diff --git "a/resemble_to_drop_frequency_40817/checkpoint-40000/trainer_state.json" "b/resemble_to_drop_frequency_40817/checkpoint-40000/trainer_state.json" new file mode 100644--- /dev/null +++ "b/resemble_to_drop_frequency_40817/checkpoint-40000/trainer_state.json" @@ -0,0 +1,6003 @@ +{ + "best_global_step": 40000, + "best_metric": 3.5598602294921875, + "best_model_checkpoint": "/scratch/cl5625/exceptions/models/resemble_to_drop_frequency_40817/checkpoint-40000", + "epoch": 11.648261401362921, + "eval_steps": 1000, + "global_step": 40000, + "is_hyper_param_search": false, + "is_local_process_zero": true, + "is_world_process_zero": true, + "log_history": [ + { + "epoch": 0.014561127613722406, + "grad_norm": 0.7783309817314148, + "learning_rate": 0.000294, + "loss": 8.4595, + "step": 50 + }, + { + "epoch": 0.029122255227444813, + "grad_norm": 0.5112840533256531, + "learning_rate": 0.0005939999999999999, + "loss": 6.7353, + "step": 100 + }, + { + "epoch": 0.04368338284116722, + "grad_norm": 0.5744251608848572, + "learning_rate": 0.0005998286713286713, + "loss": 6.3362, + "step": 150 + }, + { + "epoch": 0.058244510454889625, + "grad_norm": 0.5645453333854675, + "learning_rate": 0.0005996538461538461, + "loss": 6.1367, + "step": 200 + }, + { + "epoch": 0.07280563806861204, + "grad_norm": 0.44470539689064026, + "learning_rate": 0.0005994790209790209, + "loss": 5.9852, + "step": 250 + }, + { + "epoch": 0.08736676568233444, + "grad_norm": 0.4577213525772095, + "learning_rate": 0.0005993041958041958, + "loss": 5.8404, + "step": 300 + }, + { + "epoch": 0.10192789329605685, + "grad_norm": 0.4522819519042969, + "learning_rate": 0.0005991293706293705, + "loss": 5.7217, + "step": 350 + }, + { + "epoch": 0.11648902090977925, + "grad_norm": 0.5096995234489441, + "learning_rate": 0.0005989545454545454, + "loss": 5.6136, + "step": 400 + }, + { + "epoch": 0.13105014852350166, + "grad_norm": 0.5080946087837219, + "learning_rate": 0.0005987797202797202, + "loss": 5.4938, + "step": 450 + }, + { + "epoch": 0.14561127613722408, + "grad_norm": 0.46365445852279663, + "learning_rate": 0.000598604895104895, + "loss": 5.3922, + "step": 500 + }, + { + "epoch": 0.16017240375094646, + "grad_norm": 0.47090452909469604, + "learning_rate": 0.0005984300699300698, + "loss": 5.3192, + "step": 550 + }, + { + "epoch": 0.17473353136466888, + "grad_norm": 0.44763490557670593, + "learning_rate": 0.0005982552447552447, + "loss": 5.2508, + "step": 600 + }, + { + "epoch": 0.1892946589783913, + "grad_norm": 0.41502994298934937, + "learning_rate": 0.0005980804195804195, + "loss": 5.17, + "step": 650 + }, + { + "epoch": 0.2038557865921137, + "grad_norm": 0.4346926510334015, + "learning_rate": 0.0005979055944055943, + "loss": 5.1188, + "step": 700 + }, + { + "epoch": 0.2184169142058361, + "grad_norm": 0.4056375324726105, + "learning_rate": 0.0005977307692307691, + "loss": 5.0473, + "step": 750 + }, + { + "epoch": 0.2329780418195585, + "grad_norm": 0.40403687953948975, + "learning_rate": 0.000597555944055944, + "loss": 5.0269, + "step": 800 + }, + { + "epoch": 0.24753916943328091, + "grad_norm": 0.4618856906890869, + "learning_rate": 0.0005973811188811188, + "loss": 4.9739, + "step": 850 + }, + { + "epoch": 0.2621002970470033, + "grad_norm": 0.4969853460788727, + "learning_rate": 0.0005972062937062936, + "loss": 4.9138, + "step": 900 + }, + { + "epoch": 0.27666142466072574, + "grad_norm": 0.4236414134502411, + "learning_rate": 0.0005970314685314685, + "loss": 4.8832, + "step": 950 + }, + { + "epoch": 0.29122255227444815, + "grad_norm": 0.4577479958534241, + "learning_rate": 0.0005968566433566433, + "loss": 4.8224, + "step": 1000 + }, + { + "epoch": 0.29122255227444815, + "eval_accuracy": 0.25555888357699036, + "eval_loss": 4.742461204528809, + "eval_runtime": 179.7598, + "eval_samples_per_second": 92.574, + "eval_steps_per_second": 5.791, + "step": 1000 + }, + { + "epoch": 0.30578367988817057, + "grad_norm": 0.506449818611145, + "learning_rate": 0.0005966818181818181, + "loss": 4.7668, + "step": 1050 + }, + { + "epoch": 0.3203448075018929, + "grad_norm": 0.4153960049152374, + "learning_rate": 0.0005965069930069929, + "loss": 4.7367, + "step": 1100 + }, + { + "epoch": 0.33490593511561534, + "grad_norm": 0.45366883277893066, + "learning_rate": 0.0005963321678321677, + "loss": 4.6976, + "step": 1150 + }, + { + "epoch": 0.34946706272933775, + "grad_norm": 0.42683732509613037, + "learning_rate": 0.0005961573426573425, + "loss": 4.6572, + "step": 1200 + }, + { + "epoch": 0.36402819034306017, + "grad_norm": 0.4945414662361145, + "learning_rate": 0.0005959825174825174, + "loss": 4.6279, + "step": 1250 + }, + { + "epoch": 0.3785893179567826, + "grad_norm": 0.3974955081939697, + "learning_rate": 0.0005958076923076922, + "loss": 4.5904, + "step": 1300 + }, + { + "epoch": 0.393150445570505, + "grad_norm": 0.4485011100769043, + "learning_rate": 0.000595632867132867, + "loss": 4.584, + "step": 1350 + }, + { + "epoch": 0.4077115731842274, + "grad_norm": 0.459290087223053, + "learning_rate": 0.0005954580419580418, + "loss": 4.5296, + "step": 1400 + }, + { + "epoch": 0.4222727007979498, + "grad_norm": 0.4126909375190735, + "learning_rate": 0.0005952832167832168, + "loss": 4.5209, + "step": 1450 + }, + { + "epoch": 0.4368338284116722, + "grad_norm": 0.44859471917152405, + "learning_rate": 0.0005951083916083916, + "loss": 4.4935, + "step": 1500 + }, + { + "epoch": 0.4513949560253946, + "grad_norm": 0.4654662311077118, + "learning_rate": 0.0005949335664335664, + "loss": 4.4779, + "step": 1550 + }, + { + "epoch": 0.465956083639117, + "grad_norm": 0.41490229964256287, + "learning_rate": 0.0005947587412587413, + "loss": 4.4651, + "step": 1600 + }, + { + "epoch": 0.4805172112528394, + "grad_norm": 0.4170430302619934, + "learning_rate": 0.0005945839160839161, + "loss": 4.4403, + "step": 1650 + }, + { + "epoch": 0.49507833886656183, + "grad_norm": 0.3953550159931183, + "learning_rate": 0.0005944090909090909, + "loss": 4.4209, + "step": 1700 + }, + { + "epoch": 0.5096394664802842, + "grad_norm": 0.3975103199481964, + "learning_rate": 0.0005942342657342657, + "loss": 4.421, + "step": 1750 + }, + { + "epoch": 0.5242005940940067, + "grad_norm": 0.39017555117607117, + "learning_rate": 0.0005940594405594406, + "loss": 4.3957, + "step": 1800 + }, + { + "epoch": 0.5387617217077291, + "grad_norm": 0.3636375963687897, + "learning_rate": 0.0005938846153846153, + "loss": 4.375, + "step": 1850 + }, + { + "epoch": 0.5533228493214515, + "grad_norm": 0.39028239250183105, + "learning_rate": 0.0005937097902097902, + "loss": 4.3758, + "step": 1900 + }, + { + "epoch": 0.5678839769351739, + "grad_norm": 0.4079553186893463, + "learning_rate": 0.000593534965034965, + "loss": 4.3576, + "step": 1950 + }, + { + "epoch": 0.5824451045488963, + "grad_norm": 0.4021719694137573, + "learning_rate": 0.0005933601398601398, + "loss": 4.3378, + "step": 2000 + }, + { + "epoch": 0.5824451045488963, + "eval_accuracy": 0.2992665064277265, + "eval_loss": 4.2818121910095215, + "eval_runtime": 179.136, + "eval_samples_per_second": 92.896, + "eval_steps_per_second": 5.811, + "step": 2000 + }, + { + "epoch": 0.5970062321626187, + "grad_norm": 0.39577916264533997, + "learning_rate": 0.0005931853146853146, + "loss": 4.3205, + "step": 2050 + }, + { + "epoch": 0.6115673597763411, + "grad_norm": 0.40457814931869507, + "learning_rate": 0.0005930104895104895, + "loss": 4.2945, + "step": 2100 + }, + { + "epoch": 0.6261284873900634, + "grad_norm": 0.3639944791793823, + "learning_rate": 0.0005928356643356643, + "loss": 4.2936, + "step": 2150 + }, + { + "epoch": 0.6406896150037859, + "grad_norm": 0.3751417100429535, + "learning_rate": 0.0005926608391608391, + "loss": 4.2896, + "step": 2200 + }, + { + "epoch": 0.6552507426175083, + "grad_norm": 0.4348156452178955, + "learning_rate": 0.000592486013986014, + "loss": 4.2753, + "step": 2250 + }, + { + "epoch": 0.6698118702312307, + "grad_norm": 0.41849884390830994, + "learning_rate": 0.0005923111888111888, + "loss": 4.2645, + "step": 2300 + }, + { + "epoch": 0.6843729978449531, + "grad_norm": 0.44262415170669556, + "learning_rate": 0.0005921363636363636, + "loss": 4.2756, + "step": 2350 + }, + { + "epoch": 0.6989341254586755, + "grad_norm": 0.36123883724212646, + "learning_rate": 0.0005919615384615384, + "loss": 4.2507, + "step": 2400 + }, + { + "epoch": 0.7134952530723979, + "grad_norm": 0.39199990034103394, + "learning_rate": 0.0005917867132867133, + "loss": 4.2298, + "step": 2450 + }, + { + "epoch": 0.7280563806861203, + "grad_norm": 0.40468963980674744, + "learning_rate": 0.0005916118881118881, + "loss": 4.237, + "step": 2500 + }, + { + "epoch": 0.7426175082998427, + "grad_norm": 0.34999388456344604, + "learning_rate": 0.0005914370629370629, + "loss": 4.2304, + "step": 2550 + }, + { + "epoch": 0.7571786359135652, + "grad_norm": 0.36491668224334717, + "learning_rate": 0.0005912622377622377, + "loss": 4.2157, + "step": 2600 + }, + { + "epoch": 0.7717397635272876, + "grad_norm": 0.33732375502586365, + "learning_rate": 0.0005910874125874125, + "loss": 4.208, + "step": 2650 + }, + { + "epoch": 0.78630089114101, + "grad_norm": 0.3434152901172638, + "learning_rate": 0.0005909125874125873, + "loss": 4.2131, + "step": 2700 + }, + { + "epoch": 0.8008620187547324, + "grad_norm": 0.36287346482276917, + "learning_rate": 0.0005907377622377622, + "loss": 4.1885, + "step": 2750 + }, + { + "epoch": 0.8154231463684548, + "grad_norm": 0.37600141763687134, + "learning_rate": 0.000590562937062937, + "loss": 4.172, + "step": 2800 + }, + { + "epoch": 0.8299842739821772, + "grad_norm": 0.3740973174571991, + "learning_rate": 0.0005903881118881118, + "loss": 4.1696, + "step": 2850 + }, + { + "epoch": 0.8445454015958996, + "grad_norm": 0.36323437094688416, + "learning_rate": 0.0005902132867132867, + "loss": 4.1522, + "step": 2900 + }, + { + "epoch": 0.8591065292096219, + "grad_norm": 0.3623560667037964, + "learning_rate": 0.0005900384615384615, + "loss": 4.156, + "step": 2950 + }, + { + "epoch": 0.8736676568233444, + "grad_norm": 0.3544504940509796, + "learning_rate": 0.0005898636363636363, + "loss": 4.1566, + "step": 3000 + }, + { + "epoch": 0.8736676568233444, + "eval_accuracy": 0.31516880418545146, + "eval_loss": 4.097716331481934, + "eval_runtime": 179.1957, + "eval_samples_per_second": 92.865, + "eval_steps_per_second": 5.809, + "step": 3000 + }, + { + "epoch": 0.8882287844370668, + "grad_norm": 0.32925766706466675, + "learning_rate": 0.0005896888111888111, + "loss": 4.1427, + "step": 3050 + }, + { + "epoch": 0.9027899120507892, + "grad_norm": 0.3407750725746155, + "learning_rate": 0.000589513986013986, + "loss": 4.1328, + "step": 3100 + }, + { + "epoch": 0.9173510396645116, + "grad_norm": 0.3569066524505615, + "learning_rate": 0.0005893391608391608, + "loss": 4.1246, + "step": 3150 + }, + { + "epoch": 0.931912167278234, + "grad_norm": 0.33867526054382324, + "learning_rate": 0.0005891643356643356, + "loss": 4.1065, + "step": 3200 + }, + { + "epoch": 0.9464732948919564, + "grad_norm": 0.39542922377586365, + "learning_rate": 0.0005889895104895104, + "loss": 4.0983, + "step": 3250 + }, + { + "epoch": 0.9610344225056788, + "grad_norm": 0.35491493344306946, + "learning_rate": 0.0005888146853146853, + "loss": 4.098, + "step": 3300 + }, + { + "epoch": 0.9755955501194012, + "grad_norm": 0.37428122758865356, + "learning_rate": 0.00058863986013986, + "loss": 4.1027, + "step": 3350 + }, + { + "epoch": 0.9901566777331237, + "grad_norm": 0.33943089842796326, + "learning_rate": 0.0005884650349650349, + "loss": 4.1025, + "step": 3400 + }, + { + "epoch": 1.004659560836391, + "grad_norm": 0.33930638432502747, + "learning_rate": 0.0005882902097902097, + "loss": 4.0763, + "step": 3450 + }, + { + "epoch": 1.0192206884501136, + "grad_norm": 0.37376904487609863, + "learning_rate": 0.0005881153846153845, + "loss": 4.0184, + "step": 3500 + }, + { + "epoch": 1.033781816063836, + "grad_norm": 0.3479885160923004, + "learning_rate": 0.0005879405594405594, + "loss": 4.0166, + "step": 3550 + }, + { + "epoch": 1.0483429436775584, + "grad_norm": 0.36575818061828613, + "learning_rate": 0.0005877657342657342, + "loss": 4.0222, + "step": 3600 + }, + { + "epoch": 1.0629040712912807, + "grad_norm": 0.3828834593296051, + "learning_rate": 0.000587590909090909, + "loss": 4.019, + "step": 3650 + }, + { + "epoch": 1.0774651989050033, + "grad_norm": 0.34358036518096924, + "learning_rate": 0.0005874160839160838, + "loss": 4.0121, + "step": 3700 + }, + { + "epoch": 1.0920263265187256, + "grad_norm": 0.3743741810321808, + "learning_rate": 0.0005872412587412587, + "loss": 4.0167, + "step": 3750 + }, + { + "epoch": 1.106587454132448, + "grad_norm": 0.33027392625808716, + "learning_rate": 0.0005870664335664335, + "loss": 3.9893, + "step": 3800 + }, + { + "epoch": 1.1211485817461704, + "grad_norm": 0.3514248728752136, + "learning_rate": 0.0005868916083916083, + "loss": 3.9971, + "step": 3850 + }, + { + "epoch": 1.135709709359893, + "grad_norm": 0.33023524284362793, + "learning_rate": 0.0005867167832167831, + "loss": 4.0104, + "step": 3900 + }, + { + "epoch": 1.1502708369736152, + "grad_norm": 0.3593249022960663, + "learning_rate": 0.000586541958041958, + "loss": 4.006, + "step": 3950 + }, + { + "epoch": 1.1648319645873377, + "grad_norm": 0.3403565287590027, + "learning_rate": 0.0005863671328671328, + "loss": 4.0035, + "step": 4000 + }, + { + "epoch": 1.1648319645873377, + "eval_accuracy": 0.32484229235527606, + "eval_loss": 3.993455648422241, + "eval_runtime": 179.1123, + "eval_samples_per_second": 92.908, + "eval_steps_per_second": 