{
  "best_metric": 0.2026742845773697,
  "best_model_checkpoint": "./convext-base-more-aug/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": 24.41817283630371,
      "learning_rate": 2.999387175598269e-05,
      "loss": 2.811,
      "step": 100
    },
    {
      "epoch": 0.18,
      "grad_norm": 33.525028228759766,
      "learning_rate": 2.9975492031314045e-05,
      "loss": 1.7601,
      "step": 200
    },
    {
      "epoch": 0.27,
      "grad_norm": 34.92573165893555,
      "learning_rate": 2.994487584405244e-05,
      "loss": 1.3362,
      "step": 300
    },
    {
      "epoch": 0.36,
      "grad_norm": 67.0540771484375,
      "learning_rate": 2.990204821066006e-05,
      "loss": 1.051,
      "step": 400
    },
    {
      "epoch": 0.45,
      "grad_norm": 22.46187973022461,
      "learning_rate": 2.984704412556199e-05,
      "loss": 0.9415,
      "step": 500
    },
    {
      "epoch": 0.55,
      "grad_norm": 31.719919204711914,
      "learning_rate": 2.977990853255228e-05,
      "loss": 0.8432,
      "step": 600
    },
    {
      "epoch": 0.64,
      "grad_norm": 19.570117950439453,
      "learning_rate": 2.970069628807043e-05,
      "loss": 0.8992,
      "step": 700
    },
    {
      "epoch": 0.73,
      "grad_norm": 41.06416320800781,
      "learning_rate": 2.9609472116378222e-05,
      "loss": 0.8082,
      "step": 800
    },
    {
      "epoch": 0.82,
      "grad_norm": 38.35748291015625,
      "learning_rate": 2.9506310556673573e-05,
      "loss": 0.7754,
      "step": 900
    },
    {
      "epoch": 0.91,
      "grad_norm": 13.936028480529785,
      "learning_rate": 2.9391295902184625e-05,
      "loss": 0.7313,
      "step": 1000
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.8926441351888668,
      "eval_loss": 0.3851191997528076,
      "eval_runtime": 105.8528,
      "eval_samples_per_second": 23.759,
      "eval_steps_per_second": 1.493,
      "step": 1099
    },
    {
      "epoch": 1.0,
      "grad_norm": 33.56941223144531,
      "learning_rate": 2.9264522131293818e-05,
      "loss": 0.7834,
      "step": 1100
    },
    {
      "epoch": 1.09,
      "grad_norm": 39.84654235839844,
      "learning_rate": 2.9126092830748217e-05,
      "loss": 0.6719,
      "step": 1200
    },
    {
      "epoch": 1.18,
      "grad_norm": 39.59154510498047,
      "learning_rate": 2.897612111101888e-05,
      "loss": 0.6559,
      "step": 1300
    },
    {
      "epoch": 1.27,
      "grad_norm": 26.814891815185547,
      "learning_rate": 2.8814729513878365e-05,
      "loss": 0.6317,
      "step": 1400
    },
    {
      "epoch": 1.36,
      "grad_norm": 22.93006706237793,
      "learning_rate": 2.864204991227195e-05,
      "loss": 0.5876,
      "step": 1500
    },
    {
      "epoch": 1.46,
      "grad_norm": 33.48711013793945,
      "learning_rate": 2.8458223402564366e-05,
      "loss": 0.6032,
      "step": 1600
    },
    {
      "epoch": 1.55,
      "grad_norm": 31.577533721923828,
      "learning_rate": 2.826340018925006e-05,
      "loss": 0.6033,
      "step": 1700
    },
    {
      "epoch": 1.64,
      "grad_norm": 30.845195770263672,
      "learning_rate": 2.8057739462221215e-05,
      "loss": 0.6216,
      "step": 1800
    },
    {
      "epoch": 1.73,
      "grad_norm": 28.589920043945312,
      "learning_rate": 2.7841409266693838e-05,
      "loss": 0.5901,
      "step": 1900
    },
    {
      "epoch": 1.82,
