{
"best_metric": 0.019898300990462303,
"best_model_checkpoint": "./vit-spoof/checkpoint-2600",
"epoch": 4.0,
"eval_steps": 100,
"global_step": 3004,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.013315579227696404,
"grad_norm": 2.9065542221069336,
"learning_rate": 0.00019933422103861519,
"loss": 0.4683,
"step": 10
},
{
"epoch": 0.02663115845539281,
"grad_norm": 5.029272556304932,
"learning_rate": 0.00019866844207723036,
"loss": 0.4784,
"step": 20
},
{
"epoch": 0.03994673768308921,
"grad_norm": 1.0903401374816895,
"learning_rate": 0.00019800266311584554,
"loss": 0.2297,
"step": 30
},
{
"epoch": 0.05326231691078562,
"grad_norm": 13.801852226257324,
"learning_rate": 0.00019733688415446071,
"loss": 0.3551,
"step": 40
},
{
"epoch": 0.06657789613848203,
"grad_norm": 7.049527645111084,
"learning_rate": 0.0001966711051930759,
"loss": 0.1737,
"step": 50
},
{
"epoch": 0.07989347536617843,
"grad_norm": 4.438915729522705,
"learning_rate": 0.00019600532623169107,
"loss": 0.4895,
"step": 60
},
{
"epoch": 0.09320905459387484,
"grad_norm": 2.351810932159424,
"learning_rate": 0.00019533954727030627,
"loss": 0.1397,
"step": 70
},
{
"epoch": 0.10652463382157124,
"grad_norm": 0.2668318748474121,
"learning_rate": 0.00019467376830892145,
"loss": 0.2286,
"step": 80
},
{
"epoch": 0.11984021304926765,
"grad_norm": 0.18876244127750397,
"learning_rate": 0.00019400798934753662,
"loss": 0.0623,
"step": 90
},
{
"epoch": 0.13315579227696406,
"grad_norm": 0.08632949739694595,
"learning_rate": 0.00019334221038615183,
"loss": 0.0682,
"step": 100
},
{
"epoch": 0.13315579227696406,
"eval_accuracy": 0.9505988023952096,
"eval_loss": 0.16598451137542725,
"eval_runtime": 18.9844,
"eval_samples_per_second": 70.374,
"eval_steps_per_second": 8.797,
"step": 100
},
{
"epoch": 0.14647137150466044,
"grad_norm": 13.015838623046875,
"learning_rate": 0.000192676431424767,
"loss": 0.1278,
"step": 110
},
{
"epoch": 0.15978695073235685,
"grad_norm": 0.0676693394780159,
"learning_rate": 0.00019201065246338218,
"loss": 0.048,
"step": 120
},
{
"epoch": 0.17310252996005326,
"grad_norm": 0.08669548481702805,
"learning_rate": 0.00019134487350199735,
"loss": 0.0652,
"step": 130
},
{
"epoch": 0.18641810918774968,
"grad_norm": 2.5868160724639893,
"learning_rate": 0.00019067909454061253,
"loss": 0.1912,
"step": 140
},
{
"epoch": 0.19973368841544606,
"grad_norm": 0.046874042600393295,
"learning_rate": 0.0001900133155792277,
"loss": 0.0425,
"step": 150
},
{
"epoch": 0.21304926764314247,
"grad_norm": 0.7398509383201599,
"learning_rate": 0.00018934753661784288,
"loss": 0.0862,
"step": 160
},
{
"epoch": 0.22636484687083888,
"grad_norm": 1.8706042766571045,
"learning_rate": 0.00018868175765645806,
"loss": 0.4261,
"step": 170
},
{
"epoch": 0.2396804260985353,
"grad_norm": 8.126788139343262,
"learning_rate": 0.00018801597869507323,
"loss": 0.2851,
"step": 180
},
{
"epoch": 0.2529960053262317,
"grad_norm": 0.4172131419181824,
"learning_rate": 0.0001873501997336884,
"loss": 0.1001,
"step": 190
},
{
"epoch": 0.2663115845539281,
"grad_norm": 0.6365886926651001,
"learning_rate": 0.00018668442077230361,
"loss": 0.0711,
"step": 200
},
{
"epoch": 0.2663115845539281,
"eval_accuracy": 0.9805389221556886,
"eval_loss": 0.07567165791988373,
"eval_runtime": 19.0052,
"eval_samples_per_second": 70.297,
"eval_steps_per_second": 8.787,
"step": 200
},
{
"epoch": 0.2796271637816245,
"grad_norm": 0.057595234364271164,
"learning_rate": 0.0001860186418109188,
"loss": 0.1514,
"step": 210
},
{
"epoch": 0.2929427430093209,
"grad_norm": 8.712677955627441,
"learning_rate": 0.00018535286284953397,
"loss": 0.2079,
"step": 220
},
{
"epoch": 0.3062583222370173,
"grad_norm": 0.33387067914009094,
"learning_rate": 0.00018468708388814914,
"loss": 0.0374,
"step": 230
},
{
"epoch": 0.3195739014647137,
"grad_norm": 12.936437606811523,
"learning_rate": 0.00018402130492676432,
"loss": 0.154,
"step": 240
},
{
"epoch": 0.33288948069241014,
"grad_norm": 0.08830924332141876,
"learning_rate": 0.0001833555259653795,
"loss": 0.0712,
"step": 250
},
{
"epoch": 0.34620505992010653,
"grad_norm": 0.13204559683799744,
"learning_rate": 0.00018268974700399467,
"loss": 0.1426,
"step": 260
},
{
"epoch": 0.3595206391478029,
"grad_norm": 0.2291734367609024,
"learning_rate": 0.00018202396804260987,
"loss": 0.0472,
"step": 270
},
{
"epoch": 0.37283621837549935,
"grad_norm": 0.054096538573503494,
"learning_rate": 0.00018135818908122505,
"loss": 0.086,
"step": 280
},
{
"epoch": 0.38615179760319573,
"grad_norm": 4.264747619628906,
"learning_rate": 0.00018069241011984023,
"loss": 0.1287,
"step": 290
},
{
"epoch": 0.3994673768308921,
"grad_norm": 0.15629783272743225,
"learning_rate": 0.0001800266311584554,
"loss": 0.0112,
"step": 300
},
{
"epoch": 0.3994673768308921,
"eval_accuracy": 0.9550898203592815,
"eval_loss": 0.13134559988975525,
"eval_runtime": 18.8497,
"eval_samples_per_second": 70.876,
"eval_steps_per_second": 8.86,
"step": 300
},
{
"epoch": 0.41278295605858856,
"grad_norm": 4.84935188293457,
"learning_rate": 0.00017936085219707058,
"loss": 0.117,
"step": 310
},
{
"epoch": 0.42609853528628494,
"grad_norm": 0.580627977848053,
"learning_rate": 0.00017869507323568575,
"loss": 0.1188,
"step": 320
},
{
"epoch": 0.4394141145139814,
"grad_norm": 4.605671405792236,
"learning_rate": 0.00017802929427430096,
"loss": 0.1802,
"step": 330
},
{
"epoch": 0.45272969374167776,
"grad_norm": 0.16331535577774048,
"learning_rate": 0.00017736351531291613,
"loss": 0.1645,
"step": 340
},
{
"epoch": 0.46604527296937415,
"grad_norm": 9.795440673828125,
"learning_rate": 0.0001766977363515313,
"loss": 0.1133,
"step": 350
},
{
"epoch": 0.4793608521970706,
"grad_norm": 1.2079029083251953,
"learning_rate": 0.00017603195739014649,
"loss": 0.06,
"step": 360
},
{
"epoch": 0.49267643142476697,
"grad_norm": 0.02436252497136593,
"learning_rate": 0.00017536617842876166,
"loss": 0.0093,
"step": 370
},
{
"epoch": 0.5059920106524634,
"grad_norm": 0.19958335161209106,
"learning_rate": 0.00017470039946737684,
"loss": 0.0398,
"step": 380
