{
  "best_metric": 0.2525734305381775,
  "best_model_checkpoint": "./convnext-tiny-upgrade-384-batch-16/checkpoint-10990",
  "epoch": 10.0,
  "eval_steps": 500,
  "global_step": 10990,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.09,
      "grad_norm": 39.919612884521484,
      "learning_rate": 2.999387175598269e-05,
      "loss": 2.9864,
      "step": 100
    },
    {
      "epoch": 0.18,
      "grad_norm": 24.833358764648438,
      "learning_rate": 2.9975492031314045e-05,
      "loss": 2.0393,
      "step": 200
    },
    {
      "epoch": 0.27,
      "grad_norm": 26.59770965576172,
      "learning_rate": 2.994487584405244e-05,
      "loss": 1.5895,
      "step": 300
    },
    {
      "epoch": 0.36,
      "grad_norm": 49.16465759277344,
      "learning_rate": 2.990204821066006e-05,
      "loss": 1.232,
      "step": 400
    },
    {
      "epoch": 0.45,
      "grad_norm": 30.683462142944336,
      "learning_rate": 2.984704412556199e-05,
      "loss": 1.1329,
      "step": 500
    },
    {
      "epoch": 0.55,
      "grad_norm": 31.109006881713867,
      "learning_rate": 2.977990853255228e-05,
      "loss": 1.003,
      "step": 600
    },
    {
      "epoch": 0.64,
      "grad_norm": 30.210790634155273,
      "learning_rate": 2.970069628807043e-05,
      "loss": 0.9696,
      "step": 700
    },
    {
      "epoch": 0.73,
      "grad_norm": 29.36758041381836,
      "learning_rate": 2.9609472116378222e-05,
      "loss": 0.9048,
      "step": 800
    },
    {
      "epoch": 0.82,
      "grad_norm": 39.95523452758789,
      "learning_rate": 2.9506310556673573e-05,
      "loss": 0.8202,
      "step": 900
    },
    {
      "epoch": 0.91,
      "grad_norm": 25.76544761657715,
      "learning_rate": 2.9391295902184625e-05,
      "loss": 0.855,
      "step": 1000
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.8576540755467197,
      "eval_loss": 0.49257978796958923,
      "eval_runtime": 68.5947,
      "eval_samples_per_second": 36.665,
      "eval_steps_per_second": 2.303,
      "step": 1099
    },
    {
      "epoch": 1.0,
      "grad_norm": 39.65526580810547,
      "learning_rate": 2.9264522131293818e-05,
      "loss": 0.8411,
      "step": 1100
    },
    {
      "epoch": 1.09,
      "grad_norm": 33.58937454223633,
      "learning_rate": 2.9126092830748217e-05,
      "loss": 0.7447,
      "step": 1200
    },
    {
      "epoch": 1.18,
      "grad_norm": 32.78460693359375,
      "learning_rate": 2.897612111101888e-05,
      "loss": 0.7118,
      "step": 1300
    },
    {
      "epoch": 1.27,
      "grad_norm": 23.4544734954834,
      "learning_rate": 2.8814729513878365e-05,
      "loss": 0.6763,
      "step": 1400
    },
    {
      "epoch": 1.36,
      "grad_norm": 22.840782165527344,
      "learning_rate": 2.864204991227195e-05,
      "loss": 0.6592,
      "step": 1500
    },
    {
      "epoch": 1.46,
      "grad_norm": 30.469961166381836,
      "learning_rate": 2.8458223402564366e-05,
      "loss": 0.6426,
      "step": 1600
    },
    {
      "epoch": 1.55,
      "grad_norm": 22.303668975830078,
      "learning_rate": 2.826340018925006e-05,
      "loss": 0.6948,
      "step": 1700
    },
    {
      "epoch": 1.64,
      "grad_norm": 31.617197036743164,
      "learning_rate": 2.8057739462221215e-05,
      "loss": 0.6894,
      "step": 1800
    },
    {
      "epoch": 1.73,
      "grad_norm": 29.403879165649414,
      "learning_rate": 2.7841409266693838e-05,