      "grad_norm": 36.163124084472656,
      "learning_rate": 2.761458636589813e-05,
      "loss": 0.5612,
      "step": 2000
    },
    {
      "epoch": 1.91,
      "grad_norm": 28.885812759399414,
      "learning_rate": 2.7377456096645395e-05,
      "loss": 0.5999,
      "step": 2100
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.9145129224652088,
      "eval_loss": 0.3024557828903198,
      "eval_runtime": 106.6841,
      "eval_samples_per_second": 23.574,
      "eval_steps_per_second": 1.481,
      "step": 2198
    },
    {
      "epoch": 2.0,
      "grad_norm": 39.23602294921875,
      "learning_rate": 2.7130212217889484e-05,
      "loss": 0.5515,
      "step": 2200
    },
    {
      "epoch": 2.09,
      "grad_norm": 24.46492576599121,
      "learning_rate": 2.6873056752406504e-05,
      "loss": 0.5271,
      "step": 2300
    },
    {
      "epoch": 2.18,
      "grad_norm": 36.76384353637695,
      "learning_rate": 2.6606199821722166e-05,
      "loss": 0.5007,
      "step": 2400
    },
    {
      "epoch": 2.27,
      "grad_norm": 13.54237174987793,
      "learning_rate": 2.632985947442167e-05,
      "loss": 0.5438,
      "step": 2500
    },
    {
      "epoch": 2.37,
      "grad_norm": 15.048121452331543,
      "learning_rate": 2.6044261507982356e-05,
      "loss": 0.512,
      "step": 2600
    },
    {
      "epoch": 2.46,
      "grad_norm": 37.57391357421875,
      "learning_rate": 2.5749639284274782e-05,
      "loss": 0.4896,
      "step": 2700
    },
    {
      "epoch": 2.55,
      "grad_norm": 34.42573547363281,
      "learning_rate": 2.5446233538882924e-05,
      "loss": 0.449,
      "step": 2800
    },
    {
      "epoch": 2.64,
      "grad_norm": 32.23761749267578,
      "learning_rate": 2.513429218439932e-05,
      "loss": 0.476,
      "step": 2900
    },
    {
      "epoch": 2.73,
      "grad_norm": 15.834548950195312,
      "learning_rate": 2.4814070107855878e-05,
      "loss": 0.4739,
      "step": 3000
    },
    {
      "epoch": 2.82,
      "grad_norm": 21.600011825561523,
      "learning_rate": 2.448582896245591e-05,
      "loss": 0.4421,
      "step": 3100
    },
    {
      "epoch": 2.91,
      "grad_norm": 25.791595458984375,
      "learning_rate": 2.4149836953777488e-05,
      "loss": 0.4509,
      "step": 3200
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.9264413518886679,
      "eval_loss": 0.261316180229187,
      "eval_runtime": 106.09,
      "eval_samples_per_second": 23.706,
      "eval_steps_per_second": 1.489,
      "step": 3297
    },
    {
      "epoch": 3.0,
      "grad_norm": 24.959386825561523,
      "learning_rate": 2.3806368620622876e-05,
      "loss": 0.4837,
      "step": 3300
    },
    {
      "epoch": 3.09,
      "grad_norm": 25.834165573120117,
      "learning_rate": 2.345570461069312e-05,
      "loss": 0.4441,
      "step": 3400
    },
    {
      "epoch": 3.18,
      "grad_norm": 9.622383117675781,
      "learning_rate": 2.3098131451271016e-05,
      "loss": 0.4529,
      "step": 3500
    },
    {
      "epoch": 3.28,
      "grad_norm": 19.754833221435547,
      "learning_rate": 2.2733941315099883e-05,
      "loss": 0.4188,
      "step": 3600
    },
    {
      "epoch": 3.37,
      "grad_norm": 30.42739486694336,
      "learning_rate": 2.2363431781649483e-05,
      "loss": 0.4134,
      "step": 3700
    },
    {
      "epoch": 3.46,
      "grad_norm": 17.74396514892578,