},
{
"epoch": 0.5193075898801598,
"grad_norm": 0.8796051740646362,
"learning_rate": 0.00017403462050599201,
"loss": 0.1084,
"step": 390
},
{
"epoch": 0.5326231691078562,
"grad_norm": 11.185724258422852,
"learning_rate": 0.0001733688415446072,
"loss": 0.0544,
"step": 400
},
{
"epoch": 0.5326231691078562,
"eval_accuracy": 0.9850299401197605,
"eval_loss": 0.05561767518520355,
"eval_runtime": 19.0907,
"eval_samples_per_second": 69.982,
"eval_steps_per_second": 8.748,
"step": 400
},
{
"epoch": 0.5459387483355526,
"grad_norm": 0.045906443148851395,
"learning_rate": 0.00017270306258322237,
"loss": 0.049,
"step": 410
},
{
"epoch": 0.559254327563249,
"grad_norm": 0.03049567900598049,
"learning_rate": 0.00017203728362183754,
"loss": 0.0573,
"step": 420
},
{
"epoch": 0.5725699067909454,
"grad_norm": 8.935633659362793,
"learning_rate": 0.00017137150466045272,
"loss": 0.324,
"step": 430
},
{
"epoch": 0.5858854860186418,
"grad_norm": 0.19179384410381317,
"learning_rate": 0.00017070572569906792,
"loss": 0.0121,
"step": 440
},
{
"epoch": 0.5992010652463382,
"grad_norm": 3.0050430297851562,
"learning_rate": 0.0001700399467376831,
"loss": 0.0361,
"step": 450
},
{
"epoch": 0.6125166444740346,
"grad_norm": 0.0664597898721695,
"learning_rate": 0.00016937416777629827,
"loss": 0.1106,
"step": 460
},
{
"epoch": 0.625832223701731,
"grad_norm": 5.098622798919678,
"learning_rate": 0.00016870838881491348,
"loss": 0.1938,
"step": 470
},
{
"epoch": 0.6391478029294274,
"grad_norm": 4.372994899749756,
"learning_rate": 0.00016804260985352865,
"loss": 0.0807,
"step": 480
},
{
"epoch": 0.6524633821571239,
"grad_norm": 0.28381890058517456,
"learning_rate": 0.00016737683089214383,
"loss": 0.0463,
"step": 490
},
{
"epoch": 0.6657789613848203,
"grad_norm": 0.8805880546569824,
"learning_rate": 0.000166711051930759,
"loss": 0.0557,
"step": 500
},
{
"epoch": 0.6657789613848203,
"eval_accuracy": 0.968562874251497,
"eval_loss": 0.09653697907924652,
"eval_runtime": 18.9994,
"eval_samples_per_second": 70.318,
"eval_steps_per_second": 8.79,
"step": 500
},
{
"epoch": 0.6790945406125166,
"grad_norm": 0.3464203178882599,
"learning_rate": 0.00016604527296937418,
"loss": 0.1231,
"step": 510
},
{
"epoch": 0.6924101198402131,
"grad_norm": 0.10989736020565033,
"learning_rate": 0.00016537949400798936,
"loss": 0.0666,
"step": 520
},
{
"epoch": 0.7057256990679095,
"grad_norm": 2.2785890102386475,
"learning_rate": 0.00016471371504660453,
"loss": 0.1636,
"step": 530
},
{
"epoch": 0.7190412782956058,
"grad_norm": 0.07484255731105804,
"learning_rate": 0.0001640479360852197,
"loss": 0.0288,
"step": 540
},
{
"epoch": 0.7323568575233023,
"grad_norm": 0.02765025570988655,
"learning_rate": 0.0001633821571238349,
"loss": 0.0514,
"step": 550
},
{
"epoch": 0.7456724367509987,
"grad_norm": 0.07284240424633026,
"learning_rate": 0.00016271637816245006,
"loss": 0.0547,
"step": 560
},
{
"epoch": 0.758988015978695,
"grad_norm": 0.07628186792135239,
"learning_rate": 0.00016205059920106524,
"loss": 0.0239,
"step": 570
},
{
"epoch": 0.7723035952063915,
"grad_norm": 0.051849640905857086,
"learning_rate": 0.00016138482023968042,
"loss": 0.0568,
"step": 580
},
{
"epoch": 0.7856191744340879,
"grad_norm": 0.12565506994724274,
"learning_rate": 0.00016071904127829562,
"loss": 0.0568,
"step": 590
},
{
"epoch": 0.7989347536617842,
"grad_norm": 0.022169604897499084,
"learning_rate": 0.0001600532623169108,
"loss": 0.016,
"step": 600
},
{
"epoch": 0.7989347536617842,
"eval_accuracy": 0.9827844311377245,
"eval_loss": 0.05710091441869736,
"eval_runtime": 19.0923,
"eval_samples_per_second": 69.976,
"eval_steps_per_second": 8.747,
"step": 600
},
{
"epoch": 0.8122503328894807,
"grad_norm": 8.000443458557129,
"learning_rate": 0.00015938748335552597,
"loss": 0.0764,
"step": 610
},
{
"epoch": 0.8255659121171771,
"grad_norm": 4.5883965492248535,
"learning_rate": 0.00015872170439414115,
"loss": 0.1149,
"step": 620
},
{
"epoch": 0.8388814913448736,
"grad_norm": 0.1661900281906128,
"learning_rate": 0.00015805592543275632,
"loss": 0.0243,
"step": 630
},
{
"epoch": 0.8521970705725699,
"grad_norm": 2.3490681648254395,
"learning_rate": 0.00015739014647137153,
"loss": 0.1185,
"step": 640
},
{
"epoch": 0.8655126498002663,
"grad_norm": 0.3317602276802063,
"learning_rate": 0.0001567243675099867,
"loss": 0.0235,
"step": 650
},
{
"epoch": 0.8788282290279628,
"grad_norm": 0.038634102791547775,
"learning_rate": 0.00015605858854860188,
"loss": 0.0774,
"step": 660
},
{
"epoch": 0.8921438082556591,
"grad_norm": 0.05563969910144806,
"learning_rate": 0.00015539280958721705,
"loss": 0.0812,
"step": 670
},
{
"epoch": 0.9054593874833555,
"grad_norm": 0.029876109212636948,
"learning_rate": 0.00015472703062583223,
"loss": 0.0314,
"step": 680
},
{
"epoch": 0.918774966711052,
"grad_norm": 1.0358515977859497,
"learning_rate": 0.0001540612516644474,
"loss": 0.0538,
"step": 690
},
{
"epoch": 0.9320905459387483,
"grad_norm": 0.21664029359817505,
"learning_rate": 0.00015339547270306258,
"loss": 0.0382,
"step": 700
},
{
"epoch": 0.9320905459387483,
"eval_accuracy": 0.9925149700598802,
"eval_loss": 0.029164018109440804,
"eval_runtime": 19.0562,
"eval_samples_per_second": 70.108,
"eval_steps_per_second": 8.764,
"step": 700
},
{
"epoch": 0.9454061251664447,
"grad_norm": 0.24551372230052948,
"learning_rate": 0.00015272969374167776,
"loss": 0.0445,
"step": 710
},
{
"epoch": 0.9587217043941412,
"grad_norm": 0.024771912023425102,
"learning_rate": 0.00015206391478029296,
"loss": 0.0366,
"step": 720
},
{
"epoch": 0.9720372836218375,
"grad_norm": 0.027900271117687225,
"learning_rate": 0.00015139813581890814,
"loss": 0.0044,
"step": 730
},
{
"epoch": 0.9853528628495339,
"grad_norm": 0.023675180971622467,
"learning_rate": 0.00015073235685752331,
"loss": 0.0191,
"step": 740
},
{
"epoch": 0.9986684420772304,
"grad_norm": 0.1361338347196579,
"learning_rate": 0.0001500665778961385,
"loss": 0.0496,
"step": 750
},
{
"epoch": 1.0119840213049267,
"grad_norm": 0.016286784783005714,
"learning_rate": 0.00014940079893475367,