      "loss": 0.584,
      "step": 1900
    },
    {
      "epoch": 1.82,
      "grad_norm": 20.05999183654785,
      "learning_rate": 2.761458636589813e-05,
      "loss": 0.5967,
      "step": 2000
    },
    {
      "epoch": 1.91,
      "grad_norm": 30.453641891479492,
      "learning_rate": 2.7377456096645395e-05,
      "loss": 0.6243,
      "step": 2100
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.8910536779324055,
      "eval_loss": 0.3875384032726288,
      "eval_runtime": 68.6682,
      "eval_samples_per_second": 36.625,
      "eval_steps_per_second": 2.301,
      "step": 2198
    },
    {
      "epoch": 2.0,
      "grad_norm": 22.794034957885742,
      "learning_rate": 2.7130212217889484e-05,
      "loss": 0.5952,
      "step": 2200
    },
    {
      "epoch": 2.09,
      "grad_norm": 21.936601638793945,
      "learning_rate": 2.6873056752406504e-05,
      "loss": 0.53,
      "step": 2300
    },
    {
      "epoch": 2.18,
      "grad_norm": 26.926753997802734,
      "learning_rate": 2.6606199821722166e-05,
      "loss": 0.5591,
      "step": 2400
    },
    {
      "epoch": 2.27,
      "grad_norm": 23.042003631591797,
      "learning_rate": 2.632985947442167e-05,
      "loss": 0.5301,
      "step": 2500
    },
    {
      "epoch": 2.37,
      "grad_norm": 19.099658966064453,
      "learning_rate": 2.6044261507982356e-05,
      "loss": 0.5852,
      "step": 2600
    },
    {
      "epoch": 2.46,
      "grad_norm": 33.43634796142578,
      "learning_rate": 2.5749639284274782e-05,
      "loss": 0.4954,
      "step": 2700
    },
    {
      "epoch": 2.55,
      "grad_norm": 20.47861099243164,
      "learning_rate": 2.5446233538882924e-05,
      "loss": 0.5428,
      "step": 2800
    },
    {
      "epoch": 2.64,
      "grad_norm": 45.73374938964844,
      "learning_rate": 2.513429218439932e-05,
      "loss": 0.5164,
      "step": 2900
    },
    {
      "epoch": 2.73,
      "grad_norm": 17.044771194458008,
      "learning_rate": 2.4814070107855878e-05,
      "loss": 0.5257,
      "step": 3000
    },
    {
      "epoch": 2.82,
      "grad_norm": 23.656314849853516,
      "learning_rate": 2.448582896245591e-05,
      "loss": 0.5295,
      "step": 3100
    },
    {
      "epoch": 2.91,
      "grad_norm": 59.598411560058594,
      "learning_rate": 2.4149836953777488e-05,
      "loss": 0.4776,
      "step": 3200
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.9125248508946322,
      "eval_loss": 0.3229917287826538,
      "eval_runtime": 69.7961,
      "eval_samples_per_second": 36.034,
      "eval_steps_per_second": 2.264,
      "step": 3297
    },
    {
      "epoch": 3.0,
      "grad_norm": 14.091382026672363,
      "learning_rate": 2.3806368620622876e-05,
      "loss": 0.5338,
      "step": 3300
    },
    {
      "epoch": 3.09,
      "grad_norm": 34.719451904296875,
      "learning_rate": 2.345570461069312e-05,
      "loss": 0.4399,
      "step": 3400
    },
    {
      "epoch": 3.18,
      "grad_norm": 29.300067901611328,
      "learning_rate": 2.3098131451271016e-05,
      "loss": 0.4671,
      "step": 3500
    },
    {
      "epoch": 3.28,
      "grad_norm": 23.41453742980957,
      "learning_rate": 2.2733941315099883e-05,
      "loss": 0.4643,
      "step": 3600
    },
    {
      "epoch": 3.37,
      "grad_norm": 18.566978454589844,
      "learning_rate": 2.2363431781649483e-05,
      "loss": 0.4193,
      "step": 3700