      "learning_rate": 2.1986905593964048e-05,
      "loss": 0.4333,
      "step": 3800
    },
    {
      "epoch": 3.55,
      "grad_norm": 19.46295166015625,
      "learning_rate": 2.1604670411291174e-05,
      "loss": 0.4537,
      "step": 3900
    },
    {
      "epoch": 3.64,
      "grad_norm": 28.412992477416992,
      "learning_rate": 2.121703855769373e-05,
      "loss": 0.408,
      "step": 4000
    },
    {
      "epoch": 3.73,
      "grad_norm": 28.82063102722168,
      "learning_rate": 2.0824326766850072e-05,
      "loss": 0.4718,
      "step": 4100
    },
    {
      "epoch": 3.82,
      "grad_norm": 28.174217224121094,
      "learning_rate": 2.042685592325123e-05,
      "loss": 0.435,
      "step": 4200
    },
    {
      "epoch": 3.91,
      "grad_norm": 28.694595336914062,
      "learning_rate": 2.0024950800006463e-05,
      "loss": 0.4366,
      "step": 4300
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.9395626242544731,
      "eval_loss": 0.2299971878528595,
      "eval_runtime": 107.3748,
      "eval_samples_per_second": 23.423,
      "eval_steps_per_second": 1.471,
      "step": 4396
    },
    {
      "epoch": 4.0,
      "grad_norm": 12.610053062438965,
      "learning_rate": 1.961893979347137e-05,
      "loss": 0.3969,
      "step": 4400
    },
    {
      "epoch": 4.09,
      "grad_norm": 25.074817657470703,
      "learning_rate": 1.9209154654915524e-05,
      "loss": 0.3771,
      "step": 4500
    },
    {
      "epoch": 4.19,
      "grad_norm": 34.68314743041992,
      "learning_rate": 1.879593021944875e-05,
      "loss": 0.3735,
      "step": 4600
    },
    {
      "epoch": 4.28,
      "grad_norm": 25.962541580200195,
      "learning_rate": 1.837960413242765e-05,
      "loss": 0.3804,
      "step": 4700
    },
    {
      "epoch": 4.37,
      "grad_norm": 15.469616889953613,
      "learning_rate": 1.796051657356582e-05,
      "loss": 0.3797,
      "step": 4800
    },
    {
      "epoch": 4.46,
      "grad_norm": 10.265162467956543,
      "learning_rate": 1.7539009978973312e-05,
      "loss": 0.3265,
      "step": 4900
    },
    {
      "epoch": 4.55,
      "grad_norm": 27.874446868896484,
      "learning_rate": 1.711542876135233e-05,
      "loss": 0.3491,
      "step": 5000
    },
    {
      "epoch": 4.64,
      "grad_norm": 22.741313934326172,
      "learning_rate": 1.669011902857791e-05,
      "loss": 0.3608,
      "step": 5100
    },
    {
      "epoch": 4.73,
      "grad_norm": 46.997928619384766,
      "learning_rate": 1.6263428300893422e-05,
      "loss": 0.4026,
      "step": 5200
    },
    {
      "epoch": 4.82,
      "grad_norm": 16.42508316040039,
      "learning_rate": 1.5835705226952112e-05,
      "loss": 0.3701,
      "step": 5300
    },
    {
      "epoch": 4.91,
      "grad_norm": 20.580171585083008,
      "learning_rate": 1.540729929893649e-05,
      "loss": 0.3953,
      "step": 5400
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.9411530815109344,
      "eval_loss": 0.2278304100036621,
      "eval_runtime": 105.8697,
      "eval_samples_per_second": 23.756,
      "eval_steps_per_second": 1.492,
      "step": 5495
    },
    {
      "epoch": 5.0,
      "grad_norm": 18.614845275878906,
      "learning_rate": 1.4978560566988603e-05,
      "loss": 0.3845,
      "step": 5500
    },
    {
      "epoch": 5.1,
      "grad_norm": 19.673063278198242,
      "learning_rate": 1.454983935318433e-05,