"loss": 0.0658,
"step": 760
},
{
"epoch": 1.0252996005326231,
"grad_norm": 0.04202214255928993,
"learning_rate": 0.00014873501997336884,
"loss": 0.074,
"step": 770
},
{
"epoch": 1.0386151797603196,
"grad_norm": 0.10847309976816177,
"learning_rate": 0.00014806924101198402,
"loss": 0.0871,
"step": 780
},
{
"epoch": 1.051930758988016,
"grad_norm": 0.049633290618658066,
"learning_rate": 0.0001474034620505992,
"loss": 0.0098,
"step": 790
},
{
"epoch": 1.0652463382157125,
"grad_norm": 0.029612813144922256,
"learning_rate": 0.00014673768308921437,
"loss": 0.0281,
"step": 800
},
{
"epoch": 1.0652463382157125,
"eval_accuracy": 0.9782934131736527,
"eval_loss": 0.07025627046823502,
"eval_runtime": 18.7557,
"eval_samples_per_second": 71.232,
"eval_steps_per_second": 8.904,
"step": 800
},
{
"epoch": 1.078561917443409,
"grad_norm": 15.133095741271973,
"learning_rate": 0.00014607190412782957,
"loss": 0.1002,
"step": 810
},
{
"epoch": 1.0918774966711051,
"grad_norm": 0.051363505423069,
"learning_rate": 0.00014540612516644475,
"loss": 0.0365,
"step": 820
},
{
"epoch": 1.1051930758988016,
"grad_norm": 0.03931460902094841,
"learning_rate": 0.00014474034620505993,
"loss": 0.0033,
"step": 830
},
{
"epoch": 1.118508655126498,
"grad_norm": 0.02210480161011219,
"learning_rate": 0.0001440745672436751,
"loss": 0.0695,
"step": 840
},
{
"epoch": 1.1318242343541944,
"grad_norm": 1.2839746475219727,
"learning_rate": 0.0001434087882822903,
"loss": 0.0675,
"step": 850
},
{
"epoch": 1.1451398135818909,
"grad_norm": 0.31109488010406494,
"learning_rate": 0.00014274300932090548,
"loss": 0.0486,
"step": 860
},
{
"epoch": 1.158455392809587,
"grad_norm": 1.5788602828979492,
"learning_rate": 0.00014207723035952066,
"loss": 0.0325,
"step": 870
},
{
"epoch": 1.1717709720372835,
"grad_norm": 0.03321152552962303,
"learning_rate": 0.00014141145139813583,
"loss": 0.0796,
"step": 880
},
{
"epoch": 1.18508655126498,
"grad_norm": 12.448341369628906,
"learning_rate": 0.000140745672436751,
"loss": 0.0515,
"step": 890
},
{
"epoch": 1.1984021304926764,
"grad_norm": 0.1457500457763672,
"learning_rate": 0.0001400798934753662,
"loss": 0.023,
"step": 900
},
{
"epoch": 1.1984021304926764,
"eval_accuracy": 0.9932634730538922,
"eval_loss": 0.02668871358036995,
"eval_runtime": 18.662,
"eval_samples_per_second": 71.589,
"eval_steps_per_second": 8.949,
"step": 900
},
{
"epoch": 1.2117177097203728,
"grad_norm": 0.14423052966594696,
"learning_rate": 0.00013941411451398136,
"loss": 0.08,
"step": 910
},
{
"epoch": 1.2250332889480693,
"grad_norm": 0.055785391479730606,
"learning_rate": 0.00013874833555259654,
"loss": 0.0445,
"step": 920
},
{
"epoch": 1.2383488681757657,
"grad_norm": 0.0294383205473423,
"learning_rate": 0.00013808255659121172,
"loss": 0.0414,
"step": 930
},
{
"epoch": 1.2516644474034622,
"grad_norm": 0.0800587385892868,
"learning_rate": 0.0001374167776298269,
"loss": 0.003,
"step": 940
},
{
"epoch": 1.2649800266311584,
"grad_norm": 0.08012738078832626,
"learning_rate": 0.00013675099866844207,
"loss": 0.0042,
"step": 950
},
{
"epoch": 1.2782956058588548,
"grad_norm": 0.022230001166462898,
"learning_rate": 0.00013608521970705724,
"loss": 0.0361,
"step": 960
},
{
"epoch": 1.2916111850865513,
"grad_norm": 0.014080135151743889,
"learning_rate": 0.00013541944074567242,
"loss": 0.004,
"step": 970
},
{
"epoch": 1.3049267643142477,
"grad_norm": 0.01428599376231432,
"learning_rate": 0.00013475366178428762,
"loss": 0.0016,
"step": 980
},
{
"epoch": 1.3182423435419441,
"grad_norm": 0.01126811746507883,
"learning_rate": 0.0001340878828229028,
"loss": 0.0014,
"step": 990
},
{
"epoch": 1.3315579227696404,
"grad_norm": 0.016759509220719337,
"learning_rate": 0.00013342210386151798,
"loss": 0.0695,
"step": 1000
},
{
"epoch": 1.3315579227696404,
"eval_accuracy": 0.9887724550898204,
"eval_loss": 0.043992191553115845,
"eval_runtime": 18.8129,
"eval_samples_per_second": 71.015,
"eval_steps_per_second": 8.877,
"step": 1000
},
{
"epoch": 1.3448735019973368,
"grad_norm": 12.30389404296875,
"learning_rate": 0.00013275632490013318,
"loss": 0.0344,
"step": 1010
},
{
"epoch": 1.3581890812250332,
"grad_norm": 0.019324608147144318,
"learning_rate": 0.00013209054593874836,
"loss": 0.1322,
"step": 1020
},
{
"epoch": 1.3715046604527297,
"grad_norm": 0.08052018284797668,
"learning_rate": 0.00013142476697736353,
"loss": 0.0878,
"step": 1030
},
{
"epoch": 1.3848202396804261,
"grad_norm": 0.015013271011412144,
"learning_rate": 0.0001307589880159787,
"loss": 0.0176,
"step": 1040
},
{
"epoch": 1.3981358189081226,
"grad_norm": 1.5079394578933716,
"learning_rate": 0.00013009320905459388,
"loss": 0.0757,
"step": 1050
},
{
"epoch": 1.411451398135819,
"grad_norm": 1.3230667114257812,
"learning_rate": 0.00012942743009320906,
"loss": 0.054,
"step": 1060
},
{
"epoch": 1.4247669773635154,
"grad_norm": 0.025523126125335693,
"learning_rate": 0.00012876165113182424,
"loss": 0.0439,
"step": 1070
},
{
"epoch": 1.4380825565912116,
"grad_norm": 4.955082893371582,
"learning_rate": 0.0001280958721704394,
"loss": 0.0196,
"step": 1080
},
{
"epoch": 1.451398135818908,
"grad_norm": 0.07307817041873932,
"learning_rate": 0.0001274300932090546,
"loss": 0.0637,
"step": 1090
},
{
"epoch": 1.4647137150466045,
"grad_norm": 0.016766542568802834,
"learning_rate": 0.00012676431424766976,
"loss": 0.0215,
"step": 1100
},
{
"epoch": 1.4647137150466045,
"eval_accuracy": 0.9910179640718563,
"eval_loss": 0.031917981803417206,
"eval_runtime": 18.9879,
"eval_samples_per_second": 70.361,
"eval_steps_per_second": 8.795,
"step": 1100
},
{
"epoch": 1.478029294274301,
"grad_norm": 0.010708771646022797,
"learning_rate": 0.00012609853528628497,
"loss": 0.0036,
"step": 1110
},
{
"epoch": 1.4913448735019974,
"grad_norm": 0.010356970131397247,
"learning_rate": 0.00012543275632490014,
"loss": 0.0016,
"step": 1120
},
{
"epoch": 1.5046604527296936,
"grad_norm": 0.010580988600850105,
"learning_rate": 0.00012476697736351532,
"loss": 0.0354,
"step": 1130
},
{
"epoch": 1.51797603195739,
"grad_norm": 0.009461847133934498,