5.812, + "step": 4000 + }, + { + "epoch": 1.17939309220106, + "grad_norm": 0.3532305061817169, + "learning_rate": 0.0005861923076923076, + "loss": 4.0042, + "step": 4050 + }, + { + "epoch": 1.1939542198147826, + "grad_norm": 0.3242940902709961, + "learning_rate": 0.0005860174825174824, + "loss": 3.9969, + "step": 4100 + }, + { + "epoch": 1.2085153474285049, + "grad_norm": 0.3659493625164032, + "learning_rate": 0.0005858426573426573, + "loss": 3.9847, + "step": 4150 + }, + { + "epoch": 1.2230764750422272, + "grad_norm": 0.3628295063972473, + "learning_rate": 0.000585667832167832, + "loss": 3.9836, + "step": 4200 + }, + { + "epoch": 1.2376376026559497, + "grad_norm": 0.35134556889533997, + "learning_rate": 0.000585493006993007, + "loss": 3.9942, + "step": 4250 + }, + { + "epoch": 1.2521987302696722, + "grad_norm": 0.3528451919555664, + "learning_rate": 0.0005853181818181817, + "loss": 3.9826, + "step": 4300 + }, + { + "epoch": 1.2667598578833945, + "grad_norm": 0.3459540605545044, + "learning_rate": 0.0005851433566433565, + "loss": 3.9759, + "step": 4350 + }, + { + "epoch": 1.2813209854971168, + "grad_norm": 0.341845840215683, + "learning_rate": 0.0005849685314685315, + "loss": 3.9846, + "step": 4400 + }, + { + "epoch": 1.2958821131108393, + "grad_norm": 0.3483335077762604, + "learning_rate": 0.0005847937062937063, + "loss": 3.9729, + "step": 4450 + }, + { + "epoch": 1.3104432407245616, + "grad_norm": 0.333688884973526, + "learning_rate": 0.0005846188811188811, + "loss": 3.9507, + "step": 4500 + }, + { + "epoch": 1.3250043683382842, + "grad_norm": 0.3437134325504303, + "learning_rate": 0.0005844440559440559, + "loss": 3.9595, + "step": 4550 + }, + { + "epoch": 1.3395654959520065, + "grad_norm": 0.3335651755332947, + "learning_rate": 0.0005842692307692308, + "loss": 3.9583, + "step": 4600 + }, + { + "epoch": 1.354126623565729, + "grad_norm": 0.3338441252708435, + "learning_rate": 0.0005840944055944056, + "loss": 3.96, + "step": 4650 + }, + { + "epoch": 1.3686877511794513, + "grad_norm": 0.3445087671279907, + "learning_rate": 0.0005839195804195804, + "loss": 3.9468, + "step": 4700 + }, + { + "epoch": 1.3832488787931738, + "grad_norm": 0.34749364852905273, + "learning_rate": 0.0005837447552447552, + "loss": 3.9458, + "step": 4750 + }, + { + "epoch": 1.3978100064068961, + "grad_norm": 0.3308727443218231, + "learning_rate": 0.0005835699300699301, + "loss": 3.9474, + "step": 4800 + }, + { + "epoch": 1.4123711340206184, + "grad_norm": 0.34063777327537537, + "learning_rate": 0.0005833951048951048, + "loss": 3.9294, + "step": 4850 + }, + { + "epoch": 1.426932261634341, + "grad_norm": 0.33163222670555115, + "learning_rate": 0.0005832202797202797, + "loss": 3.9419, + "step": 4900 + }, + { + "epoch": 1.4414933892480635, + "grad_norm": 0.34162136912345886, + "learning_rate": 0.0005830454545454546, + "loss": 3.9489, + "step": 4950 + }, + { + "epoch": 1.4560545168617858, + "grad_norm": 0.34426388144493103, + "learning_rate": 0.0005828706293706293, + "loss": 3.9277, + "step": 5000 + }, + { + "epoch": 1.4560545168617858, + "eval_accuracy": 0.33148363548357623, + "eval_loss": 3.9177608489990234, + "eval_runtime": 179.2433, + "eval_samples_per_second": 92.84, + "eval_steps_per_second": 5.808, + "step": 5000 + }, + { + "epoch": 1.470615644475508, + "grad_norm": 0.3323045074939728, + "learning_rate": 0.0005826958041958042, + "loss": 3.9224, + "step": 5050 + }, + { + "epoch": 1.4851767720892306, + "grad_norm": 0.3566197156906128, + "learning_rate": 0.000582520979020979, + "loss": 3.9173, + "step": 5100 + }, + { + "epoch": 1.4997378997029531, + "grad_norm": 0.32939663529396057, + "learning_rate": 0.0005823461538461538, + "loss": 3.9372, + "step": 5150 + }, + { + "epoch": 1.5142990273166754, + "grad_norm": 0.34687715768814087, + "learning_rate": 0.0005821713286713286, + "loss": 3.9131, + "step": 5200 + }, + { + "epoch": 1.5288601549303977, + "grad_norm": 0.32731279730796814, + "learning_rate": 0.0005819965034965035, + "loss": 3.9105, + "step": 5250 + }, + { + "epoch": 1.5434212825441203, + "grad_norm": 0.33731281757354736, + "learning_rate": 0.0005818216783216783, + "loss": 3.9149, + "step": 5300 + }, + { + "epoch": 1.5579824101578428, + "grad_norm": 0.3197886645793915, + "learning_rate": 0.0005816468531468531, + "loss": 3.9157, + "step": 5350 + }, + { + "epoch": 1.572543537771565, + "grad_norm": 0.32866302132606506, + "learning_rate": 0.0005814720279720279, + "loss": 3.9101, + "step": 5400 + }, + { + "epoch": 1.5871046653852874, + "grad_norm": 0.3010786175727844, + "learning_rate": 0.0005812972027972028, + "loss": 3.9056, + "step": 5450 + }, + { + "epoch": 1.6016657929990097, + "grad_norm": 0.3262270987033844, + "learning_rate": 0.0005811223776223776, + "loss": 3.9083, + "step": 5500 + }, + { + "epoch": 1.6162269206127322, + "grad_norm": 0.3309171795845032, + "learning_rate": 0.0005809475524475524, + "loss": 3.9013, + "step": 5550 + }, + { + "epoch": 1.6307880482264547, + "grad_norm": 0.3318457305431366, + "learning_rate": 0.0005807727272727272, + "loss": 3.899, + "step": 5600 + }, + { + "epoch": 1.645349175840177, + "grad_norm": 0.33176082372665405, + "learning_rate": 0.0005805979020979021, + "loss": 3.8861, + "step": 5650 + }, + { + "epoch": 1.6599103034538993, + "grad_norm": 0.3390110731124878, + "learning_rate": 0.0005804230769230769, + "loss": 3.8957, + "step": 5700 + }, + { + "epoch": 1.6744714310676219, + "grad_norm": 0.3413689136505127, + "learning_rate": 0.0005802482517482517, + "loss": 3.8921, + "step": 5750 + }, + { + "epoch": 1.6890325586813444, + "grad_norm": 0.3367161452770233, + "learning_rate": 0.0005800734265734265, + "loss": 3.8924, + "step": 5800 + }, + { + "epoch": 1.7035936862950667, + "grad_norm": 0.3219411075115204, + "learning_rate": 0.0005798986013986013, + "loss": 3.8962, + "step": 5850 + }, + { + "epoch": 1.718154813908789, + "grad_norm": 0.3066975474357605, + "learning_rate": 0.0005797237762237762, + "loss": 3.8923, + "step": 5900 + }, + { + "epoch": 1.7327159415225115, + "grad_norm": 0.32110902667045593, + "learning_rate": 0.000579548951048951, + "loss": 3.878, + "step": 5950 + }, + { + "epoch": 1.747277069136234, + "grad_norm": 0.3308785557746887, + "learning_rate": 0.0005793741258741258, + "loss": 3.8713, + "step": 6000 + }, + { + "epoch": 1.747277069136234, + "eval_accuracy": 0.3367313255368257, + "eval_loss": 3.8610405921936035, + "eval_runtime": 179.6017, + "eval_samples_per_second": 92.655, + "eval_steps_per_second": 5.796, + "step": 6000 + }, + { + "epoch": 1.7618381967499563, + "grad_norm": 0.33315688371658325, + "learning_rate": 0.0005791993006993006, + "loss": 3.8799, + "step": 6050 + }, + { + "epoch": 1.7763993243636786, + "grad_norm": 0.3255741000175476, + "learning_rate": 0.0005790244755244755, + "loss": 3.8775, + "step": 6100 + }, + { + "epoch": 1.7909604519774012, + "grad_norm": 0.3403703570365906, + "learning_rate": 0.0005788496503496503, + "loss": 3.8766, + "step": 6150 + }, + { + "epoch": 1.8055215795911237, + "grad_norm": 0.3325391411781311, + "learning_rate": 0.0005786748251748251, + "loss": 3.8727, + "step": 6200 + }, + { + "epoch": 1.820082707204846, + "grad_norm": 0.3367559015750885, + "learning_rate": 0.0005784999999999999, + "loss": 3.8623, + "step": 6250 + }, + { + "epoch": 1.8346438348185683, + "grad_norm": 0.31256911158561707, + "learning_rate": 0.0005783251748251748, + "loss": 3.8543, + "step": 6300 + }, + { + "epoch": 1.8492049624322906, + "grad_norm": 0.3506211042404175, + "learning_rate": 0.0005781503496503496, + "loss": 3.8745, + "step": 6350 + }, + { + "epoch": 1.8637660900460131, + "grad_norm": 0.3530968725681305, + "learning_rate": 0.0005779755244755244, + "loss": 3.8703, + "step": 6400 + }, + { + "epoch": 1.8783272176597356, + "grad_norm": 0.3274037837982178, + "learning_rate": 0.0005778006993006993, + "loss": 3.8583, + "step": 6450 + }, + { + "epoch": 1.892888345273458, + "grad_norm": 0.3105003833770752, + "learning_rate": 0.000577625874125874, + "loss": 3.8621, + "step": 6500 + }, + { + "epoch": 1.9074494728871803, + "grad_norm": 0.3492962121963501, + "learning_rate": 0.0005774510489510489, + "loss": 3.8588, + "step": 6550 + }, + { + "epoch": 1.9220106005009028, + "grad_norm": 0.3350745141506195, + "learning_rate": 0.0005772762237762237, + "loss": 3.8567, + "step": 6600 + }, + { + "epoch": 1.9365717281146253, + "grad_norm": 0.332505464553833, + "learning_rate": 0.0005771013986013985, + "loss": 3.8398, + "step": 6650 + }, + { + "epoch": 1.9511328557283476, + "grad_norm": 0.31538161635398865, + "learning_rate": 0.0005769265734265733, + "loss": 3.8456, + "step": 6700 + }, + { + "epoch": 1.96569398334207, + "grad_norm": 0.31272760033607483, + "learning_rate": 0.0005767517482517482, + "loss": 3.8481, + "step": 6750 + }, + { + "epoch": 1.9802551109557924, + "grad_norm": 0.3264569044113159, + "learning_rate": 0.000576576923076923, + "loss": 3.8422, + "step": 6800 + }, + { + "epoch": 1.994816238569515, + "grad_norm": 0.3190554976463318, + "learning_rate": 0.0005764020979020978, + "loss": 3.8397, + "step": 6850 + }, + { + "epoch": 2.009319121672782, + "grad_norm": 0.3068064749240875, + "learning_rate": 0.0005762272727272726, + "loss": 3.7913, + "step": 6900 + }, + { + "epoch": 2.023880249286505, + "grad_norm": 0.32731175422668457, + "learning_rate": 0.0005760524475524475, + "loss": 3.7477, + "step": 6950 + }, + { + "epoch": 2.038441376900227, + "grad_norm": 0.32606661319732666, + "learning_rate": 0.0005758776223776223, + "loss": 3.7496, + "step": 7000 + }, + { + "epoch": 2.038441376900227, + "eval_accuracy": 0.34119357901187397, + "eval_loss": 3.819080352783203, + "eval_runtime": 179.5119, + "eval_samples_per_second": 92.701, + "eval_steps_per_second": 5.799, + "step": 7000 + }, + { + "epoch": 2.0530025045139495, + "grad_norm": 0.3195810317993164, + "learning_rate": 0.0005757027972027971, + "loss": 3.7426, + "step": 7050 + }, + { + "epoch": 2.067563632127672, + "grad_norm": 0.33737748861312866, + "learning_rate": 0.000575527972027972, + "loss": 3.7466, + "step": 7100 + }, + { + "epoch": 2.0821247597413945, + "grad_norm": 0.33379852771759033, + "learning_rate": 0.0005753531468531468, + "loss": 3.7562, + "step": 7150 + }, + { + "epoch": 2.096685887355117, + "grad_norm": 0.32753437757492065, + "learning_rate": 0.0005751783216783216, + "loss": 3.7565, + "step": 7200 + }, + { + "epoch": 2.111247014968839, + "grad_norm": 0.31312432885169983, + "learning_rate": 0.0005750034965034964, + "loss": 3.7545, + "step": 7250 + }, + { + "epoch": 2.1258081425825615, + "grad_norm": 0.34001678228378296, + "learning_rate": 0.0005748286713286712, + "loss": 3.7624, + "step": 7300 + }, + { + "epoch": 2.140369270196284, + "grad_norm": 0.32072916626930237, + "learning_rate": 0.000574653846153846, + "loss": 3.759, + "step": 7350 + }, + { + "epoch": 2.1549303978100065, + "grad_norm": 0.32128438353538513, + "learning_rate": 0.000574479020979021, + "loss": 3.7536, + "step": 7400 + }, + { + "epoch": 2.169491525423729, + "grad_norm": 0.32183223962783813, + "learning_rate": 0.0005743041958041958, + "loss": 3.7615, + "step": 7450 + }, + { + "epoch": 2.184052653037451, + "grad_norm": 0.33787763118743896, + "learning_rate": 0.0005741293706293706, + "loss": 3.7465, + "step": 7500 + }, + { + "epoch": 2.198613780651174, + "grad_norm": 0.3221549093723297, + "learning_rate": 0.0005739545454545454, + "loss": 3.7529, + "step": 7550 + }, + { + "epoch": 2.213174908264896, + "grad_norm": 0.3178912401199341, + "learning_rate": 0.0005737797202797203, + "loss": 3.7603, + "step": 7600 + }, + { + "epoch": 2.2277360358786185, + "grad_norm": 0.32143399119377136, + "learning_rate": 0.0005736048951048951, + "loss": 3.7542, + "step": 7650 + }, + { + "epoch": 2.2422971634923408, + "grad_norm": 0.32524600625038147, + "learning_rate": 0.0005734300699300699, + "loss": 3.7538, + "step": 7700 + }, + { + "epoch": 2.256858291106063, + "grad_norm": 0.3093379735946655, + "learning_rate": 0.0005732552447552448, + "loss": 3.7424, + "step": 7750 + }, + { + "epoch": 2.271419418719786, + "grad_norm": 0.3282085657119751, + "learning_rate": 0.0005730804195804196, + "loss": 3.7587, + "step": 7800 + }, + { + "epoch": 2.285980546333508, + "grad_norm": 0.3183515667915344, + "learning_rate": 0.0005729055944055944, + "loss": 3.7581, + "step": 7850 + }, + { + "epoch": 2.3005416739472304, + "grad_norm": 0.3316892683506012, + "learning_rate": 0.0005727307692307692, + "loss": 3.751, + "step": 7900 + }, + { + "epoch": 2.3151028015609527, + "grad_norm": 0.3208901286125183, + "learning_rate": 0.0005725559440559441, + "loss": 3.7519, + "step": 7950 + }, + { + "epoch": 2.3296639291746755, + "grad_norm": 0.3379865288734436, + "learning_rate": 0.0005723811188811188, + "loss": 3.7479, + "step": 8000 + }, + { + "epoch": 2.3296639291746755, + "eval_accuracy": 0.3441776264997999, + "eval_loss": 3.787078619003296, + "eval_runtime": 180.9921, + "eval_samples_per_second": 91.943, + "eval_steps_per_second": 5.752, + "step": 8000 + }, + { + "epoch": 2.3442250567883978, + "grad_norm": 0.3328506648540497, + "learning_rate": 0.0005722062937062937, + "loss": 3.7636, + "step": 8050 + }, + { + "epoch": 2.35878618440212, + "grad_norm": 0.3388136327266693, + "learning_rate": 0.0005720314685314685, + "loss": 3.7577, + "step": 8100 + }, + { + "epoch": 2.3733473120158424, + "grad_norm": 0.3213030695915222, + "learning_rate": 0.0005718566433566433, + "loss": 3.7558, + "step": 8150 + }, + { + "epoch": 2.387908439629565, + "grad_norm": 0.32116878032684326, + "learning_rate": 0.0005716818181818181, + "loss": 3.7593, + "step": 8200 + }, + { + "epoch": 2.4024695672432874, + "grad_norm": 