    },
    {
      "epoch": 3.46,
      "grad_norm": 25.832847595214844,
      "learning_rate": 2.1986905593964048e-05,
      "loss": 0.4921,
      "step": 3800
    },
    {
      "epoch": 3.55,
      "grad_norm": 25.94115447998047,
      "learning_rate": 2.1604670411291174e-05,
      "loss": 0.4778,
      "step": 3900
    },
    {
      "epoch": 3.64,
      "grad_norm": 20.953384399414062,
      "learning_rate": 2.121703855769373e-05,
      "loss": 0.4823,
      "step": 4000
    },
    {
      "epoch": 3.73,
      "grad_norm": 20.484058380126953,
      "learning_rate": 2.0824326766850072e-05,
      "loss": 0.4804,
      "step": 4100
    },
    {
      "epoch": 3.82,
      "grad_norm": 6.4666290283203125,
      "learning_rate": 2.042685592325123e-05,
      "loss": 0.4492,
      "step": 4200
    },
    {
      "epoch": 3.91,
      "grad_norm": 36.683509826660156,
      "learning_rate": 2.0024950800006463e-05,
      "loss": 0.4535,
      "step": 4300
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.9204771371769384,
      "eval_loss": 0.2854064702987671,
      "eval_runtime": 68.5965,
      "eval_samples_per_second": 36.664,
      "eval_steps_per_second": 2.303,
      "step": 4396
    },
    {
      "epoch": 4.0,
      "grad_norm": 8.934366226196289,
      "learning_rate": 1.961893979347137e-05,
      "loss": 0.3936,
      "step": 4400
    },
    {
      "epoch": 4.09,
      "grad_norm": 19.066776275634766,
      "learning_rate": 1.9209154654915524e-05,
      "loss": 0.4025,
      "step": 4500
    },
    {
      "epoch": 4.19,
      "grad_norm": 35.145851135253906,
      "learning_rate": 1.879593021944875e-05,
      "loss": 0.4269,
      "step": 4600
    },
    {
      "epoch": 4.28,
      "grad_norm": 53.010009765625,
      "learning_rate": 1.837960413242765e-05,
      "loss": 0.4041,
      "step": 4700
    },
    {
      "epoch": 4.37,
      "grad_norm": 20.637130737304688,
      "learning_rate": 1.796051657356582e-05,
      "loss": 0.3865,
      "step": 4800
    },
    {
      "epoch": 4.46,
      "grad_norm": 44.37509536743164,
      "learning_rate": 1.7539009978973312e-05,
      "loss": 0.3781,
      "step": 4900
    },
    {
      "epoch": 4.55,
      "grad_norm": 26.981489181518555,
      "learning_rate": 1.711542876135233e-05,
      "loss": 0.3876,
      "step": 5000
    },
    {
      "epoch": 4.64,
      "grad_norm": 36.75742721557617,
      "learning_rate": 1.669011902857791e-05,
      "loss": 0.3919,
      "step": 5100
    },
    {
      "epoch": 4.73,
      "grad_norm": 15.431539535522461,
      "learning_rate": 1.6263428300893422e-05,
      "loss": 0.4433,
      "step": 5200
    },
    {
      "epoch": 4.82,
      "grad_norm": 26.744518280029297,
      "learning_rate": 1.5835705226952112e-05,
      "loss": 0.4334,
      "step": 5300
    },
    {
      "epoch": 4.91,
      "grad_norm": 7.719830513000488,
      "learning_rate": 1.540729929893649e-05,
      "loss": 0.4204,
      "step": 5400
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.9168986083499006,
      "eval_loss": 0.29153481125831604,
      "eval_runtime": 68.6941,
      "eval_samples_per_second": 36.612,
      "eval_steps_per_second": 2.3,
      "step": 5495
    },
    {
      "epoch": 5.0,
      "grad_norm": 18.945358276367188,
      "learning_rate": 1.4978560566988603e-05,
      "loss": 0.4077,
      "step": 5500
    },
    {
      "epoch": 5.1,
      "grad_norm": 32.922447204589844,