      "loss": 0.3464,
      "step": 5600
    },
    {
      "epoch": 5.19,
      "grad_norm": 7.708799362182617,
      "learning_rate": 1.4121485965285485e-05,
      "loss": 0.3286,
      "step": 5700
    },
    {
      "epoch": 5.28,
      "grad_norm": 7.584132671356201,
      "learning_rate": 1.3693850410503614e-05,
      "loss": 0.3317,
      "step": 5800
    },
    {
      "epoch": 5.37,
      "grad_norm": 2.4406657218933105,
      "learning_rate": 1.326728210950942e-05,
      "loss": 0.2991,
      "step": 5900
    },
    {
      "epoch": 5.46,
      "grad_norm": 6.6432881355285645,
      "learning_rate": 1.2842129610921378e-05,
      "loss": 0.3055,
      "step": 6000
    },
    {
      "epoch": 5.55,
      "grad_norm": 38.84210968017578,
      "learning_rate": 1.2418740306506923e-05,
      "loss": 0.3784,
      "step": 6100
    },
    {
      "epoch": 5.64,
      "grad_norm": 32.56167984008789,
      "learning_rate": 1.1997460147328984e-05,
      "loss": 0.2934,
      "step": 6200
    },
    {
      "epoch": 5.73,
      "grad_norm": 17.81641960144043,
      "learning_rate": 1.1578633361069559e-05,
      "loss": 0.37,
      "step": 6300
    },
    {
      "epoch": 5.82,
      "grad_norm": 14.04667854309082,
      "learning_rate": 1.1162602170761611e-05,
      "loss": 0.285,
      "step": 6400
    },
    {
      "epoch": 5.91,
      "grad_norm": 27.593286514282227,
      "learning_rate": 1.0749706515158863e-05,
      "loss": 0.307,
      "step": 6500
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.9427435387673956,
      "eval_loss": 0.22470639646053314,
      "eval_runtime": 105.9166,
      "eval_samples_per_second": 23.745,
      "eval_steps_per_second": 1.492,
      "step": 6594
    },
    {
      "epoch": 6.01,
      "grad_norm": 0.9961395263671875,
      "learning_rate": 1.0340283770972167e-05,
      "loss": 0.3046,
      "step": 6600
    },
    {
      "epoch": 6.1,
      "grad_norm": 30.97418212890625,
      "learning_rate": 9.93466847719919e-06,
      "loss": 0.3001,
      "step": 6700
    },
    {
      "epoch": 6.19,
      "grad_norm": 18.71433448791504,
      "learning_rate": 9.533192061772919e-06,
      "loss": 0.3179,
      "step": 6800
    },
    {
      "epoch": 6.28,
      "grad_norm": 29.3361759185791,
      "learning_rate": 9.136182570752153e-06,
      "loss": 0.2813,
      "step": 6900
    },
    {
      "epoch": 6.37,
      "grad_norm": 25.199434280395508,
      "learning_rate": 8.743964400275304e-06,
      "loss": 0.2935,
      "step": 7000
    },
    {
      "epoch": 6.46,
      "grad_norm": 11.149260520935059,
      "learning_rate": 8.356858031496596e-06,
      "loss": 0.307,
      "step": 7100
    },
    {
      "epoch": 6.55,
      "grad_norm": 17.286054611206055,
      "learning_rate": 7.975179768721187e-06,
      "loss": 0.3253,
      "step": 7200
    },
    {
      "epoch": 6.64,
      "grad_norm": 23.554401397705078,
      "learning_rate": 7.599241480953112e-06,
      "loss": 0.2685,
      "step": 7300
    },
    {
      "epoch": 6.73,
      "grad_norm": 38.09795379638672,
      "learning_rate": 7.229350347067426e-06,
      "loss": 0.261,
      "step": 7400
    },
    {
      "epoch": 6.82,
      "grad_norm": 44.90071487426758,
      "learning_rate": 6.865808604814564e-06,
      "loss": 0.2815,
      "step": 7500
    },
    {
      "epoch": 6.92,
      "grad_norm": 15.631821632385254,