0.3235709071159363, + "learning_rate": 0.000571506993006993, + "loss": 3.7453, + "step": 8250 + }, + { + "epoch": 2.4170306948570097, + "grad_norm": 0.32394877076148987, + "learning_rate": 0.0005713321678321678, + "loss": 3.7653, + "step": 8300 + }, + { + "epoch": 2.431591822470732, + "grad_norm": 0.32393255829811096, + "learning_rate": 0.0005711573426573426, + "loss": 3.7586, + "step": 8350 + }, + { + "epoch": 2.4461529500844543, + "grad_norm": 0.32673174142837524, + "learning_rate": 0.0005709825174825175, + "loss": 3.7576, + "step": 8400 + }, + { + "epoch": 2.460714077698177, + "grad_norm": 0.33789971470832825, + "learning_rate": 0.0005708076923076923, + "loss": 3.7604, + "step": 8450 + }, + { + "epoch": 2.4752752053118994, + "grad_norm": 0.34079620242118835, + "learning_rate": 0.0005706328671328671, + "loss": 3.7621, + "step": 8500 + }, + { + "epoch": 2.4898363329256217, + "grad_norm": 0.34042778611183167, + "learning_rate": 0.0005704580419580419, + "loss": 3.7344, + "step": 8550 + }, + { + "epoch": 2.5043974605393444, + "grad_norm": 0.3570379614830017, + "learning_rate": 0.0005702832167832168, + "loss": 3.7493, + "step": 8600 + }, + { + "epoch": 2.5189585881530667, + "grad_norm": 0.33469274640083313, + "learning_rate": 0.0005701083916083916, + "loss": 3.7592, + "step": 8650 + }, + { + "epoch": 2.533519715766789, + "grad_norm": 0.3132188618183136, + "learning_rate": 0.0005699335664335664, + "loss": 3.7412, + "step": 8700 + }, + { + "epoch": 2.5480808433805113, + "grad_norm": 0.31495195627212524, + "learning_rate": 0.0005697587412587412, + "loss": 3.7473, + "step": 8750 + }, + { + "epoch": 2.5626419709942336, + "grad_norm": 0.34376952052116394, + "learning_rate": 0.000569583916083916, + "loss": 3.7369, + "step": 8800 + }, + { + "epoch": 2.5772030986079564, + "grad_norm": 0.3114527463912964, + "learning_rate": 0.0005694090909090908, + "loss": 3.7386, + "step": 8850 + }, + { + "epoch": 2.5917642262216787, + "grad_norm": 0.3196084499359131, + "learning_rate": 0.0005692342657342657, + "loss": 3.7507, + "step": 8900 + }, + { + "epoch": 2.606325353835401, + "grad_norm": 0.32373708486557007, + "learning_rate": 0.0005690594405594405, + "loss": 3.7424, + "step": 8950 + }, + { + "epoch": 2.6208864814491233, + "grad_norm": 0.3105376958847046, + "learning_rate": 0.0005688846153846153, + "loss": 3.7319, + "step": 9000 + }, + { + "epoch": 2.6208864814491233, + "eval_accuracy": 0.34705407188126464, + "eval_loss": 3.7568278312683105, + "eval_runtime": 180.6425, + "eval_samples_per_second": 92.121, + "eval_steps_per_second": 5.763, + "step": 9000 + }, + { + "epoch": 2.6354476090628456, + "grad_norm": 0.30499860644340515, + "learning_rate": 0.0005687097902097901, + "loss": 3.7471, + "step": 9050 + }, + { + "epoch": 2.6500087366765683, + "grad_norm": 0.33331161737442017, + "learning_rate": 0.000568534965034965, + "loss": 3.7444, + "step": 9100 + }, + { + "epoch": 2.6645698642902906, + "grad_norm": 0.318800687789917, + "learning_rate": 0.0005683601398601398, + "loss": 3.7395, + "step": 9150 + }, + { + "epoch": 2.679130991904013, + "grad_norm": 0.3077396750450134, + "learning_rate": 0.0005681853146853146, + "loss": 3.7271, + "step": 9200 + }, + { + "epoch": 2.6936921195177357, + "grad_norm": 0.32172080874443054, + "learning_rate": 0.0005680104895104895, + "loss": 3.7468, + "step": 9250 + }, + { + "epoch": 2.708253247131458, + "grad_norm": 0.3199351131916046, + "learning_rate": 0.0005678356643356643, + "loss": 3.7271, + "step": 9300 + }, + { + "epoch": 2.7228143747451803, + "grad_norm": 0.334290087223053, + "learning_rate": 0.0005676608391608391, + "loss": 3.7435, + "step": 9350 + }, + { + "epoch": 2.7373755023589026, + "grad_norm": 0.30215924978256226, + "learning_rate": 0.0005674860139860139, + "loss": 3.7213, + "step": 9400 + }, + { + "epoch": 2.751936629972625, + "grad_norm": 0.3150777220726013, + "learning_rate": 0.0005673111888111888, + "loss": 3.7354, + "step": 9450 + }, + { + "epoch": 2.7664977575863476, + "grad_norm": 0.32755008339881897, + "learning_rate": 0.0005671363636363635, + "loss": 3.731, + "step": 9500 + }, + { + "epoch": 2.78105888520007, + "grad_norm": 0.33355712890625, + "learning_rate": 0.0005669615384615384, + "loss": 3.7366, + "step": 9550 + }, + { + "epoch": 2.7956200128137922, + "grad_norm": 0.32244864106178284, + "learning_rate": 0.0005667867132867132, + "loss": 3.7394, + "step": 9600 + }, + { + "epoch": 2.8101811404275145, + "grad_norm": 0.3095114231109619, + "learning_rate": 0.000566611888111888, + "loss": 3.7366, + "step": 9650 + }, + { + "epoch": 2.824742268041237, + "grad_norm": 0.353028267621994, + "learning_rate": 0.0005664370629370628, + "loss": 3.7387, + "step": 9700 + }, + { + "epoch": 2.8393033956549596, + "grad_norm": 0.32903262972831726, + "learning_rate": 0.0005662622377622377, + "loss": 3.7375, + "step": 9750 + }, + { + "epoch": 2.853864523268682, + "grad_norm": 0.3056410253047943, + "learning_rate": 0.0005660874125874125, + "loss": 3.7238, + "step": 9800 + }, + { + "epoch": 2.868425650882404, + "grad_norm": 0.31195202469825745, + "learning_rate": 0.0005659125874125873, + "loss": 3.727, + "step": 9850 + }, + { + "epoch": 2.882986778496127, + "grad_norm": 0.3105180263519287, + "learning_rate": 0.0005657377622377622, + "loss": 3.7285, + "step": 9900 + }, + { + "epoch": 2.8975479061098492, + "grad_norm": 0.31410694122314453, + "learning_rate": 0.000565562937062937, + "loss": 3.7289, + "step": 9950 + }, + { + "epoch": 2.9121090337235715, + "grad_norm": 0.3199072480201721, + "learning_rate": 0.0005653881118881118, + "loss": 3.7207, + "step": 10000 + }, + { + "epoch": 2.9121090337235715, + "eval_accuracy": 0.34957090279108105, + "eval_loss": 3.7290003299713135, + "eval_runtime": 179.6135, + "eval_samples_per_second": 92.649, + "eval_steps_per_second": 5.796, + "step": 10000 + }, + { + "epoch": 2.926670161337294, + "grad_norm": 0.3202771544456482, + "learning_rate": 0.0005652132867132866, + "loss": 3.7316, + "step": 10050 + }, + { + "epoch": 2.941231288951016, + "grad_norm": 0.3228454887866974, + "learning_rate": 0.0005650384615384615, + "loss": 3.7221, + "step": 10100 + }, + { + "epoch": 2.955792416564739, + "grad_norm": 0.302839457988739, + "learning_rate": 0.0005648636363636363, + "loss": 3.7298, + "step": 10150 + }, + { + "epoch": 2.970353544178461, + "grad_norm": 0.3409121036529541, + "learning_rate": 0.0005646888111888111, + "loss": 3.7142, + "step": 10200 + }, + { + "epoch": 2.9849146717921835, + "grad_norm": 0.305637389421463, + "learning_rate": 0.000564513986013986, + "loss": 3.7306, + "step": 10250 + }, + { + "epoch": 2.9994757994059063, + "grad_norm": 0.32202163338661194, + "learning_rate": 0.0005643391608391607, + "loss": 3.7141, + "step": 10300 + }, + { + "epoch": 3.0139786825091734, + "grad_norm": 0.3135906457901001, + "learning_rate": 0.0005641643356643355, + "loss": 3.618, + "step": 10350 + }, + { + "epoch": 3.0285398101228957, + "grad_norm": 0.3280876576900482, + "learning_rate": 0.0005639895104895105, + "loss": 3.6231, + "step": 10400 + }, + { + "epoch": 3.0431009377366185, + "grad_norm": 0.32824140787124634, + "learning_rate": 0.0005638146853146853, + "loss": 3.6222, + "step": 10450 + }, + { + "epoch": 3.057662065350341, + "grad_norm": 0.3240225315093994, + "learning_rate": 0.0005636398601398601, + "loss": 3.6236, + "step": 10500 + }, + { + "epoch": 3.072223192964063, + "grad_norm": 0.30825719237327576, + "learning_rate": 0.000563465034965035, + "loss": 3.6228, + "step": 10550 + }, + { + "epoch": 3.0867843205777854, + "grad_norm": 0.3186241388320923, + "learning_rate": 0.0005632902097902098, + "loss": 3.6176, + "step": 10600 + }, + { + "epoch": 3.101345448191508, + "grad_norm": 0.3103659152984619, + "learning_rate": 0.0005631153846153846, + "loss": 3.6289, + "step": 10650 + }, + { + "epoch": 3.1159065758052304, + "grad_norm": 0.31888821721076965, + "learning_rate": 0.0005629405594405594, + "loss": 3.6213, + "step": 10700 + }, + { + "epoch": 3.1304677034189528, + "grad_norm": 0.31721800565719604, + "learning_rate": 0.0005627657342657343, + "loss": 3.6386, + "step": 10750 + }, + { + "epoch": 3.145028831032675, + "grad_norm": 0.3233102560043335, + "learning_rate": 0.0005625909090909091, + "loss": 3.6312, + "step": 10800 + }, + { + "epoch": 3.1595899586463974, + "grad_norm": 0.34500283002853394, + "learning_rate": 0.0005624160839160839, + "loss": 3.6397, + "step": 10850 + }, + { + "epoch": 3.17415108626012, + "grad_norm": 0.3219081461429596, + "learning_rate": 0.0005622412587412587, + "loss": 3.6393, + "step": 10900 + }, + { + "epoch": 3.1887122138738424, + "grad_norm": 0.35239356756210327, + "learning_rate": 0.0005620664335664336, + "loss": 3.6313, + "step": 10950 + }, + { + "epoch": 3.2032733414875647, + "grad_norm": 0.3073521554470062, + "learning_rate": 0.0005618916083916083, + "loss": 3.6445, + "step": 11000 + }, + { + "epoch": 3.2032733414875647, + "eval_accuracy": 0.3511596508329285, + "eval_loss": 3.7183070182800293, + "eval_runtime": 179.4026, + "eval_samples_per_second": 92.758, + "eval_steps_per_second": 5.803, + "step": 11000 + }, + { + "epoch": 3.217834469101287, + "grad_norm": 0.31943169236183167, + "learning_rate": 0.0005617167832167832, + "loss": 3.6346, + "step": 11050 + }, + { + "epoch": 3.2323955967150098, + "grad_norm": 0.32679134607315063, + "learning_rate": 0.000561541958041958, + "loss": 3.6398, + "step": 11100 + }, + { + "epoch": 3.246956724328732, + "grad_norm": 0.33268868923187256, + "learning_rate": 0.0005613671328671328, + "loss": 3.6401, + "step": 11150 + }, + { + "epoch": 3.2615178519424544, + "grad_norm": 0.33772939443588257, + "learning_rate": 0.0005611923076923077, + "loss": 3.6452, + "step": 11200 + }, + { + "epoch": 3.2760789795561767, + "grad_norm": 0.3287737965583801, + "learning_rate": 0.0005610174825174825, + "loss": 3.6403, + "step": 11250 + }, + { + "epoch": 3.2906401071698994, + "grad_norm": 0.34239116311073303, + "learning_rate": 0.0005608426573426573, + "loss": 3.6236, + "step": 11300 + }, + { + "epoch": 3.3052012347836217, + "grad_norm": 0.3230609893798828, + "learning_rate": 0.0005606678321678321, + "loss": 3.6434, + "step": 11350 + }, + { + "epoch": 3.319762362397344, + "grad_norm": 0.3213825225830078, + "learning_rate": 0.000560493006993007, + "loss": 3.6468, + "step": 11400 + }, + { + "epoch": 3.3343234900110663, + "grad_norm": 0.32964688539505005, + "learning_rate": 0.0005603181818181818, + "loss": 3.6545, + "step": 11450 + }, + { + "epoch": 3.3488846176247886, + "grad_norm": 0.330831378698349, + "learning_rate": 0.0005601433566433566, + "loss": 3.6381, + "step": 11500 + }, + { + "epoch": 3.3634457452385114, + "grad_norm": 0.3213164806365967, + "learning_rate": 0.0005599685314685314, + "loss": 3.6442, + "step": 11550 + }, + { + "epoch": 3.3780068728522337, + "grad_norm": 0.3467330038547516, + "learning_rate": 0.0005597937062937063, + "loss": 3.642, + "step": 11600 + }, + { + "epoch": 3.392568000465956, + "grad_norm": 0.31923604011535645, + "learning_rate": 0.0005596188811188811, + "loss": 3.6543, + "step": 11650 + }, + { + "epoch": 3.4071291280796787, + "grad_norm": 0.30125677585601807, + "learning_rate": 0.0005594440559440559, + "loss": 3.6587, + "step": 11700 + }, + { + "epoch": 3.421690255693401, + "grad_norm": 0.34452974796295166, + "learning_rate": 0.0005592692307692307, + "loss": 3.642, + "step": 11750 + }, + { + "epoch": 3.4362513833071233, + "grad_norm": 0.3240102231502533, + "learning_rate": 0.0005590944055944055, + "loss": 3.6498, + "step": 11800 + }, + { + "epoch": 3.4508125109208456, + "grad_norm": 0.3127278685569763, + "learning_rate": 0.0005589195804195803, + "loss": 3.65, + "step": 11850 + }, + { + "epoch": 3.465373638534568, + "grad_norm": 0.30665114521980286, + "learning_rate": 0.0005587447552447552, + "loss": 3.6551, + "step": 11900 + }, + { + "epoch": 3.4799347661482907, + "grad_norm": 0.31048598885536194, + "learning_rate": 0.00055856993006993, + "loss": 3.6496, + "step": 11950 + }, + { + "epoch": 3.494495893762013, + "grad_norm": 0.3066251873970032, + "learning_rate": 0.0005583951048951048, + "loss": 3.638, + "step": 12000 + }, + { + "epoch": 3.494495893762013, + "eval_accuracy": 0.35302451881572766, + "eval_loss": 3.7004880905151367, + "eval_runtime": 179.3816, + "eval_samples_per_second": 92.769, + "eval_steps_per_second": 5.803, + "step": 12000 + }, + { + "epoch": 3.5090570213757353, + "grad_norm": 0.33529502153396606, + "learning_rate": 0.0005582202797202797, + "loss": 3.6461, + "step": 12050 + }, + { + "epoch": 3.523618148989458, + "grad_norm": 0.3222676217556, + "learning_rate": 0.0005580454545454545, + "loss": 3.6394, + "step": 12100 + }, + { + "epoch": 3.53817927660318, + "grad_norm": 0.3227137625217438, + "learning_rate": 0.0005578706293706293, + "loss": 3.6426, + "step": 12150 + }, + { + "epoch": 3.5527404042169026, + "grad_norm": 0.2938767671585083, + "learning_rate": 0.0005576958041958041, + "loss": 3.6354, + "step": 12200 + }, + { + "epoch": 3.567301531830625, + "grad_norm": 0.32853955030441284, + "learning_rate": 0.000557520979020979, + "loss": 3.6471, + "step": 12250 + }, + { + "epoch": 3.5818626594443472, + "grad_norm": 0.322208970785141, + "learning_rate": 0.0005573461538461538, + "loss": 3.6455, + "step": 12300 + }, + { + "epoch": 3.59642378705807, + "grad_norm": 0.33103466033935547, + "learning_rate": 0.0005571713286713286, + "loss": 3.6602, + "step": 12350 + }, + { + "epoch": 3.6109849146717923, + "grad_norm": 0.3074956238269806, + "learning_rate": 0.0005569965034965034, + "loss": 3.6518, + "step": 12400 + }, + { + "epoch": 3.6255460422855146, + "grad_norm": 