      "learning_rate": 1.454983935318433e-05,
      "loss": 0.356,
      "step": 5600
    },
    {
      "epoch": 5.19,
      "grad_norm": 23.755434036254883,
      "learning_rate": 1.4121485965285485e-05,
      "loss": 0.3569,
      "step": 5700
    },
    {
      "epoch": 5.28,
      "grad_norm": 23.00724983215332,
      "learning_rate": 1.3693850410503614e-05,
      "loss": 0.379,
      "step": 5800
    },
    {
      "epoch": 5.37,
      "grad_norm": 15.597938537597656,
      "learning_rate": 1.326728210950942e-05,
      "loss": 0.3608,
      "step": 5900
    },
    {
      "epoch": 5.46,
      "grad_norm": 16.457727432250977,
      "learning_rate": 1.2842129610921378e-05,
      "loss": 0.3613,
      "step": 6000
    },
    {
      "epoch": 5.55,
      "grad_norm": 20.524658203125,
      "learning_rate": 1.2418740306506923e-05,
      "loss": 0.3631,
      "step": 6100
    },
    {
      "epoch": 5.64,
      "grad_norm": 31.883596420288086,
      "learning_rate": 1.1997460147328984e-05,
      "loss": 0.343,
      "step": 6200
    },
    {
      "epoch": 5.73,
      "grad_norm": 18.492212295532227,
      "learning_rate": 1.1578633361069559e-05,
      "loss": 0.3839,
      "step": 6300
    },
    {
      "epoch": 5.82,
      "grad_norm": 18.56053924560547,
      "learning_rate": 1.1162602170761611e-05,
      "loss": 0.3661,
      "step": 6400
    },
    {
      "epoch": 5.91,
      "grad_norm": 34.4693489074707,
      "learning_rate": 1.0749706515158863e-05,
      "loss": 0.3756,
      "step": 6500
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.9192842942345925,
      "eval_loss": 0.29144659638404846,
      "eval_runtime": 68.5475,
      "eval_samples_per_second": 36.69,
      "eval_steps_per_second": 2.305,
      "step": 6594
    },
    {
      "epoch": 6.01,
      "grad_norm": 18.55854034423828,
      "learning_rate": 1.0340283770972167e-05,
      "loss": 0.3684,
      "step": 6600
    },
    {
      "epoch": 6.1,
      "grad_norm": 23.47572135925293,
      "learning_rate": 9.93466847719919e-06,
      "loss": 0.3282,
      "step": 6700
    },
    {
      "epoch": 6.19,
      "grad_norm": 26.671112060546875,
      "learning_rate": 9.533192061772919e-06,
      "loss": 0.3611,
      "step": 6800
    },
    {
      "epoch": 6.28,
      "grad_norm": 7.420229434967041,
      "learning_rate": 9.136182570752153e-06,
      "loss": 0.3547,
      "step": 6900
    },
    {
      "epoch": 6.37,
      "grad_norm": 13.92644214630127,
      "learning_rate": 8.743964400275304e-06,
      "loss": 0.3029,
      "step": 7000
    },
    {
      "epoch": 6.46,
      "grad_norm": 12.337591171264648,
      "learning_rate": 8.356858031496596e-06,
      "loss": 0.3657,
      "step": 7100
    },
    {
      "epoch": 6.55,
      "grad_norm": 10.994024276733398,
      "learning_rate": 7.975179768721187e-06,
      "loss": 0.3179,
      "step": 7200
    },
    {
      "epoch": 6.64,
      "grad_norm": 26.956287384033203,
      "learning_rate": 7.599241480953112e-06,
      "loss": 0.3009,
      "step": 7300
    },
    {
      "epoch": 6.73,
      "grad_norm": 29.974740982055664,
      "learning_rate": 7.229350347067426e-06,
      "loss": 0.3297,
      "step": 7400
    },
    {
      "epoch": 6.82,
      "grad_norm": 11.850204467773438,
      "learning_rate": 6.865808604814564e-06,
      "loss": 0.3175,
      "step": 7500
    },
    {
      "epoch": 6.92,
      "grad_norm": 4.603498935699463,