"learning_rate": 0.0001241011984021305,
"loss": 0.0011,
"step": 1140
},
{
"epoch": 1.5312916111850865,
"grad_norm": 0.011837669648230076,
"learning_rate": 0.00012343541944074567,
"loss": 0.0434,
"step": 1150
},
{
"epoch": 1.544607190412783,
"grad_norm": 0.01360328495502472,
"learning_rate": 0.00012276964047936085,
"loss": 0.0013,
"step": 1160
},
{
"epoch": 1.5579227696404794,
"grad_norm": 0.010293539613485336,
"learning_rate": 0.00012210386151797602,
"loss": 0.0013,
"step": 1170
},
{
"epoch": 1.5712383488681758,
"grad_norm": 0.011664211750030518,
"learning_rate": 0.00012143808255659121,
"loss": 0.0013,
"step": 1180
},
{
"epoch": 1.5845539280958723,
"grad_norm": 0.03968218341469765,
"learning_rate": 0.00012077230359520639,
"loss": 0.0067,
"step": 1190
},
{
"epoch": 1.5978695073235687,
"grad_norm": 1.3615704774856567,
"learning_rate": 0.00012010652463382157,
"loss": 0.0236,
"step": 1200
},
{
"epoch": 1.5978695073235687,
"eval_accuracy": 0.9880239520958084,
"eval_loss": 0.06240475922822952,
"eval_runtime": 19.0358,
"eval_samples_per_second": 70.184,
"eval_steps_per_second": 8.773,
"step": 1200
},
{
"epoch": 1.6111850865512651,
"grad_norm": 0.3500171899795532,
"learning_rate": 0.00011944074567243674,
"loss": 0.0198,
"step": 1210
},
{
"epoch": 1.6245006657789614,
"grad_norm": 0.007627115119248629,
"learning_rate": 0.00011877496671105193,
"loss": 0.0016,
"step": 1220
},
{
"epoch": 1.6378162450066578,
"grad_norm": 0.07766377925872803,
"learning_rate": 0.00011810918774966711,
"loss": 0.001,
"step": 1230
},
{
"epoch": 1.6511318242343542,
"grad_norm": 0.008079525083303452,
"learning_rate": 0.00011744340878828231,
"loss": 0.001,
"step": 1240
},
{
"epoch": 1.6644474034620504,
"grad_norm": 14.718483924865723,
"learning_rate": 0.00011677762982689749,
"loss": 0.0543,
"step": 1250
},
{
"epoch": 1.6777629826897469,
"grad_norm": 3.2829227447509766,
"learning_rate": 0.00011611185086551266,
"loss": 0.0409,
"step": 1260
},
{
"epoch": 1.6910785619174433,
"grad_norm": 0.8293498158454895,
"learning_rate": 0.00011544607190412784,
"loss": 0.0632,
"step": 1270
},
{
"epoch": 1.7043941411451398,
"grad_norm": 0.15176649391651154,
"learning_rate": 0.00011478029294274302,
"loss": 0.0056,
"step": 1280
},
{
"epoch": 1.7177097203728362,
"grad_norm": 0.015873638913035393,
"learning_rate": 0.00011411451398135819,
"loss": 0.0032,
"step": 1290
},
{
"epoch": 1.7310252996005326,
"grad_norm": 0.011890141293406487,
"learning_rate": 0.00011344873501997337,
"loss": 0.0021,
"step": 1300
},
{
"epoch": 1.7310252996005326,
"eval_accuracy": 0.9895209580838323,
"eval_loss": 0.05610267445445061,
"eval_runtime": 18.9536,
"eval_samples_per_second": 70.488,
"eval_steps_per_second": 8.811,
"step": 1300
},
{
"epoch": 1.744340878828229,
"grad_norm": 0.009211612865328789,
"learning_rate": 0.00011278295605858856,
"loss": 0.0014,
"step": 1310
},
{
"epoch": 1.7576564580559255,
"grad_norm": 0.015255720354616642,
"learning_rate": 0.00011211717709720373,
"loss": 0.0088,
"step": 1320
},
{
"epoch": 1.770972037283622,
"grad_norm": 0.009559585712850094,
"learning_rate": 0.00011145139813581891,
"loss": 0.0203,
"step": 1330
},
{
"epoch": 1.7842876165113184,
"grad_norm": 39.935569763183594,
"learning_rate": 0.00011078561917443409,
"loss": 0.0302,
"step": 1340
},
{
"epoch": 1.7976031957390146,
"grad_norm": 0.010259670205414295,
"learning_rate": 0.00011011984021304926,
"loss": 0.0721,
"step": 1350
},
{
"epoch": 1.810918774966711,
"grad_norm": 0.013963967561721802,
"learning_rate": 0.00010945406125166447,
"loss": 0.0358,
"step": 1360
},
{
"epoch": 1.8242343541944075,
"grad_norm": 0.2606169283390045,
"learning_rate": 0.00010878828229027964,
"loss": 0.0179,
"step": 1370
},
{
"epoch": 1.8375499334221037,
"grad_norm": 21.295629501342773,
"learning_rate": 0.00010812250332889482,
"loss": 0.0677,
"step": 1380
},
{
"epoch": 1.8508655126498001,
"grad_norm": 0.014277924783527851,
"learning_rate": 0.00010745672436751,
"loss": 0.0074,
"step": 1390
},
{
"epoch": 1.8641810918774966,
"grad_norm": 0.007295672316104174,
"learning_rate": 0.00010679094540612517,
"loss": 0.0218,
"step": 1400
},
{
"epoch": 1.8641810918774966,
"eval_accuracy": 0.9820359281437125,
"eval_loss": 0.07637928426265717,
"eval_runtime": 18.728,
"eval_samples_per_second": 71.337,
"eval_steps_per_second": 8.917,
"step": 1400
},
{
"epoch": 1.877496671105193,
"grad_norm": 0.12700679898262024,
"learning_rate": 0.00010612516644474036,
"loss": 0.0066,
"step": 1410
},
{
"epoch": 1.8908122503328895,
"grad_norm": 0.5147867202758789,
"learning_rate": 0.00010545938748335554,
"loss": 0.0024,
"step": 1420
},
{
"epoch": 1.904127829560586,
"grad_norm": 0.007088396232575178,
"learning_rate": 0.00010479360852197071,
"loss": 0.0016,
"step": 1430
},
{
"epoch": 1.9174434087882823,
"grad_norm": 0.07631803303956985,
"learning_rate": 0.00010412782956058589,
"loss": 0.0016,
"step": 1440
},
{
"epoch": 1.9307589880159788,
"grad_norm": 0.02045373059809208,
"learning_rate": 0.00010346205059920106,
"loss": 0.0062,
"step": 1450
},
{
"epoch": 1.9440745672436752,
"grad_norm": 0.034513406455516815,
"learning_rate": 0.00010279627163781624,
"loss": 0.0418,
"step": 1460
},
{
"epoch": 1.9573901464713717,
"grad_norm": 0.006663164123892784,
"learning_rate": 0.00010213049267643142,
"loss": 0.0355,
"step": 1470
},
{
"epoch": 1.9707057256990679,
"grad_norm": 0.15765492618083954,
"learning_rate": 0.0001014647137150466,
"loss": 0.0708,
"step": 1480
},
{
"epoch": 1.9840213049267643,
"grad_norm": 4.438633441925049,
"learning_rate": 0.0001007989347536618,
"loss": 0.0889,
"step": 1490
},
{
"epoch": 1.9973368841544608,
"grad_norm": 0.006967665161937475,
"learning_rate": 0.00010013315579227697,
"loss": 0.0102,
"step": 1500
},
{
"epoch": 1.9973368841544608,
"eval_accuracy": 0.9910179640718563,
"eval_loss": 0.034723859280347824,
"eval_runtime": 18.7741,
"eval_samples_per_second": 71.162,
"eval_steps_per_second": 8.895,
"step": 1500
},
{
"epoch": 2.010652463382157,
"grad_norm": 0.09451789408922195,
"learning_rate": 9.946737683089215e-05,