      "learning_rate": 6.508913303862144e-06,
      "loss": 0.2967,
      "step": 7600
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.9443339960238568,
      "eval_loss": 0.21055155992507935,
      "eval_runtime": 105.5753,
      "eval_samples_per_second": 23.822,
      "eval_steps_per_second": 1.497,
      "step": 7693
    },
    {
      "epoch": 7.01,
      "grad_norm": 24.378986358642578,
      "learning_rate": 6.1589560630758656e-06,
      "loss": 0.2478,
      "step": 7700
    },
    {
      "epoch": 7.1,
      "grad_norm": 17.62042236328125,
      "learning_rate": 5.8162228322380155e-06,
      "loss": 0.2642,
      "step": 7800
    },
    {
      "epoch": 7.19,
      "grad_norm": 22.40974235534668,
      "learning_rate": 5.480993658398129e-06,
      "loss": 0.2787,
      "step": 7900
    },
    {
      "epoch": 7.28,
      "grad_norm": 10.600983619689941,
      "learning_rate": 5.153542457046737e-06,
      "loss": 0.2709,
      "step": 8000
    },
    {
      "epoch": 7.37,
      "grad_norm": 7.483508586883545,
      "learning_rate": 4.834136788299248e-06,
      "loss": 0.2679,
      "step": 8100
    },
    {
      "epoch": 7.46,
      "grad_norm": 22.654191970825195,
      "learning_rate": 4.523037638272822e-06,
      "loss": 0.237,
      "step": 8200
    },
    {
      "epoch": 7.55,
      "grad_norm": 25.131502151489258,
      "learning_rate": 4.220499205834783e-06,
      "loss": 0.2703,
      "step": 8300
    },
    {
      "epoch": 7.64,
      "grad_norm": 15.906916618347168,
      "learning_rate": 3.926768694896931e-06,
      "loss": 0.2543,
      "step": 8400
    },
    {
      "epoch": 7.73,
      "grad_norm": 27.455238342285156,
      "learning_rate": 3.6420861124254607e-06,
      "loss": 0.2543,
      "step": 8500
    },
    {
      "epoch": 7.83,
      "grad_norm": 22.13401222229004,
      "learning_rate": 3.3666840723314145e-06,
      "loss": 0.3041,
      "step": 8600
    },
    {
      "epoch": 7.92,
      "grad_norm": 15.260415077209473,
      "learning_rate": 3.1007876054020724e-06,
      "loss": 0.2549,
      "step": 8700
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.9491053677932405,
      "eval_loss": 0.2030409872531891,
      "eval_runtime": 106.5142,
      "eval_samples_per_second": 23.612,
      "eval_steps_per_second": 1.483,
      "step": 8792
    },
    {
      "epoch": 8.01,
      "grad_norm": 9.118292808532715,
      "learning_rate": 2.8446139754284486e-06,
      "loss": 0.2919,
      "step": 8800
    },
    {
      "epoch": 8.1,
      "grad_norm": 19.433610916137695,
      "learning_rate": 2.5983725016792574e-06,
      "loss": 0.2496,
      "step": 8900
    },
    {
      "epoch": 8.19,
      "grad_norm": 20.59990882873535,
      "learning_rate": 2.36226438786627e-06,
      "loss": 0.2737,
      "step": 9000
    },
    {
      "epoch": 8.28,
      "grad_norm": 16.759319305419922,
      "learning_rate": 2.1364825577409424e-06,
      "loss": 0.2583,
      "step": 9100
    },
    {
      "epoch": 8.37,
      "grad_norm": 3.51478910446167,
      "learning_rate": 1.9212114974565664e-06,
      "loss": 0.2179,
      "step": 9200
    },
    {
      "epoch": 8.46,
      "grad_norm": 8.665757179260254,
      "learning_rate": 1.7166271048247796e-06,
      "loss": 0.2125,
      "step": 9300
    },
    {
      "epoch": 8.55,
      "grad_norm": 40.940914154052734,
      "learning_rate": 1.5228965455896054e-06,
      "loss": 0.2357,