0.30890724062919617, + "learning_rate": 0.0005568216783216783, + "loss": 3.6496, + "step": 12450 + }, + { + "epoch": 3.640107169899237, + "grad_norm": 0.3310304582118988, + "learning_rate": 0.000556646853146853, + "loss": 3.6425, + "step": 12500 + }, + { + "epoch": 3.654668297512959, + "grad_norm": 0.33728861808776855, + "learning_rate": 0.0005564720279720279, + "loss": 3.6406, + "step": 12550 + }, + { + "epoch": 3.669229425126682, + "grad_norm": 0.3305113911628723, + "learning_rate": 0.0005562972027972027, + "loss": 3.6467, + "step": 12600 + }, + { + "epoch": 3.6837905527404042, + "grad_norm": 0.3080269992351532, + "learning_rate": 0.0005561223776223775, + "loss": 3.6374, + "step": 12650 + }, + { + "epoch": 3.6983516803541265, + "grad_norm": 0.3238011598587036, + "learning_rate": 0.0005559475524475524, + "loss": 3.6418, + "step": 12700 + }, + { + "epoch": 3.7129128079678493, + "grad_norm": 0.33371302485466003, + "learning_rate": 0.0005557727272727272, + "loss": 3.6393, + "step": 12750 + }, + { + "epoch": 3.7274739355815716, + "grad_norm": 0.313056081533432, + "learning_rate": 0.000555597902097902, + "loss": 3.6459, + "step": 12800 + }, + { + "epoch": 3.742035063195294, + "grad_norm": 0.3184046745300293, + "learning_rate": 0.0005554230769230768, + "loss": 3.6449, + "step": 12850 + }, + { + "epoch": 3.756596190809016, + "grad_norm": 0.33030468225479126, + "learning_rate": 0.0005552482517482517, + "loss": 3.6424, + "step": 12900 + }, + { + "epoch": 3.7711573184227385, + "grad_norm": 0.311156302690506, + "learning_rate": 0.0005550734265734265, + "loss": 3.6538, + "step": 12950 + }, + { + "epoch": 3.7857184460364612, + "grad_norm": 0.3100126087665558, + "learning_rate": 0.0005548986013986013, + "loss": 3.644, + "step": 13000 + }, + { + "epoch": 3.7857184460364612, + "eval_accuracy": 0.35456269974743493, + "eval_loss": 3.684239149093628, + "eval_runtime": 179.5691, + "eval_samples_per_second": 92.672, + "eval_steps_per_second": 5.797, + "step": 13000 + }, + { + "epoch": 3.8002795736501835, + "grad_norm": 0.31739217042922974, + "learning_rate": 0.0005547237762237761, + "loss": 3.6415, + "step": 13050 + }, + { + "epoch": 3.814840701263906, + "grad_norm": 0.35627153515815735, + "learning_rate": 0.000554548951048951, + "loss": 3.6551, + "step": 13100 + }, + { + "epoch": 3.829401828877628, + "grad_norm": 0.3268809914588928, + "learning_rate": 0.0005543741258741258, + "loss": 3.6476, + "step": 13150 + }, + { + "epoch": 3.8439629564913504, + "grad_norm": 0.3053591549396515, + "learning_rate": 0.0005541993006993006, + "loss": 3.6348, + "step": 13200 + }, + { + "epoch": 3.858524084105073, + "grad_norm": 0.3270852565765381, + "learning_rate": 0.0005540244755244756, + "loss": 3.6359, + "step": 13250 + }, + { + "epoch": 3.8730852117187955, + "grad_norm": 0.31783512234687805, + "learning_rate": 0.0005538496503496502, + "loss": 3.6464, + "step": 13300 + }, + { + "epoch": 3.887646339332518, + "grad_norm": 0.3072042763233185, + "learning_rate": 0.0005536748251748252, + "loss": 3.6407, + "step": 13350 + }, + { + "epoch": 3.9022074669462405, + "grad_norm": 0.3247385323047638, + "learning_rate": 0.0005535, + "loss": 3.644, + "step": 13400 + }, + { + "epoch": 3.916768594559963, + "grad_norm": 0.3208458423614502, + "learning_rate": 0.0005533251748251748, + "loss": 3.6403, + "step": 13450 + }, + { + "epoch": 3.931329722173685, + "grad_norm": 0.3018021285533905, + "learning_rate": 0.0005531503496503496, + "loss": 3.6399, + "step": 13500 + }, + { + "epoch": 3.9458908497874075, + "grad_norm": 0.3152608871459961, + "learning_rate": 0.0005529755244755245, + "loss": 3.6424, + "step": 13550 + }, + { + "epoch": 3.9604519774011298, + "grad_norm": 0.3139157295227051, + "learning_rate": 0.0005528006993006993, + "loss": 3.6339, + "step": 13600 + }, + { + "epoch": 3.9750131050148525, + "grad_norm": 0.3041459321975708, + "learning_rate": 0.0005526258741258741, + "loss": 3.6422, + "step": 13650 + }, + { + "epoch": 3.989574232628575, + "grad_norm": 0.3212553560733795, + "learning_rate": 0.0005524510489510489, + "loss": 3.6308, + "step": 13700 + }, + { + "epoch": 4.004077115731842, + "grad_norm": 0.31951627135276794, + "learning_rate": 0.0005522762237762238, + "loss": 3.6008, + "step": 13750 + }, + { + "epoch": 4.018638243345564, + "grad_norm": 0.32649630308151245, + "learning_rate": 0.0005521013986013986, + "loss": 3.5126, + "step": 13800 + }, + { + "epoch": 4.033199370959287, + "grad_norm": 0.3360797166824341, + "learning_rate": 0.0005519265734265734, + "loss": 3.539, + "step": 13850 + }, + { + "epoch": 4.04776049857301, + "grad_norm": 0.3214600682258606, + "learning_rate": 0.0005517517482517482, + "loss": 3.5333, + "step": 13900 + }, + { + "epoch": 4.062321626186732, + "grad_norm": 0.3207820951938629, + "learning_rate": 0.0005515769230769231, + "loss": 3.5418, + "step": 13950 + }, + { + "epoch": 4.076882753800454, + "grad_norm": 0.3166014850139618, + "learning_rate": 0.0005514020979020979, + "loss": 3.5463, + "step": 14000 + }, + { + "epoch": 4.076882753800454, + "eval_accuracy": 0.3557597290825915, + "eval_loss": 3.6768152713775635, + "eval_runtime": 179.3506, + "eval_samples_per_second": 92.785, + "eval_steps_per_second": 5.804, + "step": 14000 + }, + { + "epoch": 4.091443881414177, + "grad_norm": 0.32194244861602783, + "learning_rate": 0.0005512272727272727, + "loss": 3.5486, + "step": 14050 + }, + { + "epoch": 4.106005009027899, + "grad_norm": 0.33398452401161194, + "learning_rate": 0.0005510524475524475, + "loss": 3.5615, + "step": 14100 + }, + { + "epoch": 4.120566136641622, + "grad_norm": 0.3263501226902008, + "learning_rate": 0.0005508776223776223, + "loss": 3.5564, + "step": 14150 + }, + { + "epoch": 4.135127264255344, + "grad_norm": 0.31160303950309753, + "learning_rate": 0.0005507027972027972, + "loss": 3.5539, + "step": 14200 + }, + { + "epoch": 4.149688391869066, + "grad_norm": 0.34740275144577026, + "learning_rate": 0.000550527972027972, + "loss": 3.5539, + "step": 14250 + }, + { + "epoch": 4.164249519482789, + "grad_norm": 0.34102705121040344, + "learning_rate": 0.0005503531468531468, + "loss": 3.5511, + "step": 14300 + }, + { + "epoch": 4.178810647096511, + "grad_norm": 0.3336825370788574, + "learning_rate": 0.0005501783216783216, + "loss": 3.5628, + "step": 14350 + }, + { + "epoch": 4.193371774710234, + "grad_norm": 0.3288195729255676, + "learning_rate": 0.0005500034965034965, + "loss": 3.5549, + "step": 14400 + }, + { + "epoch": 4.207932902323956, + "grad_norm": 0.3203940987586975, + "learning_rate": 0.0005498286713286713, + "loss": 3.5672, + "step": 14450 + }, + { + "epoch": 4.222494029937678, + "grad_norm": 0.3168569505214691, + "learning_rate": 0.0005496538461538461, + "loss": 3.5574, + "step": 14500 + }, + { + "epoch": 4.237055157551401, + "grad_norm": 0.3419305980205536, + "learning_rate": 0.0005494790209790209, + "loss": 3.5574, + "step": 14550 + }, + { + "epoch": 4.251616285165123, + "grad_norm": 0.329912006855011, + "learning_rate": 0.0005493041958041958, + "loss": 3.5602, + "step": 14600 + }, + { + "epoch": 4.266177412778846, + "grad_norm": 0.3159247636795044, + "learning_rate": 0.0005491293706293706, + "loss": 3.5608, + "step": 14650 + }, + { + "epoch": 4.280738540392568, + "grad_norm": 0.3186332583427429, + "learning_rate": 0.0005489545454545454, + "loss": 3.5694, + "step": 14700 + }, + { + "epoch": 4.29529966800629, + "grad_norm": 0.3384833335876465, + "learning_rate": 0.0005487797202797203, + "loss": 3.5734, + "step": 14750 + }, + { + "epoch": 4.309860795620013, + "grad_norm": 0.3567145764827728, + "learning_rate": 0.000548604895104895, + "loss": 3.5744, + "step": 14800 + }, + { + "epoch": 4.324421923233735, + "grad_norm": 0.32064342498779297, + "learning_rate": 0.0005484300699300699, + "loss": 3.5708, + "step": 14850 + }, + { + "epoch": 4.338983050847458, + "grad_norm": 0.32063060998916626, + "learning_rate": 0.0005482552447552447, + "loss": 3.566, + "step": 14900 + }, + { + "epoch": 4.35354417846118, + "grad_norm": 0.334367573261261, + "learning_rate": 0.0005480804195804195, + "loss": 3.5726, + "step": 14950 + }, + { + "epoch": 4.368105306074902, + "grad_norm": 0.34719032049179077, + "learning_rate": 0.0005479055944055943, + "loss": 3.5757, + "step": 15000 + }, + { + "epoch": 4.368105306074902, + "eval_accuracy": 0.35707223958555667, + "eval_loss": 3.6637015342712402, + "eval_runtime": 179.1964, + "eval_samples_per_second": 92.865, + "eval_steps_per_second": 5.809, + "step": 15000 + }, + { + "epoch": 4.382666433688625, + "grad_norm": 0.31356847286224365, + "learning_rate": 0.0005477307692307692, + "loss": 3.577, + "step": 15050 + }, + { + "epoch": 4.397227561302348, + "grad_norm": 0.321928471326828, + "learning_rate": 0.000547555944055944, + "loss": 3.5658, + "step": 15100 + }, + { + "epoch": 4.41178868891607, + "grad_norm": 0.33381205797195435, + "learning_rate": 0.0005473811188811188, + "loss": 3.5686, + "step": 15150 + }, + { + "epoch": 4.426349816529792, + "grad_norm": 0.3244903087615967, + "learning_rate": 0.0005472062937062936, + "loss": 3.5741, + "step": 15200 + }, + { + "epoch": 4.440910944143514, + "grad_norm": 0.3154377341270447, + "learning_rate": 0.0005470314685314685, + "loss": 3.5816, + "step": 15250 + }, + { + "epoch": 4.455472071757237, + "grad_norm": 0.3369635343551636, + "learning_rate": 0.0005468566433566433, + "loss": 3.5686, + "step": 15300 + }, + { + "epoch": 4.47003319937096, + "grad_norm": 0.3178083002567291, + "learning_rate": 0.0005466818181818181, + "loss": 3.589, + "step": 15350 + }, + { + "epoch": 4.4845943269846815, + "grad_norm": 0.3455277681350708, + "learning_rate": 0.000546506993006993, + "loss": 3.572, + "step": 15400 + }, + { + "epoch": 4.499155454598404, + "grad_norm": 0.3200279176235199, + "learning_rate": 0.0005463321678321678, + "loss": 3.5712, + "step": 15450 + }, + { + "epoch": 4.513716582212126, + "grad_norm": 0.30433082580566406, + "learning_rate": 0.0005461573426573426, + "loss": 3.5659, + "step": 15500 + }, + { + "epoch": 4.528277709825849, + "grad_norm": 0.32228726148605347, + "learning_rate": 0.0005459825174825174, + "loss": 3.5753, + "step": 15550 + }, + { + "epoch": 4.542838837439572, + "grad_norm": 0.3227219581604004, + "learning_rate": 0.0005458076923076922, + "loss": 3.581, + "step": 15600 + }, + { + "epoch": 4.5573999650532935, + "grad_norm": 0.31072357296943665, + "learning_rate": 0.000545632867132867, + "loss": 3.5821, + "step": 15650 + }, + { + "epoch": 4.571961092667016, + "grad_norm": 0.3410278260707855, + "learning_rate": 0.0005454580419580419, + "loss": 3.5746, + "step": 15700 + }, + { + "epoch": 4.586522220280738, + "grad_norm": 0.3314943313598633, + "learning_rate": 0.0005452832167832167, + "loss": 3.585, + "step": 15750 + }, + { + "epoch": 4.601083347894461, + "grad_norm": 0.31863710284233093, + "learning_rate": 0.0005451083916083915, + "loss": 3.576, + "step": 15800 + }, + { + "epoch": 4.615644475508184, + "grad_norm": 0.3237667977809906, + "learning_rate": 0.0005449335664335663, + "loss": 3.5803, + "step": 15850 + }, + { + "epoch": 4.630205603121905, + "grad_norm": 0.31655800342559814, + "learning_rate": 0.0005447587412587412, + "loss": 3.5793, + "step": 15900 + }, + { + "epoch": 4.644766730735628, + "grad_norm": 0.3285786509513855, + "learning_rate": 0.000544583916083916, + "loss": 3.5886, + "step": 15950 + }, + { + "epoch": 4.659327858349351, + "grad_norm": 0.3272700607776642, + "learning_rate": 0.0005444090909090908, + "loss": 3.5779, + "step": 16000 + }, + { + "epoch": 4.659327858349351, + "eval_accuracy": 0.3580464208423046, + "eval_loss": 3.6513254642486572, + "eval_runtime": 179.2925, + "eval_samples_per_second": 92.815, + "eval_steps_per_second": 5.806, + "step": 16000 + }, + { + "epoch": 4.673888985963073, + "grad_norm": 0.3093654215335846, + "learning_rate": 0.0005442342657342657, + "loss": 3.5759, + "step": 16050 + }, + { + "epoch": 4.6884501135767955, + "grad_norm": 0.3177140951156616, + "learning_rate": 0.0005440594405594405, + "loss": 3.5715, + "step": 16100 + }, + { + "epoch": 4.703011241190518, + "grad_norm": 0.32796481251716614, + "learning_rate": 0.0005438846153846153, + "loss": 3.5836, + "step": 16150 + }, + { + "epoch": 4.71757236880424, + "grad_norm": 0.32273897528648376, + "learning_rate": 0.0005437097902097901, + "loss": 3.5795, + "step": 16200 + }, + { + "epoch": 4.732133496417963, + "grad_norm": 0.3069426715373993, + "learning_rate": 0.0005435349650349651, + "loss": 3.5762, + "step": 16250 + }, + { + "epoch": 4.746694624031685, + "grad_norm": 0.3387124240398407, + "learning_rate": 0.0005433601398601397, + "loss": 3.5745, + "step": 16300 + }, + { + "epoch": 4.7612557516454075, + "grad_norm": 0.3111344873905182, + "learning_rate": 0.0005431853146853147, + "loss": 3.5698, + "step": 16350 + }, + { + "epoch": 4.77581687925913, + "grad_norm": 0.311276912689209, + "learning_rate": 0.0005430104895104895, + "loss": 3.5799, + "step": 16400 + }, + { + "epoch": 4.790378006872852, + "grad_norm": 0.3338645398616791, + "learning_rate": 0.0005428356643356643, + "loss": 3.5812, + "step": 16450 + }, + { + "epoch": 4.804939134486575, + "grad_norm": Infinity, + "learning_rate": 0.0005426608391608391, + "loss": 3.58, + "step": 16500 + }, + { + "epoch": 4.819500262100297, + "grad_norm": 0.3227423429489136, + "learning_rate": 0.000542486013986014, + "loss": 3.5685, + "step": 16550 + }, + { + "epoch": 4.834061389714019, + "grad_norm": 0.32446616888046265, + "learning_rate": 0.0005423111888111888, + "loss": 3.5818, + "step": 16600 + }, + { + "epoch": 4.848622517327742, + "grad_norm": 0.3579038977622986, + "learning_rate": 