3.3974390029907227,
      "learning_rate": 3.687821612349914e-06,
      "loss": 0.3147,
      "step": 5400
    },
    {
      "epoch": 4.631217838765009,
      "eval_accuracy": 0.9141999141999142,
      "eval_loss": 0.24695377051830292,
      "eval_runtime": 13.4806,
      "eval_samples_per_second": 172.915,
      "eval_steps_per_second": 21.661,
      "step": 5400
    },
    {
      "epoch": 4.67409948542024,
      "grad_norm": 26.67882537841797,
      "learning_rate": 3.259005145797599e-06,
      "loss": 0.3364,
      "step": 5450
    },
    {
      "epoch": 4.716981132075472,
      "grad_norm": 0.8814979791641235,
      "learning_rate": 2.830188679245283e-06,
      "loss": 0.2858,
      "step": 5500
    },
    {
      "epoch": 4.716981132075472,
      "eval_accuracy": 0.9103389103389103,
      "eval_loss": 0.24419519305229187,
      "eval_runtime": 13.9344,
      "eval_samples_per_second": 167.284,
      "eval_steps_per_second": 20.955,
      "step": 5500
    },
    {
      "epoch": 4.759862778730703,
      "grad_norm": 13.250667572021484,
      "learning_rate": 2.4013722126929674e-06,
      "loss": 0.3674,
      "step": 5550
    },
    {
      "epoch": 4.802744425385935,
      "grad_norm": 11.3861722946167,
      "learning_rate": 1.972555746140652e-06,
      "loss": 0.3664,
      "step": 5600
    },
    {
      "epoch": 4.802744425385935,
      "eval_accuracy": 0.9064779064779065,
      "eval_loss": 0.2746657431125641,
      "eval_runtime": 13.6304,
      "eval_samples_per_second": 171.015,
      "eval_steps_per_second": 21.423,
      "step": 5600
    },
    {
      "epoch": 4.845626072041166,
      "grad_norm": 3.5501041412353516,
      "learning_rate": 1.5437392795883364e-06,
      "loss": 0.3152,
      "step": 5650
    },
    {
      "epoch": 4.888507718696398,
      "grad_norm": 24.59325408935547,
      "learning_rate": 1.1149228130360207e-06,
      "loss": 0.3549,
      "step": 5700
    },
    {
      "epoch": 4.888507718696398,
      "eval_accuracy": 0.9099099099099099,
      "eval_loss": 0.26241084933280945,
      "eval_runtime": 15.5182,
      "eval_samples_per_second": 150.211,
      "eval_steps_per_second": 18.817,
      "step": 5700
    },
    {
      "epoch": 4.931389365351629,
      "grad_norm": 32.23326110839844,
      "learning_rate": 6.86106346483705e-07,
      "loss": 0.3796,
      "step": 5750
    },
    {
      "epoch": 4.974271012006861,
      "grad_norm": 30.20063591003418,
      "learning_rate": 2.572898799313894e-07,
      "loss": 0.3656,
      "step": 5800
    },
    {
      "epoch": 4.974271012006861,
      "eval_accuracy": 0.9047619047619048,
      "eval_loss": 0.27714791893959045,
      "eval_runtime": 14.0154,
      "eval_samples_per_second": 166.317,
      "eval_steps_per_second": 20.834,
      "step": 5800
    }
  ],
  "logging_steps": 50,
  "max_steps": 5830,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 5,
  "save_steps": 100,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 2.7169922825846784e+17,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}