      "step": 9400
    },
    {
      "epoch": 8.64,
      "grad_norm": 26.388017654418945,
      "learning_rate": 1.3401781168364591e-06,
      "loss": 0.2593,
      "step": 9500
    },
    {
      "epoch": 8.74,
      "grad_norm": 4.975929260253906,
      "learning_rate": 1.1686211176477208e-06,
      "loss": 0.2489,
      "step": 9600
    },
    {
      "epoch": 8.83,
      "grad_norm": 39.22834014892578,
      "learning_rate": 1.00836572711058e-06,
      "loss": 0.2468,
      "step": 9700
    },
    {
      "epoch": 8.92,
      "grad_norm": 3.737492322921753,
      "learning_rate": 8.595428897768071e-07,
      "loss": 0.275,
      "step": 9800
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.9463220675944334,
      "eval_loss": 0.20357173681259155,
      "eval_runtime": 106.7748,
      "eval_samples_per_second": 23.554,
      "eval_steps_per_second": 1.48,
      "step": 9891
    },
    {
      "epoch": 9.01,
      "grad_norm": 44.20964050292969,
      "learning_rate": 7.222742086680756e-07,
      "loss": 0.2398,
      "step": 9900
    },
    {
      "epoch": 9.1,
      "grad_norm": 0.32034990191459656,
      "learning_rate": 5.966718459142196e-07,
      "loss": 0.2214,
      "step": 10000
    },
    {
      "epoch": 9.19,
      "grad_norm": 22.17435646057129,
      "learning_rate": 4.82838431105655e-07,
      "loss": 0.2647,
      "step": 10100
    },
    {
      "epoch": 9.28,
      "grad_norm": 8.493167877197266,
      "learning_rate": 3.808669774348167e-07,
      "loss": 0.2548,
      "step": 10200
    },
    {
      "epoch": 9.37,
      "grad_norm": 17.518217086791992,
      "learning_rate": 2.908408056951578e-07,
      "loss": 0.2497,
      "step": 10300
    },
    {
      "epoch": 9.46,
      "grad_norm": 18.134353637695312,
      "learning_rate": 2.1283347619979243e-07,
      "loss": 0.2265,
      "step": 10400
    },
    {
      "epoch": 9.55,
      "grad_norm": 21.909814834594727,
      "learning_rate": 1.4690872867542892e-07,
      "loss": 0.2371,
      "step": 10500
    },
    {
      "epoch": 9.65,
      "grad_norm": 25.709300994873047,
      "learning_rate": 9.312043018067762e-08,
      "loss": 0.2599,
      "step": 10600
    },
    {
      "epoch": 9.74,
      "grad_norm": 13.61241340637207,
      "learning_rate": 5.1512531091333914e-08,
      "loss": 0.235,
      "step": 10700
    },
    {
      "epoch": 9.83,
      "grad_norm": 20.345327377319336,
      "learning_rate": 2.211902918855313e-08,
      "loss": 0.2686,
      "step": 10800
    },
    {
      "epoch": 9.92,
      "grad_norm": 20.140609741210938,
      "learning_rate": 4.963941879295164e-09,
      "loss": 0.24,
      "step": 10900
    },
    {
      "epoch": 10.0,
      "eval_accuracy": 0.947117296222664,
      "eval_loss": 0.2026742845773697,
      "eval_runtime": 105.8999,
      "eval_samples_per_second": 23.749,
      "eval_steps_per_second": 1.492,
      "step": 10990
    },
    {
      "epoch": 10.0,
      "step": 10990,
      "total_flos": 4.09349935387607e+19,
      "train_loss": 0.44345692520905233,
      "train_runtime": 20905.109,
      "train_samples_per_second": 8.41,
      "train_steps_per_second": 0.526
    }
  ],
  "logging_steps": 100,
  "max_steps": 10990,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 10,
  "save_steps": 500,
  "total_flos": 4.09349935387607e+19,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": null
}