0.0005421363636363636, + "loss": 3.5862, + "step": 16650 + }, + { + "epoch": 4.863183644941464, + "grad_norm": 0.3191075026988983, + "learning_rate": 0.0005419615384615385, + "loss": 3.5853, + "step": 16700 + }, + { + "epoch": 4.877744772555187, + "grad_norm": 0.3210720717906952, + "learning_rate": 0.0005417867132867133, + "loss": 3.576, + "step": 16750 + }, + { + "epoch": 4.892305900168909, + "grad_norm": 0.3151341676712036, + "learning_rate": 0.0005416118881118881, + "loss": 3.5861, + "step": 16800 + }, + { + "epoch": 4.906867027782631, + "grad_norm": 0.31691774725914, + "learning_rate": 0.0005414370629370629, + "loss": 3.5741, + "step": 16850 + }, + { + "epoch": 4.921428155396354, + "grad_norm": 0.318175345659256, + "learning_rate": 0.0005412622377622378, + "loss": 3.5768, + "step": 16900 + }, + { + "epoch": 4.935989283010076, + "grad_norm": 0.3188709616661072, + "learning_rate": 0.0005410874125874126, + "loss": 3.576, + "step": 16950 + }, + { + "epoch": 4.950550410623799, + "grad_norm": 0.33716508746147156, + "learning_rate": 0.0005409125874125874, + "loss": 3.5792, + "step": 17000 + }, + { + "epoch": 4.950550410623799, + "eval_accuracy": 0.3592166378492938, + "eval_loss": 3.6367878913879395, + "eval_runtime": 179.3955, + "eval_samples_per_second": 92.762, + "eval_steps_per_second": 5.803, + "step": 17000 + }, + { + "epoch": 4.9651115382375215, + "grad_norm": 0.32607048749923706, + "learning_rate": 0.0005407377622377622, + "loss": 3.5881, + "step": 17050 + }, + { + "epoch": 4.979672665851243, + "grad_norm": 0.31015822291374207, + "learning_rate": 0.000540562937062937, + "loss": 3.5824, + "step": 17100 + }, + { + "epoch": 4.994233793464966, + "grad_norm": 0.3248412609100342, + "learning_rate": 0.0005403881118881118, + "loss": 3.5774, + "step": 17150 + }, + { + "epoch": 5.008736676568233, + "grad_norm": 0.317827045917511, + "learning_rate": 0.0005402132867132867, + "loss": 3.5138, + "step": 17200 + }, + { + "epoch": 5.023297804181956, + "grad_norm": 0.347983717918396, + "learning_rate": 0.0005400384615384615, + "loss": 3.4802, + "step": 17250 + }, + { + "epoch": 5.037858931795678, + "grad_norm": 0.3314956724643707, + "learning_rate": 0.0005398636363636363, + "loss": 3.4751, + "step": 17300 + }, + { + "epoch": 5.052420059409401, + "grad_norm": 0.34507569670677185, + "learning_rate": 0.0005396888111888111, + "loss": 3.4727, + "step": 17350 + }, + { + "epoch": 5.066981187023123, + "grad_norm": 0.31889015436172485, + "learning_rate": 0.000539513986013986, + "loss": 3.4871, + "step": 17400 + }, + { + "epoch": 5.081542314636845, + "grad_norm": 0.3292185962200165, + "learning_rate": 0.0005393391608391608, + "loss": 3.4871, + "step": 17450 + }, + { + "epoch": 5.096103442250568, + "grad_norm": 0.3279465436935425, + "learning_rate": 0.0005391643356643356, + "loss": 3.4936, + "step": 17500 + }, + { + "epoch": 5.110664569864291, + "grad_norm": 0.3154655694961548, + "learning_rate": 0.0005389895104895105, + "loss": 3.4811, + "step": 17550 + }, + { + "epoch": 5.125225697478013, + "grad_norm": 0.33116382360458374, + "learning_rate": 0.0005388146853146853, + "loss": 3.501, + "step": 17600 + }, + { + "epoch": 5.139786825091735, + "grad_norm": 0.3599289357662201, + "learning_rate": 0.0005386398601398601, + "loss": 3.4892, + "step": 17650 + }, + { + "epoch": 5.154347952705457, + "grad_norm": 0.32488229870796204, + "learning_rate": 0.0005384650349650349, + "loss": 3.4957, + "step": 17700 + }, + { + "epoch": 5.16890908031918, + "grad_norm": 0.3252679109573364, + "learning_rate": 0.0005382902097902098, + "loss": 3.4945, + "step": 17750 + }, + { + "epoch": 5.183470207932903, + "grad_norm": 0.31277021765708923, + "learning_rate": 0.0005381153846153845, + "loss": 3.5013, + "step": 17800 + }, + { + "epoch": 5.1980313355466246, + "grad_norm": 0.34036797285079956, + "learning_rate": 0.0005379405594405594, + "loss": 3.507, + "step": 17850 + }, + { + "epoch": 5.212592463160347, + "grad_norm": 0.33982405066490173, + "learning_rate": 0.0005377657342657342, + "loss": 3.5075, + "step": 17900 + }, + { + "epoch": 5.227153590774069, + "grad_norm": 0.33606475591659546, + "learning_rate": 0.000537590909090909, + "loss": 3.4981, + "step": 17950 + }, + { + "epoch": 5.241714718387792, + "grad_norm": 0.31139010190963745, + "learning_rate": 0.0005374160839160838, + "loss": 3.5045, + "step": 18000 + }, + { + "epoch": 5.241714718387792, + "eval_accuracy": 0.35992128464920126, + "eval_loss": 3.639274835586548, + "eval_runtime": 179.4938, + "eval_samples_per_second": 92.711, + "eval_steps_per_second": 5.8, + "step": 18000 + }, + { + "epoch": 5.256275846001515, + "grad_norm": 0.355369508266449, + "learning_rate": 0.0005372412587412587, + "loss": 3.51, + "step": 18050 + }, + { + "epoch": 5.2708369736152365, + "grad_norm": 0.3412383496761322, + "learning_rate": 0.0005370664335664335, + "loss": 3.5198, + "step": 18100 + }, + { + "epoch": 5.285398101228959, + "grad_norm": 0.3562864065170288, + "learning_rate": 0.0005368916083916083, + "loss": 3.5179, + "step": 18150 + }, + { + "epoch": 5.299959228842681, + "grad_norm": 0.3295607268810272, + "learning_rate": 0.0005367167832167832, + "loss": 3.5125, + "step": 18200 + }, + { + "epoch": 5.314520356456404, + "grad_norm": 0.301607608795166, + "learning_rate": 0.000536541958041958, + "loss": 3.5117, + "step": 18250 + }, + { + "epoch": 5.329081484070127, + "grad_norm": 0.3317857086658478, + "learning_rate": 0.0005363671328671328, + "loss": 3.5257, + "step": 18300 + }, + { + "epoch": 5.3436426116838485, + "grad_norm": 0.31705543398857117, + "learning_rate": 0.0005361923076923076, + "loss": 3.5187, + "step": 18350 + }, + { + "epoch": 5.358203739297571, + "grad_norm": 0.3254964351654053, + "learning_rate": 0.0005360174825174825, + "loss": 3.5162, + "step": 18400 + }, + { + "epoch": 5.372764866911294, + "grad_norm": 0.31116539239883423, + "learning_rate": 0.0005358426573426573, + "loss": 3.512, + "step": 18450 + }, + { + "epoch": 5.387325994525016, + "grad_norm": 0.3314308524131775, + "learning_rate": 0.0005356678321678321, + "loss": 3.5199, + "step": 18500 + }, + { + "epoch": 5.401887122138739, + "grad_norm": 0.3243958652019501, + "learning_rate": 0.0005354930069930069, + "loss": 3.5183, + "step": 18550 + }, + { + "epoch": 5.41644824975246, + "grad_norm": 0.3620099723339081, + "learning_rate": 0.0005353181818181817, + "loss": 3.5152, + "step": 18600 + }, + { + "epoch": 5.431009377366183, + "grad_norm": 0.30956247448921204, + "learning_rate": 0.0005351433566433565, + "loss": 3.534, + "step": 18650 + }, + { + "epoch": 5.445570504979906, + "grad_norm": 0.33031249046325684, + "learning_rate": 0.0005349685314685314, + "loss": 3.511, + "step": 18700 + }, + { + "epoch": 5.460131632593628, + "grad_norm": 0.32894542813301086, + "learning_rate": 0.0005347937062937062, + "loss": 3.5217, + "step": 18750 + }, + { + "epoch": 5.4746927602073505, + "grad_norm": 0.36818283796310425, + "learning_rate": 0.000534618881118881, + "loss": 3.5188, + "step": 18800 + }, + { + "epoch": 5.489253887821073, + "grad_norm": 0.34024327993392944, + "learning_rate": 0.0005344440559440559, + "loss": 3.5226, + "step": 18850 + }, + { + "epoch": 5.503815015434795, + "grad_norm": 0.3341643214225769, + "learning_rate": 0.0005342692307692307, + "loss": 3.5196, + "step": 18900 + }, + { + "epoch": 5.518376143048518, + "grad_norm": 0.3215097188949585, + "learning_rate": 0.0005340944055944055, + "loss": 3.5239, + "step": 18950 + }, + { + "epoch": 5.53293727066224, + "grad_norm": 0.3380003273487091, + "learning_rate": 0.0005339195804195803, + "loss": 3.5331, + "step": 19000 + }, + { + "epoch": 5.53293727066224, + "eval_accuracy": 0.36069837177433284, + "eval_loss": 3.6282050609588623, + "eval_runtime": 179.5413, + "eval_samples_per_second": 92.686, + "eval_steps_per_second": 5.798, + "step": 19000 + }, + { + "epoch": 5.5474983982759625, + "grad_norm": 0.33044567704200745, + "learning_rate": 0.0005337447552447552, + "loss": 3.5279, + "step": 19050 + }, + { + "epoch": 5.562059525889685, + "grad_norm": 0.33487701416015625, + "learning_rate": 0.00053356993006993, + "loss": 3.527, + "step": 19100 + }, + { + "epoch": 5.576620653503407, + "grad_norm": 0.33163321018218994, + "learning_rate": 0.0005333951048951048, + "loss": 3.5208, + "step": 19150 + }, + { + "epoch": 5.59118178111713, + "grad_norm": 0.31409138441085815, + "learning_rate": 0.0005332202797202796, + "loss": 3.5294, + "step": 19200 + }, + { + "epoch": 5.605742908730852, + "grad_norm": 0.32347944378852844, + "learning_rate": 0.0005330454545454546, + "loss": 3.5235, + "step": 19250 + }, + { + "epoch": 5.620304036344574, + "grad_norm": 0.330264687538147, + "learning_rate": 0.0005328706293706292, + "loss": 3.5207, + "step": 19300 + }, + { + "epoch": 5.634865163958297, + "grad_norm": 0.3169718384742737, + "learning_rate": 0.0005326958041958042, + "loss": 3.5237, + "step": 19350 + }, + { + "epoch": 5.649426291572019, + "grad_norm": 0.3469524085521698, + "learning_rate": 0.000532520979020979, + "loss": 3.523, + "step": 19400 + }, + { + "epoch": 5.663987419185742, + "grad_norm": 0.3340586721897125, + "learning_rate": 0.0005323461538461538, + "loss": 3.5288, + "step": 19450 + }, + { + "epoch": 5.6785485467994645, + "grad_norm": 0.3267485499382019, + "learning_rate": 0.0005321713286713287, + "loss": 3.5171, + "step": 19500 + }, + { + "epoch": 5.693109674413186, + "grad_norm": 0.3368189036846161, + "learning_rate": 0.0005319965034965035, + "loss": 3.5313, + "step": 19550 + }, + { + "epoch": 5.707670802026909, + "grad_norm": 0.35651063919067383, + "learning_rate": 0.0005318216783216783, + "loss": 3.5366, + "step": 19600 + }, + { + "epoch": 5.722231929640631, + "grad_norm": 0.3218107223510742, + "learning_rate": 0.0005316468531468531, + "loss": 3.5248, + "step": 19650 + }, + { + "epoch": 5.736793057254354, + "grad_norm": 0.32096830010414124, + "learning_rate": 0.000531472027972028, + "loss": 3.5275, + "step": 19700 + }, + { + "epoch": 5.7513541848680765, + "grad_norm": 0.31366539001464844, + "learning_rate": 0.0005312972027972028, + "loss": 3.5388, + "step": 19750 + }, + { + "epoch": 5.765915312481798, + "grad_norm": 0.31841805577278137, + "learning_rate": 0.0005311223776223776, + "loss": 3.5238, + "step": 19800 + }, + { + "epoch": 5.780476440095521, + "grad_norm": 0.3162098824977875, + "learning_rate": 0.0005309475524475524, + "loss": 3.5399, + "step": 19850 + }, + { + "epoch": 5.795037567709244, + "grad_norm": 0.34669944643974304, + "learning_rate": 0.0005307727272727273, + "loss": 3.5248, + "step": 19900 + }, + { + "epoch": 5.809598695322966, + "grad_norm": 0.32545655965805054, + "learning_rate": 0.0005305979020979021, + "loss": 3.5316, + "step": 19950 + }, + { + "epoch": 5.824159822936688, + "grad_norm": 0.3101173937320709, + "learning_rate": 0.0005304230769230769, + "loss": 3.5289, + "step": 20000 + }, + { + "epoch": 5.824159822936688, + "eval_accuracy": 0.3619148047680316, + "eval_loss": 3.6172592639923096, + "eval_runtime": 179.6401, + "eval_samples_per_second": 92.635, + "eval_steps_per_second": 5.795, + "step": 20000 + }, + { + "epoch": 5.83872095055041, + "grad_norm": 0.30793216824531555, + "learning_rate": 0.0005302482517482517, + "loss": 3.5286, + "step": 20050 + }, + { + "epoch": 5.853282078164133, + "grad_norm": 0.301077663898468, + "learning_rate": 0.0005300734265734265, + "loss": 3.5269, + "step": 20100 + }, + { + "epoch": 5.867843205777856, + "grad_norm": 0.3162022531032562, + "learning_rate": 0.0005298986013986013, + "loss": 3.5304, + "step": 20150 + }, + { + "epoch": 5.882404333391578, + "grad_norm": 0.3171811103820801, + "learning_rate": 0.0005297237762237762, + "loss": 3.5286, + "step": 20200 + }, + { + "epoch": 5.8969654610053, + "grad_norm": 0.33995500206947327, + "learning_rate": 0.000529548951048951, + "loss": 3.5373, + "step": 20250 + }, + { + "epoch": 5.911526588619022, + "grad_norm": 0.32125958800315857, + "learning_rate": 0.0005293741258741258, + "loss": 3.5379, + "step": 20300 + }, + { + "epoch": 5.926087716232745, + "grad_norm": 0.3155665397644043, + "learning_rate": 0.0005291993006993007, + "loss": 3.5392, + "step": 20350 + }, + { + "epoch": 5.940648843846468, + "grad_norm": 0.31241375207901, + "learning_rate": 0.0005290244755244755, + "loss": 3.5348, + "step": 20400 + }, + { + "epoch": 5.95520997146019, + "grad_norm": 0.2965865731239319, + "learning_rate": 0.0005288496503496503, + "loss": 3.5434, + "step": 20450 + }, + { + "epoch": 5.969771099073912, + "grad_norm": 0.304505854845047, + "learning_rate": 0.0005286748251748251, + "loss": 3.5298, + "step": 20500 + }, + { + "epoch": 5.984332226687634, + "grad_norm": 0.3139273226261139, + "learning_rate": 0.0005285, + "loss": 3.5242, + "step": 20550 + }, + { + "epoch": 5.998893354301357, + "grad_norm": 0.33269649744033813, + "learning_rate": 0.0005283251748251748, + "loss": 3.5254, + "step": 20600 + }, + { + "epoch": 6.013396237404625, + "grad_norm": 0.32686853408813477, + "learning_rate": 0.0005281503496503496, + "loss": 3.4371, + "step": 20650 + }, + { + "epoch": 6.027957365018347, + "grad_norm": 0.3370496928691864, + "learning_rate": 0.0005279755244755244, + "loss": 3.4257, + "step": 20700 + }, + { + "epoch": 6.04251849263207, + "grad_norm": 0.35626882314682007, + "learning_rate": 0.0005278006993006993, + "loss": 3.4318, + "step": 20750 + }, + { + "epoch": 6.0570796202457915, + "grad_norm": 0.3384275734424591, + "learning_rate": 0.000527625874125874, + "loss": 3.4331, + "step": 20800 + }, + { + "epoch": 6.071640747859514, + "grad_norm": 0.33348149061203003, + "learning_rate": 0.0005274510489510489, + "loss": 3.4296, + "step": 20850 + }, + { + "epoch": 6.086201875473237, + "grad_norm": 0.3619717061519623, + "learning_rate": 0.0005272762237762238, + "loss": 3.4429, + "step": 20900 + }, + { + "epoch": 6.100763003086959, + "grad_norm": 0.33701643347740173, + "learning_rate": 0.0005271013986013985, + "loss": 3.4477, + "step": 20950 + }, + { + "epoch": 6.115324130700682, + "grad_norm": 0.3502415418624878, + "learning_rate": 0.0005269265734265734, + "loss": 3.4422, + "step": 21000 + }, + { + "epoch": 6.115324130700682, + "eval_accuracy": 0.3622793583527634, + "eval_loss": 3.619753122329712, + "eval_runtime": 179.5794, + "eval_samples_per_second": 92.667, + "eval_steps_per_second": 5.797, + "step": 21000 + }, + { + "epoch": 6.1298852583144035, + "grad_norm": 0.32658058404922485, + "learning_rate": 0.0005267517482517482, + "loss": 3.4422, + "step": 21050 + }, + { + "epoch": 6.144446385928126, + "grad_norm": 0.3188963234424591, + "learning_rate": 0.000526576923076923, + "loss": 3.449, + "step": 21100 + }, + { + "epoch": 6.159007513541849, + "grad_norm": 0.3194127678871155, + "learning_rate": 0.0005264020979020978, + "loss": 3.4458, + "step": 21150 + }, + { + "epoch": 6.173568641155571, + "grad_norm": 0.3370357155799866, + "learning_rate": 0.0005262272727272727, + "loss": 3.4573, + "step": 21200 + }, + { + "epoch": 6.1881297687692935, + "grad_norm": 0.3297397196292877, + "learning_rate": 0.0005260524475524475, + "loss": 3.4578, + "step": 21250 + }, + { + "epoch": 6.202690896383016, + "grad_norm": 0.33779558539390564, + "learning_rate": 0.0005258776223776223, + "loss": 3.4566, + "step": 21300 + }, + { + "epoch": 6.217252023996738, + "grad_norm": 0.34972041845321655, + "learning_rate": 0.0005257027972027971, + "loss": 3.4734, + "step": 21350 + }, + { + "epoch": 6.231813151610461, + "grad_norm": 0.3232443332672119, + "learning_rate": 0.000525527972027972, + "loss": 3.4658, + "step": 21400 + }, + { + "epoch": 6.246374279224183, + "grad_norm": 0.35180407762527466, + "learning_rate": 0.0005253531468531468, + "loss": 3.4586, + "step": 21450 + }, + { + "epoch": 6.2609354068379055, + "grad_norm": 0.3353635370731354, + "learning_rate": 0.0005251783216783216, + "loss": 3.4643, + "step": 21500 + }, + { + "epoch": 6.275496534451628, + "grad_norm": 0.32790806889533997, + "learning_rate": 0.0005250034965034965, + "loss": 3.4607, + "step": 21550 + }, + { + "epoch": 6.29005766206535, + "grad_norm": 0.32156580686569214, + "learning_rate": 0.0005248286713286712, + "loss": 3.472, + "step": 21600 + }, + { + "epoch": 6.304618789679073, + "grad_norm": 0.3282783329486847, + "learning_rate": 0.0005246538461538461, + "loss": 3.4571, + "step": 21650 + }, + { + "epoch": 6.319179917292795, + "grad_norm": 0.3265901803970337, + "learning_rate": 0.0005244790209790209, + "loss": 3.4671, + "step": 21700 + }, + { + "epoch": 6.3337410449065175, + "grad_norm": 0.35069194436073303, + "learning_rate": 0.0005243041958041957, + "loss": 3.4685, + "step": 21750 + }, + { + "epoch": 6.34830217252024, + "grad_norm": 0.32015976309776306, + "learning_rate": 0.0005241293706293705, + "loss": 3.4721, + "step": 21800 + }, + { + "epoch": 6.362863300133962, + "grad_norm": 0.31217172741889954, + "learning_rate": 0.0005239545454545454, + "loss": 3.4812, + "step": 21850 + }, + { + "epoch": 6.377424427747685, + "grad_norm": 0.33282867074012756, + "learning_rate": 0.0005237797202797202, + "loss": 3.4727, + "step": 21900 + }, + { + "epoch": 6.391985555361408, + "grad_norm": 0.33826690912246704, + "learning_rate": 0.000523604895104895, + "loss": 3.467, + "step": 21950 + }, + { + "epoch": 6.406546682975129, + "grad_norm": 0.3465859293937683, + "learning_rate": 0.0005234300699300698, + "loss": 3.4741, + "step": 22000 + }, + { + "epoch": 6.406546682975129, + "eval_accuracy": 0.36281983844161103, + "eval_loss": 3.614673376083374, + "eval_runtime": 179.7434, + "eval_samples_per_second": 92.582, + "eval_steps_per_second": 5.792, + "step": 22000 + }, + { + "epoch": 6.421107810588852, + "grad_norm": 0.3159697353839874, + "learning_rate": 0.0005232552447552447, + "loss": 3.4675, + "step": 22050 + }, + { + "epoch": 6.435668938202574, + "grad_norm": 0.32068994641304016, + "learning_rate": 0.0005230804195804195, + "loss": 3.4754, + "step": 22100 + }, + { + "epoch": 6.450230065816297, + "grad_norm": 0.3233058750629425, + "learning_rate": 0.0005229055944055943, + "loss": 3.4766, + "step": 22150 + }, + { + "epoch": 6.4647911934300195, + "grad_norm": 0.3281400501728058, + "learning_rate": 0.0005227307692307691, + "loss": 3.4763, + "step": 22200 + }, + { + "epoch": 6.479352321043741, + "grad_norm": 0.3255999684333801, + "learning_rate": 0.0005225559440559441, + "loss": 3.4758, + "step": 22250 + }, + { + "epoch": 6.493913448657464, + "grad_norm": 0.3177129328250885, + "learning_rate": 0.0005223811188811189, + "loss": 3.4863, + "step": 22300 + }, + { + "epoch": 6.508474576271187, + "grad_norm": 0.3292180895805359, + "learning_rate": 0.0005222062937062937, + "loss": 3.4909, + "step": 22350 + }, + { + "epoch": 6.523035703884909, + "grad_norm": 0.316459059715271, + "learning_rate": 0.0005220314685314686, + "loss": 3.4942, + "step": 22400 + }, + { + "epoch": 6.5375968314986315, + "grad_norm": 0.3164568543434143, + "learning_rate": 0.0005218566433566433, + "loss": 3.4985, + "step": 22450 + }, + { + "epoch": 6.552157959112353, + "grad_norm": 0.3300154209136963, + "learning_rate": 0.0005216818181818182, + "loss": 3.4829, + "step": 22500 + }, + { + "epoch": 6.566719086726076, + "grad_norm": 0.3508749306201935, + "learning_rate": 0.000521506993006993, + "loss": 3.4796, + "step": 22550 + }, + { + "epoch": 6.581280214339799, + "grad_norm": 0.32922637462615967, + "learning_rate": 0.0005213321678321678, + "loss": 3.4965, + "step": 22600 + }, + { + "epoch": 6.595841341953521, + "grad_norm": 0.3307730555534363, + "learning_rate": 0.0005211573426573426, + "loss": 3.4825, + "step": 22650 + }, + { + "epoch": 6.610402469567243, + "grad_norm": 0.3202342092990875, + "learning_rate": 0.0005209825174825175, + "loss": 3.4849, + "step": 22700 + }, + { + "epoch": 6.624963597180965, + "grad_norm": 0.3231477737426758, + "learning_rate": 0.0005208076923076923, + "loss": 3.4861, + "step": 22750 + }, + { + "epoch": 6.639524724794688, + "grad_norm": 0.32124945521354675, + "learning_rate": 0.0005206328671328671, + "loss": 3.4998, + "step": 22800 + }, + { + "epoch": 6.654085852408411, + "grad_norm": 0.35052698850631714, + "learning_rate": 0.0005204580419580419, + "loss": 3.4907, + "step": 22850 + }, + { + "epoch": 6.668646980022133, + "grad_norm": 0.3346177637577057, + "learning_rate": 0.0005202832167832168, + "loss": 3.4881, + "step": 22900 + }, + { + "epoch": 6.683208107635855, + "grad_norm": 0.30874842405319214, + "learning_rate": 0.0005201083916083916, + "loss": 3.4943, + "step": 22950 + }, + { + "epoch": 6.697769235249577, + "grad_norm": 0.3318243622779846, + "learning_rate": 0.0005199335664335664, + "loss": 3.4904, + "step": 23000 + }, + { + "epoch": 6.697769235249577, + "eval_accuracy": 0.3636449055224106, + "eval_loss": 3.6039533615112305, + "eval_runtime": 179.591, + "eval_samples_per_second": 92.661, + "eval_steps_per_second": 5.797, + "step": 23000 + }, + { + "epoch": 6.7123303628633, + "grad_norm": 0.3388109803199768, + "learning_rate": 0.0005197587412587413, + "loss": 3.5026, + "step": 23050 + }, + { + "epoch": 6.726891490477023, + "grad_norm": 0.29118219017982483, + "learning_rate": 0.0005195839160839161, + "loss": 3.4907, + "step": 23100 + }, + { + "epoch": 6.741452618090745, + "grad_norm": 0.35532817244529724, + "learning_rate": 0.0005194090909090909, + "loss": 3.4896, + "step": 23150 + }, + { + "epoch": 6.756013745704467, + "grad_norm": 0.3368262052536011, + "learning_rate": 0.0005192342657342657, + "loss": 3.4822, + "step": 23200 + }, + { + "epoch": 6.77057487331819, + "grad_norm": 0.33118659257888794, + "learning_rate": 0.0005190594405594405, + "loss": 3.4901, + "step": 23250 + }, + { + "epoch": 6.785136000931912, + "grad_norm": 0.3284784257411957, + "learning_rate": 0.0005188846153846153, + "loss": 3.4991, + "step": 23300 + }, + { + "epoch": 6.799697128545635, + "grad_norm": 0.3228989243507385, + "learning_rate": 0.0005187097902097902, + "loss": 3.4946, + "step": 23350 + }, + { + "epoch": 6.814258256159357, + "grad_norm": 0.31879404187202454, + "learning_rate": 0.000518534965034965, + "loss": 3.4786, + "step": 23400 + }, + { + "epoch": 6.828819383773079, + "grad_norm": 0.3200366795063019, + "learning_rate": 0.0005183601398601398, + "loss": 3.4905, + "step": 23450 + }, + { + "epoch": 6.843380511386802, + "grad_norm": 0.33181434869766235, + "learning_rate": 0.0005181853146853146, + "loss": 3.4843, + "step": 23500 + }, + { + "epoch": 6.857941639000524, + "grad_norm": 0.3142688572406769, + "learning_rate": 0.0005180104895104895, + "loss": 3.495, + "step": 23550 + }, + { + "epoch": 6.872502766614247, + "grad_norm": 0.34620487689971924, + "learning_rate": 0.0005178356643356643, + "loss": 3.4968, + "step": 23600 + }, + { + "epoch": 6.887063894227969, + "grad_norm": 0.32809966802597046, + "learning_rate": 0.0005176608391608391, + "loss": 3.4826, + "step": 23650 + }, + { + "epoch": 6.901625021841691, + "grad_norm": 0.30796363949775696, + "learning_rate": 0.000517486013986014, + "loss": 3.492, + "step": 23700 + }, + { + "epoch": 6.916186149455414, + "grad_norm": 0.3204144835472107, + "learning_rate": 0.0005173111888111888, + "loss": 3.4919, + "step": 23750 + }, + { + "epoch": 6.930747277069136, + "grad_norm": 0.3202858865261078, + "learning_rate": 0.0005171363636363636, + "loss": 3.4953, + "step": 23800 + }, + { + "epoch": 6.945308404682859, + "grad_norm": 0.32442307472229004, + "learning_rate": 0.0005169615384615384, + "loss": 3.4914, + "step": 23850 + }, + { + "epoch": 6.959869532296581, + "grad_norm": 0.3379514515399933, + "learning_rate": 0.0005167867132867133, + "loss": 3.4977, + "step": 23900 + }, + { + "epoch": 6.974430659910303, + "grad_norm": 0.35075902938842773, + "learning_rate": 0.000516611888111888, + "loss": 3.5005, + "step": 23950 + }, + { + "epoch": 6.988991787524026, + "grad_norm": 0.33145564794540405, + "learning_rate": 0.0005164370629370629, + "loss": 3.501, + "step": 24000 + }, + { + "epoch": 6.988991787524026, + "eval_accuracy": 0.36449290419966907, + "eval_loss": 3.5937235355377197, + "eval_runtime": 179.4257, + "eval_samples_per_second": 92.746, + "eval_steps_per_second": 5.802, + "step": 24000 + }, + { + "epoch": 7.003494670627293, + "grad_norm": 0.32831260561943054, + "learning_rate": 0.0005162622377622377, + "loss": 3.4776, + "step": 24050 + }, + { + "epoch": 7.018055798241016, + "grad_norm": 0.337808221578598, + "learning_rate": 0.0005160874125874125, + "loss": 3.388, + "step": 24100 + }, + { + "epoch": 7.032616925854738, + "grad_norm": 0.31366923451423645, + "learning_rate": 0.0005159125874125873, + "loss": 3.3968, + "step": 24150 + }, + { + "epoch": 7.0471780534684605, + "grad_norm": 0.34105607867240906, + "learning_rate": 0.0005157377622377622, + "loss": 3.3822, + "step": 24200 + }, + { + "epoch": 7.061739181082183, + "grad_norm": 0.3203040063381195, + "learning_rate": 0.000515562937062937, + "loss": 3.3963, + "step": 24250 + }, + { + "epoch": 7.076300308695905, + "grad_norm": 0.329708993434906, + "learning_rate": 0.0005153881118881118, + "loss": 3.4021, + "step": 24300 + }, + { + "epoch": 7.090861436309628, + "grad_norm": 0.33745211362838745, + "learning_rate": 0.0005152132867132867, + "loss": 3.3907, + "step": 24350 + }, + { + "epoch": 7.105422563923351, + "grad_norm": 0.33179929852485657, + "learning_rate": 0.0005150384615384615, + "loss": 3.3883, + "step": 24400 + }, + { + "epoch": 7.1199836915370724, + "grad_norm": 0.3846040368080139, + "learning_rate": 0.0005148636363636363, + "loss": 3.4137, + "step": 24450 + }, + { + "epoch": 7.134544819150795, + "grad_norm": 0.3433890640735626, + "learning_rate": 0.0005146888111888111, + "loss": 3.4165, + "step": 24500 + }, + { + "epoch": 7.149105946764517, + "grad_norm": 0.33867666125297546, + "learning_rate": 0.000514513986013986, + "loss": 3.4308, + "step": 24550 + }, + { + "epoch": 7.16366707437824, + "grad_norm": 0.3347140848636627, + "learning_rate": 0.0005143391608391608, + "loss": 3.4058, + "step": 24600 + }, + { + "epoch": 7.1782282019919625, + "grad_norm": 0.357448011636734, + "learning_rate": 0.0005141643356643356, + "loss": 3.419, + "step": 24650 + }, + { + "epoch": 7.192789329605684, + "grad_norm": 0.33567482233047485, + "learning_rate": 0.0005139895104895104, + "loss": 3.4155, + "step": 24700 + }, + { + "epoch": 7.207350457219407, + "grad_norm": 0.35268595814704895, + "learning_rate": 0.0005138146853146852, + "loss": 3.424, + "step": 24750 + }, + { + "epoch": 7.22191158483313, + "grad_norm": 0.33590567111968994, + "learning_rate": 0.00051363986013986, + "loss": 3.4261, + "step": 24800 + }, + { + "epoch": 7.236472712446852, + "grad_norm": 0.3247842788696289, + "learning_rate": 0.0005134650349650349, + "loss": 3.4193, + "step": 24850 + }, + { + "epoch": 7.2510338400605745, + "grad_norm": 0.3556370735168457, + "learning_rate": 0.0005132902097902097, + "loss": 3.4235, + "step": 24900 + }, + { + "epoch": 7.265594967674296, + "grad_norm": 0.34552738070487976, + "learning_rate": 0.0005131153846153845, + "loss": 3.4244, + "step": 24950 + }, + { + "epoch": 7.280156095288019, + "grad_norm": 0.31991732120513916, + "learning_rate": 0.0005129405594405594, + "loss": 3.4339, + "step": 25000 + }, + { + "epoch": 7.280156095288019, + "eval_accuracy": 0.3640523823517963, + "eval_loss": 3.603454828262329, + "eval_runtime": 179.613, + "eval_samples_per_second": 