      "learning_rate": 6.508913303862144e-06,
      "loss": 0.3603,
      "step": 7600
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.9272365805168986,
      "eval_loss": 0.26446533203125,
      "eval_runtime": 68.6296,
      "eval_samples_per_second": 36.646,
      "eval_steps_per_second": 2.302,
      "step": 7693
    },
    {
      "epoch": 7.01,
      "grad_norm": 24.59402084350586,
      "learning_rate": 6.1589560630758656e-06,
      "loss": 0.2715,
      "step": 7700
    },
    {
      "epoch": 7.1,
      "grad_norm": 37.25328063964844,
      "learning_rate": 5.8162228322380155e-06,
      "loss": 0.3105,
      "step": 7800
    },
    {
      "epoch": 7.19,
      "grad_norm": 17.36973762512207,
      "learning_rate": 5.480993658398129e-06,
      "loss": 0.3269,
      "step": 7900
    },
    {
      "epoch": 7.28,
      "grad_norm": 12.080437660217285,
      "learning_rate": 5.153542457046737e-06,
      "loss": 0.3391,
      "step": 8000
    },
    {
      "epoch": 7.37,
      "grad_norm": 32.013824462890625,
      "learning_rate": 4.834136788299248e-06,
      "loss": 0.3345,
      "step": 8100
    },
    {
      "epoch": 7.46,
      "grad_norm": 36.49650573730469,
      "learning_rate": 4.523037638272822e-06,
      "loss": 0.311,
      "step": 8200
    },
    {
      "epoch": 7.55,
      "grad_norm": 60.167484283447266,
      "learning_rate": 4.220499205834783e-06,
      "loss": 0.2984,
      "step": 8300
    },
    {
      "epoch": 7.64,
      "grad_norm": 22.172080993652344,
      "learning_rate": 3.926768694896931e-06,
      "loss": 0.3295,
      "step": 8400
    },
    {
      "epoch": 7.73,
      "grad_norm": 10.280725479125977,
      "learning_rate": 3.6420861124254607e-06,
      "loss": 0.2991,
      "step": 8500
    },
    {
      "epoch": 7.83,
      "grad_norm": 31.965059280395508,
      "learning_rate": 3.3666840723314145e-06,
      "loss": 0.3375,
      "step": 8600
    },
    {
      "epoch": 7.92,
      "grad_norm": 15.358006477355957,
      "learning_rate": 3.1007876054020724e-06,
      "loss": 0.2885,
      "step": 8700
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.9280318091451292,
      "eval_loss": 0.2599093019962311,
      "eval_runtime": 70.1401,
      "eval_samples_per_second": 35.857,
      "eval_steps_per_second": 2.253,
      "step": 8792
    },
    {
      "epoch": 8.01,
      "grad_norm": 18.399208068847656,
      "learning_rate": 2.8446139754284486e-06,
      "loss": 0.2717,
      "step": 8800
    },
    {
      "epoch": 8.1,
      "grad_norm": 16.944734573364258,
      "learning_rate": 2.5983725016792574e-06,
      "loss": 0.2986,
      "step": 8900
    },
    {
      "epoch": 8.19,
      "grad_norm": 18.830663681030273,
      "learning_rate": 2.36226438786627e-06,
      "loss": 0.2738,
      "step": 9000
    },
    {
      "epoch": 8.28,
      "grad_norm": 19.215288162231445,
      "learning_rate": 2.1364825577409424e-06,
      "loss": 0.2496,
      "step": 9100
    },
    {
      "epoch": 8.37,
      "grad_norm": 3.411406993865967,
      "learning_rate": 1.9212114974565664e-06,
      "loss": 0.2954,
      "step": 9200
    },
    {
      "epoch": 8.46,
      "grad_norm": 24.457778930664062,
      "learning_rate": 1.7166271048247796e-06,
      "loss": 0.2678,
      "step": 9300
    },
    {
      "epoch": 8.55,
      "grad_norm": 7.872967720031738,
      "learning_rate": 1.5228965455896054e-06,