"loss": 0.0049,
"step": 1510
},
{
"epoch": 2.0239680426098534,
"grad_norm": 0.008569799363613129,
"learning_rate": 9.880159786950732e-05,
"loss": 0.0173,
"step": 1520
},
{
"epoch": 2.03728362183755,
"grad_norm": 0.3698938488960266,
"learning_rate": 9.813581890812251e-05,
"loss": 0.0013,
"step": 1530
},
{
"epoch": 2.0505992010652463,
"grad_norm": 0.016449321061372757,
"learning_rate": 9.747003994673769e-05,
"loss": 0.0015,
"step": 1540
},
{
"epoch": 2.0639147802929427,
"grad_norm": 0.06819990277290344,
"learning_rate": 9.680426098535287e-05,
"loss": 0.001,
"step": 1550
},
{
"epoch": 2.077230359520639,
"grad_norm": 0.006248504854738712,
"learning_rate": 9.613848202396804e-05,
"loss": 0.0011,
"step": 1560
},
{
"epoch": 2.0905459387483356,
"grad_norm": 0.005318405572324991,
"learning_rate": 9.547270306258322e-05,
"loss": 0.0277,
"step": 1570
},
{
"epoch": 2.103861517976032,
"grad_norm": 0.02135417051613331,
"learning_rate": 9.480692410119841e-05,
"loss": 0.0007,
"step": 1580
},
{
"epoch": 2.1171770972037285,
"grad_norm": 13.763138771057129,
"learning_rate": 9.414114513981358e-05,
"loss": 0.0163,
"step": 1590
},
{
"epoch": 2.130492676431425,
"grad_norm": 3.230034112930298,
"learning_rate": 9.347536617842877e-05,
"loss": 0.0668,
"step": 1600
},
{
"epoch": 2.130492676431425,
"eval_accuracy": 0.9745508982035929,
"eval_loss": 0.1236691027879715,
"eval_runtime": 18.9343,
"eval_samples_per_second": 70.56,
"eval_steps_per_second": 8.82,
"step": 1600
},
{
"epoch": 2.1438082556591214,
"grad_norm": 0.0055719888769090176,
"learning_rate": 9.280958721704395e-05,
"loss": 0.001,
"step": 1610
},
{
"epoch": 2.157123834886818,
"grad_norm": 0.007174134254455566,
"learning_rate": 9.214380825565913e-05,
"loss": 0.0353,
"step": 1620
},
{
"epoch": 2.170439414114514,
"grad_norm": 0.005854015704244375,
"learning_rate": 9.14780292942743e-05,
"loss": 0.0011,
"step": 1630
},
{
"epoch": 2.1837549933422102,
"grad_norm": 0.03955959528684616,
"learning_rate": 9.081225033288948e-05,
"loss": 0.0082,
"step": 1640
},
{
"epoch": 2.1970705725699067,
"grad_norm": 0.007123819552361965,
"learning_rate": 9.014647137150465e-05,
"loss": 0.0009,
"step": 1650
},
{
"epoch": 2.210386151797603,
"grad_norm": 0.15122471749782562,
"learning_rate": 8.948069241011984e-05,
"loss": 0.0025,
"step": 1660
},
{
"epoch": 2.2237017310252996,
"grad_norm": 0.0059026069939136505,
"learning_rate": 8.881491344873502e-05,
"loss": 0.0211,
"step": 1670
},
{
"epoch": 2.237017310252996,
"grad_norm": 0.005344110075384378,
"learning_rate": 8.814913448735021e-05,
"loss": 0.0215,
"step": 1680
},
{
"epoch": 2.2503328894806924,
"grad_norm": 0.007038838230073452,
"learning_rate": 8.748335552596539e-05,
"loss": 0.0012,
"step": 1690
},
{
"epoch": 2.263648468708389,
"grad_norm": 0.004509239457547665,
"learning_rate": 8.681757656458056e-05,
"loss": 0.0022,
"step": 1700
},
{
"epoch": 2.263648468708389,
"eval_accuracy": 0.9872754491017964,
"eval_loss": 0.0549810491502285,
"eval_runtime": 19.1819,
"eval_samples_per_second": 69.649,
"eval_steps_per_second": 8.706,
"step": 1700
},
{
"epoch": 2.2769640479360853,
"grad_norm": 0.04736039415001869,
"learning_rate": 8.615179760319574e-05,
"loss": 0.0015,
"step": 1710
},
{
"epoch": 2.2902796271637818,
"grad_norm": 6.543766498565674,
"learning_rate": 8.548601864181093e-05,
"loss": 0.0414,
"step": 1720
},
{
"epoch": 2.303595206391478,
"grad_norm": 0.004299995955079794,
"learning_rate": 8.48202396804261e-05,
"loss": 0.0008,
"step": 1730
},
{
"epoch": 2.316910785619174,
"grad_norm": 0.005761590786278248,
"learning_rate": 8.415446071904128e-05,
"loss": 0.0586,
"step": 1740
},
{
"epoch": 2.3302263648468706,
"grad_norm": 0.007165808696299791,
"learning_rate": 8.348868175765646e-05,
"loss": 0.0007,
"step": 1750
},
{
"epoch": 2.343541944074567,
"grad_norm": 0.0066890171729028225,
"learning_rate": 8.282290279627163e-05,
"loss": 0.0374,
"step": 1760
},
{
"epoch": 2.3568575233022635,
"grad_norm": 0.005914334207773209,
"learning_rate": 8.215712383488682e-05,
"loss": 0.0009,
"step": 1770
},
{
"epoch": 2.37017310252996,
"grad_norm": 0.009822635911405087,
"learning_rate": 8.1491344873502e-05,
"loss": 0.0364,
"step": 1780
},
{
"epoch": 2.3834886817576564,
"grad_norm": 0.008548504672944546,
"learning_rate": 8.082556591211719e-05,
"loss": 0.043,
"step": 1790
},
{
"epoch": 2.396804260985353,
"grad_norm": 0.007220366504043341,
"learning_rate": 8.015978695073236e-05,
"loss": 0.0156,
"step": 1800
},
{
"epoch": 2.396804260985353,
"eval_accuracy": 0.9895209580838323,
"eval_loss": 0.03286071494221687,
"eval_runtime": 19.1709,
"eval_samples_per_second": 69.689,
"eval_steps_per_second": 8.711,
"step": 1800
},
{
"epoch": 2.4101198402130493,
"grad_norm": 0.006717954762279987,
"learning_rate": 7.949400798934754e-05,
"loss": 0.0009,
"step": 1810
},
{
"epoch": 2.4234354194407457,
"grad_norm": 0.035514891147613525,
"learning_rate": 7.882822902796272e-05,
"loss": 0.0009,
"step": 1820
},
{
"epoch": 2.436750998668442,
"grad_norm": 0.0049186451360583305,
"learning_rate": 7.81624500665779e-05,
"loss": 0.0046,
"step": 1830
},
{
"epoch": 2.4500665778961386,
"grad_norm": 0.03821125254034996,
"learning_rate": 7.749667110519307e-05,
"loss": 0.0069,
"step": 1840
},
{
"epoch": 2.463382157123835,
"grad_norm": 3.3322513103485107,
"learning_rate": 7.683089214380826e-05,
"loss": 0.0022,
"step": 1850
},
{
"epoch": 2.4766977363515315,
"grad_norm": 0.004409165121614933,
"learning_rate": 7.616511318242345e-05,
"loss": 0.0023,
"step": 1860
},
{
"epoch": 2.490013315579228,
"grad_norm": 0.004281247034668922,
"learning_rate": 7.549933422103862e-05,
"loss": 0.0258,
"step": 1870
},
{
"epoch": 2.5033288948069243,
"grad_norm": 0.012024540454149246,
"learning_rate": 7.48335552596538e-05,
"loss": 0.0006,
"step": 1880
},
{
"epoch": 2.5166444740346208,
"grad_norm": 0.004248355980962515,
"learning_rate": 7.416777629826898e-05,
"loss": 0.0005,
"step": 1890
},
{
"epoch": 2.5299600532623168,
"grad_norm": 0.004529563244432211,
"learning_rate": 7.350199733688415e-05,
"loss": 0.0006,
"step": 1900