92.649, + "eval_steps_per_second": 5.796, + "step": 25000 + }, + { + "epoch": 7.294717222901742, + "grad_norm": 0.32443350553512573, + "learning_rate": 0.0005127657342657342, + "loss": 3.4309, + "step": 25050 + }, + { + "epoch": 7.309278350515464, + "grad_norm": 0.346880704164505, + "learning_rate": 0.000512590909090909, + "loss": 3.4366, + "step": 25100 + }, + { + "epoch": 7.3238394781291865, + "grad_norm": 0.33890458941459656, + "learning_rate": 0.0005124160839160838, + "loss": 3.4332, + "step": 25150 + }, + { + "epoch": 7.338400605742908, + "grad_norm": 0.32790401577949524, + "learning_rate": 0.0005122412587412588, + "loss": 3.4405, + "step": 25200 + }, + { + "epoch": 7.352961733356631, + "grad_norm": 0.31733909249305725, + "learning_rate": 0.0005120664335664336, + "loss": 3.4426, + "step": 25250 + }, + { + "epoch": 7.367522860970354, + "grad_norm": 0.3427425026893616, + "learning_rate": 0.0005118916083916084, + "loss": 3.4297, + "step": 25300 + }, + { + "epoch": 7.382083988584076, + "grad_norm": 0.3265109062194824, + "learning_rate": 0.0005117167832167832, + "loss": 3.44, + "step": 25350 + }, + { + "epoch": 7.396645116197798, + "grad_norm": 0.32305601239204407, + "learning_rate": 0.0005115419580419581, + "loss": 3.4458, + "step": 25400 + }, + { + "epoch": 7.411206243811521, + "grad_norm": 0.3397580087184906, + "learning_rate": 0.0005113671328671328, + "loss": 3.4476, + "step": 25450 + }, + { + "epoch": 7.425767371425243, + "grad_norm": 0.33656251430511475, + "learning_rate": 0.0005111923076923077, + "loss": 3.4502, + "step": 25500 + }, + { + "epoch": 7.440328499038966, + "grad_norm": 0.3773593008518219, + "learning_rate": 0.0005110174825174825, + "loss": 3.4454, + "step": 25550 + }, + { + "epoch": 7.454889626652688, + "grad_norm": 0.31921395659446716, + "learning_rate": 0.0005108426573426573, + "loss": 3.4475, + "step": 25600 + }, + { + "epoch": 7.46945075426641, + "grad_norm": 0.3150818347930908, + "learning_rate": 0.0005106678321678321, + "loss": 3.4446, + "step": 25650 + }, + { + "epoch": 7.484011881880133, + "grad_norm": 0.3486284017562866, + "learning_rate": 0.000510493006993007, + "loss": 3.436, + "step": 25700 + }, + { + "epoch": 7.498573009493855, + "grad_norm": 0.33967137336730957, + "learning_rate": 0.0005103181818181818, + "loss": 3.4475, + "step": 25750 + }, + { + "epoch": 7.513134137107578, + "grad_norm": 0.34054380655288696, + "learning_rate": 0.0005101433566433566, + "loss": 3.4485, + "step": 25800 + }, + { + "epoch": 7.5276952647213005, + "grad_norm": 0.34166619181632996, + "learning_rate": 0.0005099685314685315, + "loss": 3.4543, + "step": 25850 + }, + { + "epoch": 7.542256392335022, + "grad_norm": 0.3142751455307007, + "learning_rate": 0.0005097937062937063, + "loss": 3.4471, + "step": 25900 + }, + { + "epoch": 7.556817519948745, + "grad_norm": 0.32970210909843445, + "learning_rate": 0.0005096188811188811, + "loss": 3.439, + "step": 25950 + }, + { + "epoch": 7.571378647562467, + "grad_norm": 0.33252274990081787, + "learning_rate": 0.0005094440559440559, + "loss": 3.455, + "step": 26000 + }, + { + "epoch": 7.571378647562467, + "eval_accuracy": 0.3646296705929088, + "eval_loss": 3.592409610748291, + "eval_runtime": 179.2443, + "eval_samples_per_second": 92.84, + "eval_steps_per_second": 5.808, + "step": 26000 + }, + { + "epoch": 7.58593977517619, + "grad_norm": 0.3485804498195648, + "learning_rate": 0.0005092692307692308, + "loss": 3.4633, + "step": 26050 + }, + { + "epoch": 7.600500902789912, + "grad_norm": 0.3297301232814789, + "learning_rate": 0.0005090944055944056, + "loss": 3.4468, + "step": 26100 + }, + { + "epoch": 7.615062030403634, + "grad_norm": 0.35925644636154175, + "learning_rate": 0.0005089195804195804, + "loss": 3.4504, + "step": 26150 + }, + { + "epoch": 7.629623158017357, + "grad_norm": 0.3311382234096527, + "learning_rate": 0.0005087447552447552, + "loss": 3.4616, + "step": 26200 + }, + { + "epoch": 7.644184285631079, + "grad_norm": 0.3223942220211029, + "learning_rate": 0.00050856993006993, + "loss": 3.4614, + "step": 26250 + }, + { + "epoch": 7.658745413244802, + "grad_norm": 0.36384764313697815, + "learning_rate": 0.0005083951048951048, + "loss": 3.4568, + "step": 26300 + }, + { + "epoch": 7.673306540858524, + "grad_norm": 0.3114417493343353, + "learning_rate": 0.0005082202797202797, + "loss": 3.4632, + "step": 26350 + }, + { + "epoch": 7.687867668472246, + "grad_norm": 0.3382061719894409, + "learning_rate": 0.0005080454545454545, + "loss": 3.4573, + "step": 26400 + }, + { + "epoch": 7.702428796085969, + "grad_norm": 0.3447924554347992, + "learning_rate": 0.0005078706293706293, + "loss": 3.4569, + "step": 26450 + }, + { + "epoch": 7.716989923699691, + "grad_norm": 0.31006723642349243, + "learning_rate": 0.0005076958041958042, + "loss": 3.4533, + "step": 26500 + }, + { + "epoch": 7.731551051313414, + "grad_norm": 0.3572918176651001, + "learning_rate": 0.000507520979020979, + "loss": 3.4495, + "step": 26550 + }, + { + "epoch": 7.746112178927136, + "grad_norm": 0.346057653427124, + "learning_rate": 0.0005073461538461538, + "loss": 3.4537, + "step": 26600 + }, + { + "epoch": 7.760673306540858, + "grad_norm": 0.3390602469444275, + "learning_rate": 0.0005071713286713286, + "loss": 3.4611, + "step": 26650 + }, + { + "epoch": 7.775234434154581, + "grad_norm": 0.33900779485702515, + "learning_rate": 0.0005069965034965035, + "loss": 3.4721, + "step": 26700 + }, + { + "epoch": 7.789795561768304, + "grad_norm": 0.3299110233783722, + "learning_rate": 0.0005068216783216783, + "loss": 3.4536, + "step": 26750 + }, + { + "epoch": 7.8043566893820255, + "grad_norm": 0.3273201882839203, + "learning_rate": 0.0005066468531468531, + "loss": 3.4491, + "step": 26800 + }, + { + "epoch": 7.818917816995748, + "grad_norm": 0.3403856158256531, + "learning_rate": 0.0005064720279720279, + "loss": 3.469, + "step": 26850 + }, + { + "epoch": 7.833478944609471, + "grad_norm": 0.3232671618461609, + "learning_rate": 0.0005062972027972028, + "loss": 3.4688, + "step": 26900 + }, + { + "epoch": 7.848040072223193, + "grad_norm": 0.33079832792282104, + "learning_rate": 0.0005061223776223775, + "loss": 3.4656, + "step": 26950 + }, + { + "epoch": 7.862601199836916, + "grad_norm": 0.34660980105400085, + "learning_rate": 0.0005059475524475524, + "loss": 3.4594, + "step": 27000 + }, + { + "epoch": 7.862601199836916, + "eval_accuracy": 0.3659460618275824, + "eval_loss": 3.583092451095581, + "eval_runtime": 179.2875, + "eval_samples_per_second": 92.817, + "eval_steps_per_second": 5.806, + "step": 27000 + }, + { + "epoch": 7.8771623274506375, + "grad_norm": 0.33932608366012573, + "learning_rate": 0.0005057727272727272, + "loss": 3.4598, + "step": 27050 + }, + { + "epoch": 7.89172345506436, + "grad_norm": 0.32174038887023926, + "learning_rate": 0.000505597902097902, + "loss": 3.4676, + "step": 27100 + }, + { + "epoch": 7.906284582678083, + "grad_norm": 0.319063276052475, + "learning_rate": 0.0005054230769230769, + "loss": 3.4729, + "step": 27150 + }, + { + "epoch": 7.920845710291805, + "grad_norm": 0.3107272684574127, + "learning_rate": 0.0005052482517482517, + "loss": 3.4608, + "step": 27200 + }, + { + "epoch": 7.935406837905528, + "grad_norm": 0.30821242928504944, + "learning_rate": 0.0005050734265734265, + "loss": 3.4744, + "step": 27250 + }, + { + "epoch": 7.9499679655192494, + "grad_norm": 0.35152456164360046, + "learning_rate": 0.0005048986013986013, + "loss": 3.4609, + "step": 27300 + }, + { + "epoch": 7.964529093132972, + "grad_norm": 0.35388267040252686, + "learning_rate": 0.0005047237762237762, + "loss": 3.4589, + "step": 27350 + }, + { + "epoch": 7.979090220746695, + "grad_norm": 0.3130147159099579, + "learning_rate": 0.000504548951048951, + "loss": 3.4608, + "step": 27400 + }, + { + "epoch": 7.993651348360417, + "grad_norm": 0.33657023310661316, + "learning_rate": 0.0005043741258741258, + "loss": 3.4614, + "step": 27450 + }, + { + "epoch": 8.008154231463685, + "grad_norm": 0.3313499391078949, + "learning_rate": 0.0005041993006993006, + "loss": 3.3894, + "step": 27500 + }, + { + "epoch": 8.022715359077408, + "grad_norm": 0.3206072449684143, + "learning_rate": 0.0005040244755244755, + "loss": 3.3532, + "step": 27550 + }, + { + "epoch": 8.037276486691129, + "grad_norm": 0.342159241437912, + "learning_rate": 0.0005038496503496503, + "loss": 3.3603, + "step": 27600 + }, + { + "epoch": 8.051837614304851, + "grad_norm": 0.3359133303165436, + "learning_rate": 0.0005036748251748251, + "loss": 3.3489, + "step": 27650 + }, + { + "epoch": 8.066398741918574, + "grad_norm": 0.3396371603012085, + "learning_rate": 0.0005034999999999999, + "loss": 3.3671, + "step": 27700 + }, + { + "epoch": 8.080959869532297, + "grad_norm": 0.34200868010520935, + "learning_rate": 0.0005033251748251747, + "loss": 3.3696, + "step": 27750 + }, + { + "epoch": 8.09552099714602, + "grad_norm": 0.3584854304790497, + "learning_rate": 0.0005031503496503496, + "loss": 3.3658, + "step": 27800 + }, + { + "epoch": 8.11008212475974, + "grad_norm": 0.33276915550231934, + "learning_rate": 0.0005029755244755244, + "loss": 3.3749, + "step": 27850 + }, + { + "epoch": 8.124643252373463, + "grad_norm": 0.3365272879600525, + "learning_rate": 0.0005028006993006992, + "loss": 3.3894, + "step": 27900 + }, + { + "epoch": 8.139204379987186, + "grad_norm": 0.3202180564403534, + "learning_rate": 0.000502625874125874, + "loss": 3.3796, + "step": 27950 + }, + { + "epoch": 8.153765507600909, + "grad_norm": 0.3322695791721344, + "learning_rate": 0.000502451048951049, + "loss": 3.3875, + "step": 28000 + }, + { + "epoch": 8.153765507600909, + "eval_accuracy": 0.3657153346878263, + "eval_loss": 3.5931077003479004, + "eval_runtime": 179.2021, + "eval_samples_per_second": 92.862, + "eval_steps_per_second": 5.809, + "step": 28000 + }, + { + "epoch": 8.168326635214632, + "grad_norm": 0.34390944242477417, + "learning_rate": 0.0005022762237762237, + "loss": 3.3888, + "step": 28050 + }, + { + "epoch": 8.182887762828354, + "grad_norm": 0.3357495665550232, + "learning_rate": 0.0005021013986013985, + "loss": 3.4004, + "step": 28100 + }, + { + "epoch": 8.197448890442075, + "grad_norm": 0.36402395367622375, + "learning_rate": 0.0005019265734265733, + "loss": 3.3936, + "step": 28150 + }, + { + "epoch": 8.212010018055798, + "grad_norm": 0.35054436326026917, + "learning_rate": 0.0005017517482517483, + "loss": 3.3989, + "step": 28200 + }, + { + "epoch": 8.22657114566952, + "grad_norm": 0.36492618918418884, + "learning_rate": 0.0005015769230769231, + "loss": 3.3924, + "step": 28250 + }, + { + "epoch": 8.241132273283243, + "grad_norm": 0.34601396322250366, + "learning_rate": 0.0005014020979020979, + "loss": 3.4028, + "step": 28300 + }, + { + "epoch": 8.255693400896966, + "grad_norm": 0.337088942527771, + "learning_rate": 0.0005012272727272727, + "loss": 3.3941, + "step": 28350 + }, + { + "epoch": 8.270254528510687, + "grad_norm": 0.3266392648220062, + "learning_rate": 0.0005010524475524476, + "loss": 3.3942, + "step": 28400 + }, + { + "epoch": 8.28481565612441, + "grad_norm": 0.3120972812175751, + "learning_rate": 0.0005008776223776223, + "loss": 3.3984, + "step": 28450 + }, + { + "epoch": 8.299376783738133, + "grad_norm": 0.342134952545166, + "learning_rate": 0.0005007027972027972, + "loss": 3.4024, + "step": 28500 + }, + { + "epoch": 8.313937911351855, + "grad_norm": 0.32515040040016174, + "learning_rate": 0.000500527972027972, + "loss": 3.4032, + "step": 28550 + }, + { + "epoch": 8.328499038965578, + "grad_norm": 0.3492473363876343, + "learning_rate": 0.0005003531468531468, + "loss": 3.4017, + "step": 28600 + }, + { + "epoch": 8.3430601665793, + "grad_norm": 0.3267183005809784, + "learning_rate": 0.0005001783216783217, + "loss": 3.3989, + "step": 28650 + }, + { + "epoch": 8.357621294193022, + "grad_norm": 0.35122978687286377, + "learning_rate": 0.0005000034965034965, + "loss": 3.4109, + "step": 28700 + }, + { + "epoch": 8.372182421806745, + "grad_norm": 0.32366734743118286, + "learning_rate": 0.0004998286713286713, + "loss": 3.4049, + "step": 28750 + }, + { + "epoch": 8.386743549420467, + "grad_norm": 0.3331584632396698, + "learning_rate": 0.0004996538461538461, + "loss": 3.4121, + "step": 28800 + }, + { + "epoch": 8.40130467703419, + "grad_norm": 0.34323281049728394, + "learning_rate": 0.000499479020979021, + "loss": 3.4004, + "step": 28850 + }, + { + "epoch": 8.415865804647911, + "grad_norm": 0.3543521761894226, + "learning_rate": 0.0004993041958041958, + "loss": 3.4122, + "step": 28900 + }, + { + "epoch": 8.430426932261634, + "grad_norm": 0.3661775290966034, + "learning_rate": 0.0004991293706293706, + "loss": 3.4084, + "step": 28950 + }, + { + "epoch": 8.444988059875357, + "grad_norm": 0.34312599897384644, + "learning_rate": 0.0004989545454545454, + "loss": 3.4211, + "step": 29000 + }, + { + "epoch": 8.444988059875357, + "eval_accuracy": 0.36638529009821896, + "eval_loss": 3.586408853530884, + "eval_runtime": 179.1019, + "eval_samples_per_second": 92.914, + "eval_steps_per_second": 5.812, + "step": 29000 + }, + { + "epoch": 8.45954918748908, + "grad_norm": 0.34694284200668335, + "learning_rate": 0.0004987797202797203, + "loss": 3.3969, + "step": 29050 + }, + { + "epoch": 8.474110315102802, + "grad_norm": 0.3572796881198883, + "learning_rate": 0.0004986048951048951, + "loss": 3.4092, + "step": 29100 + }, + { + "epoch": 8.488671442716523, + "grad_norm": 0.3393023610115051, + "learning_rate": 0.0004984300699300699, + "loss": 3.4146, + "step": 29150 + }, + { + "epoch": 8.503232570330246, + "grad_norm": 0.37056249380111694, + "learning_rate": 0.0004982552447552448, + "loss": 