      "loss": 0.2968,
      "step": 9400
    },
    {
      "epoch": 8.64,
      "grad_norm": 27.78696060180664,
      "learning_rate": 1.3401781168364591e-06,
      "loss": 0.3052,
      "step": 9500
    },
    {
      "epoch": 8.74,
      "grad_norm": 35.54745101928711,
      "learning_rate": 1.1686211176477208e-06,
      "loss": 0.2943,
      "step": 9600
    },
    {
      "epoch": 8.83,
      "grad_norm": 19.328489303588867,
      "learning_rate": 1.00836572711058e-06,
      "loss": 0.3274,
      "step": 9700
    },
    {
      "epoch": 8.92,
      "grad_norm": 35.09791564941406,
      "learning_rate": 8.595428897768071e-07,
      "loss": 0.2753,
      "step": 9800
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.9292246520874752,
      "eval_loss": 0.25652942061424255,
      "eval_runtime": 69.0906,
      "eval_samples_per_second": 36.401,
      "eval_steps_per_second": 2.287,
      "step": 9891
    },
    {
      "epoch": 9.01,
      "grad_norm": 34.884788513183594,
      "learning_rate": 7.222742086680756e-07,
      "loss": 0.2635,
      "step": 9900
    },
    {
      "epoch": 9.1,
      "grad_norm": 4.236999988555908,
      "learning_rate": 5.966718459142196e-07,
      "loss": 0.2781,
      "step": 10000
    },
    {
      "epoch": 9.19,
      "grad_norm": 22.74588966369629,
      "learning_rate": 4.82838431105655e-07,
      "loss": 0.2801,
      "step": 10100
    },
    {
      "epoch": 9.28,
      "grad_norm": 31.477947235107422,
      "learning_rate": 3.808669774348167e-07,
      "loss": 0.2562,
      "step": 10200
    },
    {
      "epoch": 9.37,
      "grad_norm": 26.103023529052734,
      "learning_rate": 2.908408056951578e-07,
      "loss": 0.3041,
      "step": 10300
    },
    {
      "epoch": 9.46,
      "grad_norm": 20.699371337890625,
      "learning_rate": 2.1283347619979243e-07,
      "loss": 0.2755,
      "step": 10400
    },
    {
      "epoch": 9.55,
      "grad_norm": 30.313541412353516,
      "learning_rate": 1.4690872867542892e-07,
      "loss": 0.3044,
      "step": 10500
    },
    {
      "epoch": 9.65,
      "grad_norm": 31.09404754638672,
      "learning_rate": 9.312043018067762e-08,
      "loss": 0.3142,
      "step": 10600
    },
    {
      "epoch": 9.74,
      "grad_norm": 23.177568435668945,
      "learning_rate": 5.1512531091333914e-08,
      "loss": 0.2853,
      "step": 10700
    },
    {
      "epoch": 9.83,
      "grad_norm": 45.65806198120117,
      "learning_rate": 2.211902918855313e-08,
      "loss": 0.3018,
      "step": 10800
    },
    {
      "epoch": 9.92,
      "grad_norm": 26.684669494628906,
      "learning_rate": 4.963941879295164e-09,
      "loss": 0.2902,
      "step": 10900
    },
    {
      "epoch": 10.0,
      "eval_accuracy": 0.9292246520874752,
      "eval_loss": 0.2525734305381775,
      "eval_runtime": 68.818,
      "eval_samples_per_second": 36.546,
      "eval_steps_per_second": 2.296,
      "step": 10990
    },
    {
      "epoch": 10.0,
      "step": 10990,
      "total_flos": 1.301428412334932e+19,
      "train_loss": 0.49425973189321837,
      "train_runtime": 10500.3736,
      "train_samples_per_second": 16.743,
      "train_steps_per_second": 1.047
    }
  ],
  "logging_steps": 100,
  "max_steps": 10990,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 10,
  "save_steps": 500,
  "total_flos": 1.301428412334932e+19,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": null
}