},
{
"epoch": 2.5299600532623168,
"eval_accuracy": 0.9947604790419161,
"eval_loss": 0.02548411302268505,
"eval_runtime": 18.997,
"eval_samples_per_second": 70.327,
"eval_steps_per_second": 8.791,
"step": 1900
},
{
"epoch": 2.543275632490013,
"grad_norm": 0.00455145537853241,
"learning_rate": 7.283621837549934e-05,
"loss": 0.0005,
"step": 1910
},
{
"epoch": 2.5565912117177096,
"grad_norm": 0.004023432265967131,
"learning_rate": 7.217043941411452e-05,
"loss": 0.0005,
"step": 1920
},
{
"epoch": 2.569906790945406,
"grad_norm": 0.004490096587687731,
"learning_rate": 7.15046604527297e-05,
"loss": 0.0007,
"step": 1930
},
{
"epoch": 2.5832223701731025,
"grad_norm": 0.0075076790526509285,
"learning_rate": 7.083888149134487e-05,
"loss": 0.0012,
"step": 1940
},
{
"epoch": 2.596537949400799,
"grad_norm": 0.0037563215009868145,
"learning_rate": 7.017310252996006e-05,
"loss": 0.0004,
"step": 1950
},
{
"epoch": 2.6098535286284954,
"grad_norm": 0.0038872328586876392,
"learning_rate": 6.950732356857524e-05,
"loss": 0.0405,
"step": 1960
},
{
"epoch": 2.623169107856192,
"grad_norm": 0.005599846597760916,
"learning_rate": 6.884154460719041e-05,
"loss": 0.0006,
"step": 1970
},
{
"epoch": 2.6364846870838883,
"grad_norm": 0.003918104339390993,
"learning_rate": 6.81757656458056e-05,
"loss": 0.0009,
"step": 1980
},
{
"epoch": 2.6498002663115847,
"grad_norm": 0.0037299636751413345,
"learning_rate": 6.750998668442078e-05,
"loss": 0.0004,
"step": 1990
},
{
"epoch": 2.6631158455392807,
"grad_norm": 0.00351333268918097,
"learning_rate": 6.684420772303596e-05,
"loss": 0.0004,
"step": 2000
},
{
"epoch": 2.6631158455392807,
"eval_accuracy": 0.9910179640718563,
"eval_loss": 0.044503793120384216,
"eval_runtime": 19.2349,
"eval_samples_per_second": 69.457,
"eval_steps_per_second": 8.682,
"step": 2000
},
{
"epoch": 2.676431424766977,
"grad_norm": 0.0037041015457361937,
"learning_rate": 6.617842876165113e-05,
"loss": 0.0005,
"step": 2010
},
{
"epoch": 2.6897470039946736,
"grad_norm": 0.0033594798296689987,
"learning_rate": 6.551264980026631e-05,
"loss": 0.0004,
"step": 2020
},
{
"epoch": 2.70306258322237,
"grad_norm": 0.0033981057349592447,
"learning_rate": 6.484687083888148e-05,
"loss": 0.0004,
"step": 2030
},
{
"epoch": 2.7163781624500665,
"grad_norm": 0.006549084093421698,
"learning_rate": 6.418109187749667e-05,
"loss": 0.0005,
"step": 2040
},
{
"epoch": 2.729693741677763,
"grad_norm": 0.0034731528721749783,
"learning_rate": 6.351531291611186e-05,
"loss": 0.0004,
"step": 2050
},
{
"epoch": 2.7430093209054593,
"grad_norm": 0.003631744533777237,
"learning_rate": 6.284953395472704e-05,
"loss": 0.0004,
"step": 2060
},
{
"epoch": 2.756324900133156,
"grad_norm": 0.00373335974290967,
"learning_rate": 6.218375499334222e-05,
"loss": 0.0004,
"step": 2070
},
{
"epoch": 2.7696404793608522,
"grad_norm": 0.003112897975370288,
"learning_rate": 6.151797603195739e-05,
"loss": 0.0004,
"step": 2080
},
{
"epoch": 2.7829560585885487,
"grad_norm": 0.0032353100832551718,
"learning_rate": 6.085219707057257e-05,
"loss": 0.0221,
"step": 2090
},
{
"epoch": 2.796271637816245,
"grad_norm": 0.003041923977434635,
"learning_rate": 6.018641810918775e-05,
"loss": 0.0011,
"step": 2100
},
{
"epoch": 2.796271637816245,
"eval_accuracy": 0.9805389221556886,
"eval_loss": 0.09204710274934769,
"eval_runtime": 18.9463,
"eval_samples_per_second": 70.515,
"eval_steps_per_second": 8.814,
"step": 2100
},
{
"epoch": 2.8095872170439415,
"grad_norm": 0.003456325735896826,
"learning_rate": 5.9520639147802933e-05,
"loss": 0.0416,
"step": 2110
},
{
"epoch": 2.822902796271638,
"grad_norm": 0.004117018077522516,
"learning_rate": 5.8854860186418116e-05,
"loss": 0.0004,
"step": 2120
},
{
"epoch": 2.8362183754993344,
"grad_norm": 0.00375298666767776,
"learning_rate": 5.818908122503329e-05,
"loss": 0.0004,
"step": 2130
},
{
"epoch": 2.849533954727031,
"grad_norm": 4.614629745483398,
"learning_rate": 5.752330226364847e-05,
"loss": 0.0117,
"step": 2140
},
{
"epoch": 2.8628495339547273,
"grad_norm": 0.006098424084484577,
"learning_rate": 5.685752330226365e-05,
"loss": 0.001,
"step": 2150
},
{
"epoch": 2.8761651131824233,
"grad_norm": 0.20828425884246826,
"learning_rate": 5.619174434087883e-05,
"loss": 0.0007,
"step": 2160
},
{
"epoch": 2.8894806924101197,
"grad_norm": 0.0036955540999770164,
"learning_rate": 5.552596537949402e-05,
"loss": 0.0004,
"step": 2170
},
{
"epoch": 2.902796271637816,
"grad_norm": 0.003313822206109762,
"learning_rate": 5.4860186418109194e-05,
"loss": 0.0004,
"step": 2180
},
{
"epoch": 2.9161118508655126,
"grad_norm": 0.0031216980423778296,
"learning_rate": 5.419440745672437e-05,
"loss": 0.0007,
"step": 2190
},
{
"epoch": 2.929427430093209,
"grad_norm": 0.002954344032332301,
"learning_rate": 5.352862849533955e-05,
"loss": 0.0004,
"step": 2200
},
{
"epoch": 2.929427430093209,
"eval_accuracy": 0.9917664670658682,
"eval_loss": 0.04039301723241806,
"eval_runtime": 18.9671,
"eval_samples_per_second": 70.438,
"eval_steps_per_second": 8.805,
"step": 2200
},
{
"epoch": 2.9427430093209055,
"grad_norm": 107.526123046875,
"learning_rate": 5.286284953395473e-05,
"loss": 0.0429,
"step": 2210
},
{
"epoch": 2.956058588548602,
"grad_norm": 0.0032062442041933537,
"learning_rate": 5.2197070572569905e-05,
"loss": 0.0004,
"step": 2220
},
{
"epoch": 2.9693741677762984,
"grad_norm": 0.2084268033504486,
"learning_rate": 5.153129161118508e-05,
"loss": 0.015,
"step": 2230
},
{
"epoch": 2.982689747003995,
"grad_norm": 0.004362909123301506,
"learning_rate": 5.086551264980027e-05,
"loss": 0.0005,
"step": 2240
},
{
"epoch": 2.996005326231691,
"grad_norm": 0.0029051396995782852,
"learning_rate": 5.0199733688415454e-05,
"loss": 0.0008,
"step": 2250
},
{
"epoch": 3.0093209054593877,
"grad_norm": 0.0038269031792879105,
"learning_rate": 4.953395472703063e-05,
"loss": 0.0007,
"step": 2260
},
{
"epoch": 3.0226364846870837,
"grad_norm": 0.0031527807004749775,
"learning_rate": 4.8868175765645806e-05,
"loss": 0.0004,
"step": 2270
},
{
"epoch": 3.03595206391478,
"grad_norm": 0.0027656673919409513,
"learning_rate": 4.820239680426098e-05,
"loss": 0.0003,
"step": 2280
},
{
"epoch": 3.0492676431424766,
"grad_norm": 0.0027427240274846554,
"learning_rate": 4.753661784287617e-05,
"loss": 0.0004,
"step": 2290
},
{
"epoch": 3.062583222370173,
"grad_norm": 0.004418679978698492,
"learning_rate": 4.687083888149135e-05,
"loss": 0.0003,
"step": 2300
},
{
"epoch": 3.062583222370173,
"eval_accuracy": 0.9917664670658682,
"eval_loss": 0.037005726248025894,
"eval_runtime": 18.852,
"eval_samples_per_second": 70.868,
"eval_steps_per_second": 8.858,
"step": 2300
},
{
"epoch": 3.0758988015978694,
"grad_norm": 0.003011292079463601,
"learning_rate": 4.6205059920106524e-05,
"loss": 0.0005,
"step": 2310
},
{
"epoch": 3.089214380825566,
"grad_norm": 0.002671412192285061,
"learning_rate": 4.553928095872171e-05,
"loss": 0.0003,
"step": 2320
},
{
"epoch": 3.1025299600532623,
"grad_norm": 0.002712644636631012,
"learning_rate": 4.487350199733688e-05,
"loss": 0.0003,
"step": 2330
},
{
"epoch": 3.1158455392809588,
"grad_norm": 0.00504153361544013,
"learning_rate": 4.4207723035952066e-05,
"loss": 0.0003,
"step": 2340
},
{
"epoch": 3.129161118508655,
"grad_norm": 0.003144737333059311,
"learning_rate": 4.354194407456725e-05,
"loss": 0.0003,
"step": 2350
},
{
"epoch": 3.1424766977363516,
"grad_norm": 0.002637204946950078,
"learning_rate": 4.2876165113182425e-05,
"loss": 0.0003,
"step": 2360
},
{
"epoch": 3.155792276964048,
"grad_norm": 0.0026598216500133276,
"learning_rate": 4.22103861517976e-05,
"loss": 0.0003,
"step": 2370
},
{
"epoch": 3.1691078561917445,
"grad_norm": 0.002664501080289483,
"learning_rate": 4.154460719041279e-05,
"loss": 0.0003,
"step": 2380
},
{
"epoch": 3.182423435419441,
"grad_norm": 0.0027599988970905542,
"learning_rate": 4.087882822902797e-05,
"loss": 0.0003,
"step": 2390
},
{
"epoch": 3.195739014647137,
"grad_norm": 0.002599115017801523,
"learning_rate": 4.021304926764314e-05,
"loss": 0.0003,
"step": 2400
},
{
"epoch": 3.195739014647137,
"eval_accuracy": 0.9917664670658682,
"eval_loss": 0.03919079899787903,
"eval_runtime": 18.8696,
"eval_samples_per_second": 70.802,
"eval_steps_per_second": 8.85,
"step": 2400
},
{
"epoch": 3.2090545938748334,
"grad_norm": 0.0025887268129736185,
"learning_rate": 3.954727030625832e-05,
"loss": 0.0163,
"step": 2410
},
{
"epoch": 3.22237017310253,
"grad_norm": 0.0027751659508794546,
"learning_rate": 3.88814913448735e-05,
"loss": 0.0003,
"step": 2420
},
{
"epoch": 3.2356857523302263,
"grad_norm": 0.002749471226707101,
"learning_rate": 3.8215712383488685e-05,
"loss": 0.0003,
"step": 2430
},
{
"epoch": 3.2490013315579227,
"grad_norm": 0.002491929102689028,
"learning_rate": 3.754993342210386e-05,
"loss": 0.0012,
"step": 2440
},
{
"epoch": 3.262316910785619,
"grad_norm": 0.0031888161320239305,
"learning_rate": 3.6884154460719044e-05,
"loss": 0.0003,
"step": 2450
},
{
"epoch": 3.2756324900133156,
"grad_norm": 0.004251156002283096,
"learning_rate": 3.621837549933422e-05,
"loss": 0.0003,
"step": 2460
},
{
"epoch": 3.288948069241012,
"grad_norm": 0.0027103093452751637,
"learning_rate": 3.55525965379494e-05,
"loss": 0.0004,
"step": 2470
},
{
"epoch": 3.3022636484687085,
"grad_norm": 0.0026021709199994802,
"learning_rate": 3.4886817576564586e-05,
"loss": 0.0003,
"step": 2480
},
{
"epoch": 3.315579227696405,
"grad_norm": 0.002954940777271986,
"learning_rate": 3.422103861517976e-05,
"loss": 0.0003,
"step": 2490
},
{
"epoch": 3.3288948069241013,
"grad_norm": 0.0026502846740186214,
"learning_rate": 3.355525965379494e-05,
"loss": 0.0003,
"step": 2500
},
{
"epoch": 3.3288948069241013,
"eval_accuracy": 0.9932634730538922,
"eval_loss": 0.03057168610394001,
"eval_runtime": 19.2139,
"eval_samples_per_second": 69.533,
"eval_steps_per_second": 8.692,
"step": 2500
},
{
"epoch": 3.3422103861517978,
"grad_norm": 0.0026077169459313154,
"learning_rate": 3.288948069241012e-05,
"loss": 0.0003,
"step": 2510
},
{
"epoch": 3.3555259653794938,
"grad_norm": 0.002518726047128439,
"learning_rate": 3.2223701731025304e-05,
"loss": 0.0003,
"step": 2520
},
{
"epoch": 3.36884154460719,
"grad_norm": 0.002446199534460902,
"learning_rate": 3.155792276964048e-05,
"loss": 0.0003,
"step": 2530
},
{
"epoch": 3.3821571238348866,
"grad_norm": 0.01649678498506546,
"learning_rate": 3.0892143808255656e-05,
"loss": 0.0003,
"step": 2540
},
{
"epoch": 3.395472703062583,
"grad_norm": 0.002374894917011261,
"learning_rate": 3.0226364846870843e-05,
"loss": 0.0003,
"step": 2550
},
{
"epoch": 3.4087882822902795,
"grad_norm": 0.0023247618228197098,
"learning_rate": 2.956058588548602e-05,
"loss": 0.0003,
"step": 2560
},
{
"epoch": 3.422103861517976,
"grad_norm": 0.0022955993190407753,
"learning_rate": 2.88948069241012e-05,
"loss": 0.0003,
"step": 2570
},
{
"epoch": 3.4354194407456724,
"grad_norm": 0.0023930929601192474,
"learning_rate": 2.822902796271638e-05,
"loss": 0.0194,
"step": 2580
},
{
"epoch": 3.448735019973369,
"grad_norm": 0.0025017596781253815,
"learning_rate": 2.756324900133156e-05,
"loss": 0.0003,
"step": 2590
},
{
"epoch": 3.4620505992010653,
"grad_norm": 0.0025145343970507383,
"learning_rate": 2.6897470039946737e-05,
"loss": 0.0003,
"step": 2600
},
{
"epoch": 3.4620505992010653,
"eval_accuracy": 0.9955089820359282,
"eval_loss": 0.019898300990462303,
"eval_runtime": 19.1238,
"eval_samples_per_second": 69.861,
"eval_steps_per_second": 8.733,
"step": 2600
},
{
"epoch": 3.4753661784287617,
"grad_norm": 0.0024944059550762177,
"learning_rate": 2.623169107856192e-05,
"loss": 0.0003,
"step": 2610
},
{
"epoch": 3.488681757656458,
"grad_norm": 0.002342221327126026,
"learning_rate": 2.55659121171771e-05,
"loss": 0.0005,
"step": 2620
},
{
"epoch": 3.5019973368841546,
"grad_norm": 0.03469083085656166,
"learning_rate": 2.4900133155792276e-05,
"loss": 0.0003,
"step": 2630
},
{
"epoch": 3.515312916111851,
"grad_norm": 0.0023358033504337072,
"learning_rate": 2.423435419440746e-05,
"loss": 0.0003,
"step": 2640
},
{
"epoch": 3.5286284953395475,
"grad_norm": 0.0022878609597682953,
"learning_rate": 2.3568575233022638e-05,
"loss": 0.0003,
"step": 2650
},
{
"epoch": 3.541944074567244,
"grad_norm": 0.0026905399281531572,
"learning_rate": 2.2902796271637818e-05,
"loss": 0.0003,
"step": 2660
},
{
"epoch": 3.5552596537949404,
"grad_norm": 0.0026699311565607786,
"learning_rate": 2.2237017310252997e-05,
"loss": 0.0003,
"step": 2670
},
{
"epoch": 3.5685752330226364,
"grad_norm": 0.0023570535704493523,
"learning_rate": 2.1571238348868177e-05,
"loss": 0.0003,
"step": 2680
},
{
"epoch": 3.581890812250333,
"grad_norm": 0.0024163059424608946,
"learning_rate": 2.0905459387483356e-05,
"loss": 0.0003,
"step": 2690
},
{
"epoch": 3.5952063914780292,
"grad_norm": 0.002381423255428672,
"learning_rate": 2.0239680426098536e-05,
"loss": 0.0003,
"step": 2700
},
{
"epoch": 3.5952063914780292,
"eval_accuracy": 0.9947604790419161,
"eval_loss": 0.024509532377123833,
"eval_runtime": 19.0731,
"eval_samples_per_second": 70.046,
"eval_steps_per_second": 8.756,
"step": 2700
},
{
"epoch": 3.6085219707057257,
"grad_norm": 0.0022262686397880316,
"learning_rate": 1.9573901464713715e-05,
"loss": 0.0003,
"step": 2710
},
{
"epoch": 3.621837549933422,
"grad_norm": 0.0022425020579248667,
"learning_rate": 1.8908122503328895e-05,
"loss": 0.0003,
"step": 2720
},
{
"epoch": 3.6351531291611185,
"grad_norm": 0.0022420468740165234,
"learning_rate": 1.8242343541944078e-05,
"loss": 0.0003,
"step": 2730
},
{
"epoch": 3.648468708388815,
"grad_norm": 0.0021876932587474585,
"learning_rate": 1.7576564580559254e-05,
"loss": 0.0003,
"step": 2740
},
{
"epoch": 3.6617842876165114,
"grad_norm": 0.0023544211871922016,
"learning_rate": 1.6910785619174437e-05,
"loss": 0.0002,
"step": 2750
},
{
"epoch": 3.675099866844208,
"grad_norm": 0.002206595614552498,
"learning_rate": 1.6245006657789616e-05,
"loss": 0.0004,
"step": 2760
},
{
"epoch": 3.688415446071904,
"grad_norm": 0.002322066342458129,
"learning_rate": 1.5579227696404792e-05,
"loss": 0.0002,
"step": 2770
},
{
"epoch": 3.7017310252996003,
"grad_norm": 0.0022114135790616274,
"learning_rate": 1.4913448735019975e-05,
"loss": 0.0003,
"step": 2780
},
{
"epoch": 3.7150466045272967,
"grad_norm": 0.0022223354317247868,
"learning_rate": 1.4247669773635153e-05,
"loss": 0.0004,
"step": 2790
},
{
"epoch": 3.728362183754993,
"grad_norm": 0.002155436435714364,
"learning_rate": 1.3581890812250334e-05,
"loss": 0.0003,
"step": 2800
},
{
"epoch": 3.728362183754993,
"eval_accuracy": 0.9947604790419161,
"eval_loss": 0.025128666311502457,
"eval_runtime": 18.7023,
"eval_samples_per_second": 71.435,
"eval_steps_per_second": 8.929,
"step": 2800
},
{
"epoch": 3.7416777629826896,
"grad_norm": 0.0022344435565173626,
"learning_rate": 1.2916111850865514e-05,
"loss": 0.0003,
"step": 2810
},
{
"epoch": 3.754993342210386,
"grad_norm": 0.04637530446052551,
"learning_rate": 1.2250332889480692e-05,
"loss": 0.0003,
"step": 2820
},
{
"epoch": 3.7683089214380825,
"grad_norm": 0.002139828633517027,
"learning_rate": 1.1584553928095873e-05,
"loss": 0.0003,
"step": 2830
},
{
"epoch": 3.781624500665779,
"grad_norm": 0.0020825008396059275,
"learning_rate": 1.0918774966711052e-05,
"loss": 0.0003,
"step": 2840
},
{
"epoch": 3.7949400798934754,
"grad_norm": 0.0020064804702997208,
"learning_rate": 1.0252996005326232e-05,
"loss": 0.0003,
"step": 2850
},
{
"epoch": 3.808255659121172,
"grad_norm": 0.002234045183286071,
"learning_rate": 9.587217043941411e-06,
"loss": 0.0002,
"step": 2860
},
{
"epoch": 3.8215712383488682,
"grad_norm": 0.0021161320619285107,
"learning_rate": 8.921438082556593e-06,
"loss": 0.0003,
"step": 2870
},
{
"epoch": 3.8348868175765647,
"grad_norm": 0.0022624481935054064,
"learning_rate": 8.255659121171772e-06,
"loss": 0.0002,
"step": 2880
},
{
"epoch": 3.848202396804261,
"grad_norm": 0.0020989372860640287,
"learning_rate": 7.589880159786951e-06,
"loss": 0.0002,
"step": 2890
},
{
"epoch": 3.8615179760319576,
"grad_norm": 0.002201867988333106,
"learning_rate": 6.92410119840213e-06,
"loss": 0.0002,
"step": 2900
},
{
"epoch": 3.8615179760319576,
"eval_accuracy": 0.9947604790419161,
"eval_loss": 0.025865301489830017,
"eval_runtime": 18.6017,
"eval_samples_per_second": 71.821,
"eval_steps_per_second": 8.978,
"step": 2900
},
{
"epoch": 3.874833555259654,
"grad_norm": 0.002228468656539917,
"learning_rate": 6.258322237017311e-06,
"loss": 0.0002,
"step": 2910
},
{
"epoch": 3.8881491344873504,
"grad_norm": 0.0022046160884201527,
"learning_rate": 5.59254327563249e-06,
"loss": 0.0003,
"step": 2920
},
{
"epoch": 3.9014647137150464,
"grad_norm": 0.002218181500211358,
"learning_rate": 4.92676431424767e-06,
"loss": 0.0002,
"step": 2930
},
{
"epoch": 3.914780292942743,
"grad_norm": 0.001987821888178587,
"learning_rate": 4.26098535286285e-06,
"loss": 0.0002,
"step": 2940
},
{
"epoch": 3.9280958721704393,
"grad_norm": 0.0021837118547409773,
"learning_rate": 3.5952063914780293e-06,
"loss": 0.0002,
"step": 2950
},
{
"epoch": 3.9414114513981358,
"grad_norm": 0.002148882020264864,
"learning_rate": 2.9294274300932092e-06,
"loss": 0.0002,
"step": 2960
},
{
"epoch": 3.954727030625832,
"grad_norm": 0.002326791873201728,
"learning_rate": 2.2636484687083888e-06,
"loss": 0.0002,
"step": 2970
},
{
"epoch": 3.9680426098535286,
"grad_norm": 0.001993759535253048,
"learning_rate": 1.5978695073235687e-06,
"loss": 0.0002,
"step": 2980
},
{
"epoch": 3.981358189081225,
"grad_norm": 0.0020522773265838623,
"learning_rate": 9.320905459387485e-07,
"loss": 0.0002,
"step": 2990
},
{
"epoch": 3.9946737683089215,
"grad_norm": 0.0021599766332656145,
"learning_rate": 2.6631158455392814e-07,
"loss": 0.0002,
"step": 3000
},
{
"epoch": 3.9946737683089215,
"eval_accuracy": 0.9947604790419161,
"eval_loss": 0.02588406577706337,
"eval_runtime": 18.7868,
"eval_samples_per_second": 71.114,
"eval_steps_per_second": 8.889,
"step": 3000
}
],
"logging_steps": 10,
"max_steps": 3004,
"num_input_tokens_seen": 0,
"num_train_epochs": 4,
"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": 3.7242650208772915e+18,
"train_batch_size": 16,
"trial_name": null,
"trial_params": null
}