3.4238, + "step": 29200 + }, + { + "epoch": 8.517793697943969, + "grad_norm": 0.35526904463768005, + "learning_rate": 0.0004980804195804195, + "loss": 3.4251, + "step": 29250 + }, + { + "epoch": 8.532354825557691, + "grad_norm": 0.3318297863006592, + "learning_rate": 0.0004979055944055944, + "loss": 3.4226, + "step": 29300 + }, + { + "epoch": 8.546915953171414, + "grad_norm": 0.35030457377433777, + "learning_rate": 0.0004977307692307692, + "loss": 3.4224, + "step": 29350 + }, + { + "epoch": 8.561477080785137, + "grad_norm": 0.32418620586395264, + "learning_rate": 0.000497555944055944, + "loss": 3.4247, + "step": 29400 + }, + { + "epoch": 8.576038208398858, + "grad_norm": 0.3271182179450989, + "learning_rate": 0.0004973811188811188, + "loss": 3.4219, + "step": 29450 + }, + { + "epoch": 8.59059933601258, + "grad_norm": 0.3600592017173767, + "learning_rate": 0.0004972062937062937, + "loss": 3.4175, + "step": 29500 + }, + { + "epoch": 8.605160463626303, + "grad_norm": 0.35106921195983887, + "learning_rate": 0.0004970314685314685, + "loss": 3.4422, + "step": 29550 + }, + { + "epoch": 8.619721591240026, + "grad_norm": 0.34820082783699036, + "learning_rate": 0.0004968566433566433, + "loss": 3.4341, + "step": 29600 + }, + { + "epoch": 8.634282718853749, + "grad_norm": 0.34959983825683594, + "learning_rate": 0.0004966818181818181, + "loss": 3.43, + "step": 29650 + }, + { + "epoch": 8.64884384646747, + "grad_norm": 0.3466774523258209, + "learning_rate": 0.000496506993006993, + "loss": 3.4361, + "step": 29700 + }, + { + "epoch": 8.663404974081192, + "grad_norm": 0.3761068880558014, + "learning_rate": 0.0004963321678321678, + "loss": 3.4235, + "step": 29750 + }, + { + "epoch": 8.677966101694915, + "grad_norm": 0.3419499099254608, + "learning_rate": 0.0004961573426573426, + "loss": 3.4138, + "step": 29800 + }, + { + "epoch": 8.692527229308638, + "grad_norm": 0.3618893027305603, + "learning_rate": 0.0004959825174825175, + "loss": 3.4213, + "step": 29850 + }, + { + "epoch": 8.70708835692236, + "grad_norm": 0.3574729561805725, + "learning_rate": 0.0004958076923076923, + "loss": 3.4374, + "step": 29900 + }, + { + "epoch": 8.721649484536082, + "grad_norm": 0.3435271978378296, + "learning_rate": 0.0004956328671328671, + "loss": 3.4261, + "step": 29950 + }, + { + "epoch": 8.736210612149804, + "grad_norm": 0.33574849367141724, + "learning_rate": 0.0004954580419580419, + "loss": 3.4212, + "step": 30000 + }, + { + "epoch": 8.736210612149804, + "eval_accuracy": 0.3668438044294672, + "eval_loss": 3.5769741535186768, + "eval_runtime": 179.0988, + "eval_samples_per_second": 92.915, + "eval_steps_per_second": 5.812, + "step": 30000 + }, + { + "epoch": 8.750771739763527, + "grad_norm": 0.34257417917251587, + "learning_rate": 0.0004952832167832167, + "loss": 3.4339, + "step": 30050 + }, + { + "epoch": 8.76533286737725, + "grad_norm": 0.3249179422855377, + "learning_rate": 0.0004951083916083915, + "loss": 3.4359, + "step": 30100 + }, + { + "epoch": 8.779893994990973, + "grad_norm": 0.3594135642051697, + "learning_rate": 0.0004949335664335664, + "loss": 3.4437, + "step": 30150 + }, + { + "epoch": 8.794455122604695, + "grad_norm": 0.33656150102615356, + "learning_rate": 0.0004947587412587412, + "loss": 3.4324, + "step": 30200 + }, + { + "epoch": 8.809016250218416, + "grad_norm": 0.3542347848415375, + "learning_rate": 0.000494583916083916, + "loss": 3.4424, + "step": 30250 + }, + { + "epoch": 8.82357737783214, + "grad_norm": 0.30834895372390747, + "learning_rate": 0.0004944090909090908, + "loss": 3.4323, + "step": 30300 + }, + { + "epoch": 8.838138505445862, + "grad_norm": 0.36263927817344666, + "learning_rate": 0.0004942342657342657, + "loss": 3.426, + "step": 30350 + }, + { + "epoch": 8.852699633059585, + "grad_norm": 0.3364308774471283, + "learning_rate": 0.0004940594405594405, + "loss": 3.4404, + "step": 30400 + }, + { + "epoch": 8.867260760673307, + "grad_norm": 0.32633042335510254, + "learning_rate": 0.0004938846153846153, + "loss": 3.4464, + "step": 30450 + }, + { + "epoch": 8.881821888287028, + "grad_norm": 0.3413340449333191, + "learning_rate": 0.0004937097902097901, + "loss": 3.4286, + "step": 30500 + }, + { + "epoch": 8.896383015900751, + "grad_norm": 0.3368505537509918, + "learning_rate": 0.000493534965034965, + "loss": 3.4311, + "step": 30550 + }, + { + "epoch": 8.910944143514474, + "grad_norm": 0.34167155623435974, + "learning_rate": 0.0004933601398601398, + "loss": 3.4368, + "step": 30600 + }, + { + "epoch": 8.925505271128197, + "grad_norm": 0.35151201486587524, + "learning_rate": 0.0004931853146853146, + "loss": 3.4379, + "step": 30650 + }, + { + "epoch": 8.94006639874192, + "grad_norm": 0.35220181941986084, + "learning_rate": 0.0004930104895104895, + "loss": 3.4369, + "step": 30700 + }, + { + "epoch": 8.95462752635564, + "grad_norm": 0.33892005681991577, + "learning_rate": 0.0004928356643356642, + "loss": 3.4285, + "step": 30750 + }, + { + "epoch": 8.969188653969363, + "grad_norm": 0.33421051502227783, + "learning_rate": 0.0004926608391608391, + "loss": 3.4423, + "step": 30800 + }, + { + "epoch": 8.983749781583086, + "grad_norm": 0.3366270661354065, + "learning_rate": 0.0004924860139860139, + "loss": 3.4395, + "step": 30850 + }, + { + "epoch": 8.998310909196809, + "grad_norm": 0.32839125394821167, + "learning_rate": 0.0004923111888111887, + "loss": 3.4313, + "step": 30900 + }, + { + "epoch": 9.012813792300076, + "grad_norm": 0.33133724331855774, + "learning_rate": 0.0004921363636363635, + "loss": 3.3272, + "step": 30950 + }, + { + "epoch": 9.027374919913798, + "grad_norm": 0.3523022532463074, + "learning_rate": 0.0004919615384615384, + "loss": 3.321, + "step": 31000 + }, + { + "epoch": 9.027374919913798, + "eval_accuracy": 0.3670043256046797, + "eval_loss": 3.583221912384033, + "eval_runtime": 179.238, + "eval_samples_per_second": 92.843, + "eval_steps_per_second": 5.808, + "step": 31000 + }, + { + "epoch": 9.041936047527521, + "grad_norm": 0.3598332703113556, + "learning_rate": 0.0004917867132867132, + "loss": 3.3145, + "step": 31050 + }, + { + "epoch": 9.056497175141242, + "grad_norm": 0.35396015644073486, + "learning_rate": 0.000491611888111888, + "loss": 3.3272, + "step": 31100 + }, + { + "epoch": 9.071058302754965, + "grad_norm": 0.34974604845046997, + "learning_rate": 0.0004914370629370628, + "loss": 3.3405, + "step": 31150 + }, + { + "epoch": 9.085619430368688, + "grad_norm": 0.3377837836742401, + "learning_rate": 0.0004912622377622378, + "loss": 3.3413, + "step": 31200 + }, + { + "epoch": 9.10018055798241, + "grad_norm": 0.33599475026130676, + "learning_rate": 0.0004910874125874126, + "loss": 3.3398, + "step": 31250 + }, + { + "epoch": 9.114741685596133, + "grad_norm": 0.34810763597488403, + "learning_rate": 0.0004909125874125874, + "loss": 3.3503, + "step": 31300 + }, + { + "epoch": 9.129302813209854, + "grad_norm": 0.3428230583667755, + "learning_rate": 0.0004907377622377623, + "loss": 3.3445, + "step": 31350 + }, + { + "epoch": 9.143863940823577, + "grad_norm": 0.3467947542667389, + "learning_rate": 0.0004905629370629371, + "loss": 3.352, + "step": 31400 + }, + { + "epoch": 9.1584250684373, + "grad_norm": 0.33666664361953735, + "learning_rate": 0.0004903881118881119, + "loss": 3.3478, + "step": 31450 + }, + { + "epoch": 9.172986196051022, + "grad_norm": 0.3342663645744324, + "learning_rate": 0.0004902132867132867, + "loss": 3.3398, + "step": 31500 + }, + { + "epoch": 9.187547323664745, + "grad_norm": 0.3539953827857971, + "learning_rate": 0.0004900384615384615, + "loss": 3.368, + "step": 31550 + }, + { + "epoch": 9.202108451278466, + "grad_norm": 0.3389320373535156, + "learning_rate": 0.0004898636363636363, + "loss": 3.3578, + "step": 31600 + }, + { + "epoch": 9.216669578892189, + "grad_norm": 0.34655994176864624, + "learning_rate": 0.0004896888111888112, + "loss": 3.3618, + "step": 31650 + }, + { + "epoch": 9.231230706505912, + "grad_norm": 0.3649561405181885, + "learning_rate": 0.000489513986013986, + "loss": 3.382, + "step": 31700 + }, + { + "epoch": 9.245791834119634, + "grad_norm": 0.3527652323246002, + "learning_rate": 0.0004893391608391608, + "loss": 3.3908, + "step": 31750 + }, + { + "epoch": 9.260352961733357, + "grad_norm": 0.3279936909675598, + "learning_rate": 0.0004891643356643356, + "loss": 3.3818, + "step": 31800 + }, + { + "epoch": 9.27491408934708, + "grad_norm": 0.36550813913345337, + "learning_rate": 0.0004889895104895105, + "loss": 3.3697, + "step": 31850 + }, + { + "epoch": 9.2894752169608, + "grad_norm": 0.33970963954925537, + "learning_rate": 0.0004888146853146853, + "loss": 3.374, + "step": 31900 + }, + { + "epoch": 9.304036344574524, + "grad_norm": 0.3586331009864807, + "learning_rate": 0.0004886398601398601, + "loss": 3.3784, + "step": 31950 + }, + { + "epoch": 9.318597472188246, + "grad_norm": 0.37601354718208313, + "learning_rate": 0.000488465034965035, + "loss": 3.366, + "step": 32000 + }, + { + "epoch": 9.318597472188246, + "eval_accuracy": 0.3672680977629228, + "eval_loss": 3.581895589828491, + "eval_runtime": 179.132, + "eval_samples_per_second": 92.898, + "eval_steps_per_second": 5.811, + "step": 32000 + }, + { + "epoch": 9.333158599801969, + "grad_norm": 0.3502708673477173, + "learning_rate": 0.0004882902097902098, + "loss": 3.3771, + "step": 32050 + }, + { + "epoch": 9.347719727415692, + "grad_norm": 0.3333316743373871, + "learning_rate": 0.0004881153846153846, + "loss": 3.3919, + "step": 32100 + }, + { + "epoch": 9.362280855029413, + "grad_norm": 0.3387976884841919, + "learning_rate": 0.0004879405594405594, + "loss": 3.3867, + "step": 32150 + }, + { + "epoch": 9.376841982643136, + "grad_norm": 0.3522228002548218, + "learning_rate": 0.00048776573426573424, + "loss": 3.3816, + "step": 32200 + }, + { + "epoch": 9.391403110256858, + "grad_norm": 0.3376480042934418, + "learning_rate": 0.00048759090909090904, + "loss": 3.3864, + "step": 32250 + }, + { + "epoch": 9.405964237870581, + "grad_norm": 0.35908243060112, + "learning_rate": 0.0004874160839160839, + "loss": 3.3842, + "step": 32300 + }, + { + "epoch": 9.420525365484304, + "grad_norm": 0.3469899892807007, + "learning_rate": 0.0004872412587412587, + "loss": 3.3992, + "step": 32350 + }, + { + "epoch": 9.435086493098025, + "grad_norm": 0.3562922775745392, + "learning_rate": 0.00048706643356643354, + "loss": 3.3887, + "step": 32400 + }, + { + "epoch": 9.449647620711747, + "grad_norm": 0.3371868431568146, + "learning_rate": 0.00048689160839160834, + "loss": 3.3948, + "step": 32450 + }, + { + "epoch": 9.46420874832547, + "grad_norm": 0.33593714237213135, + "learning_rate": 0.0004867167832167832, + "loss": 3.386, + "step": 32500 + }, + { + "epoch": 9.478769875939193, + "grad_norm": 0.3544469475746155, + "learning_rate": 0.00048654195804195794, + "loss": 3.3953, + "step": 32550 + }, + { + "epoch": 9.493331003552916, + "grad_norm": 0.3525102138519287, + "learning_rate": 0.00048636713286713285, + "loss": 3.3902, + "step": 32600 + }, + { + "epoch": 9.507892131166638, + "grad_norm": 0.3648473024368286, + "learning_rate": 0.0004861923076923077, + "loss": 3.3969, + "step": 32650 + }, + { + "epoch": 9.52245325878036, + "grad_norm": 0.3423263132572174, + "learning_rate": 0.00048601748251748245, + "loss": 3.4029, + "step": 32700 + }, + { + "epoch": 9.537014386394082, + "grad_norm": 0.3576188087463379, + "learning_rate": 0.0004858426573426573, + "loss": 3.3979, + "step": 32750 + }, + { + "epoch": 9.551575514007805, + "grad_norm": 0.36224931478500366, + "learning_rate": 0.0004856678321678321, + "loss": 3.3971, + "step": 32800 + }, + { + "epoch": 9.566136641621528, + "grad_norm": 0.3539172112941742, + "learning_rate": 0.00048549300699300696, + "loss": 3.4046, + "step": 32850 + }, + { + "epoch": 9.58069776923525, + "grad_norm": 0.35503363609313965, + "learning_rate": 0.00048531818181818176, + "loss": 3.4084, + "step": 32900 + }, + { + "epoch": 9.595258896848971, + "grad_norm": 0.34804868698120117, + "learning_rate": 0.0004851433566433566, + "loss": 3.3934, + "step": 32950 + }, + { + "epoch": 9.609820024462694, + "grad_norm": 0.36221036314964294, + "learning_rate": 0.0004849685314685314, + "loss": 3.396, + "step": 33000 + }, + { + "epoch": 9.609820024462694, + "eval_accuracy": 0.36790406737138404, + "eval_loss": 3.570464611053467, + "eval_runtime": 179.3584, + "eval_samples_per_second": 92.781, + "eval_steps_per_second": 5.804, + "step": 33000 + }, + { + "epoch": 9.624381152076417, + "grad_norm": 0.33956465125083923, + "learning_rate": 0.00048479370629370627, + "loss": 3.4138, + "step": 33050 + }, + { + "epoch": 9.63894227969014, + "grad_norm": 0.3517622649669647, + "learning_rate": 0.00048461888111888106, + "loss": 3.4007, + "step": 33100 + }, + { + "epoch": 9.653503407303862, + "grad_norm": 0.3565088212490082, + "learning_rate": 0.0004844440559440559, + "loss": 3.3961, + "step": 33150 + }, + { + "epoch": 9.668064534917583, + "grad_norm": 0.3277449309825897, + "learning_rate": 0.0004842692307692307, + "loss": 3.4079, + "step": 33200 + }, + { + "epoch": 9.682625662531306, + "grad_norm": 0.3381253778934479, + "learning_rate": 0.00048409440559440557, + "loss": 3.4089, + "step": 33250 + }, + { + "epoch": 9.697186790145029, + "grad_norm": 0.3353286385536194, + "learning_rate": 0.0004839195804195803, + "loss": 3.3954, + "step": 33300 + }, + { + "epoch": 9.711747917758752, + "grad_norm": 0.3465847969055176, + "learning_rate": 0.0004837447552447552, + "loss": 3.4017, + "step": 33350 + }, + { + "epoch": 9.726309045372474, + "grad_norm": 0.35246995091438293, + "learning_rate": 0.0004835699300699301, + "loss": 3.4131, + "step": 33400 + }, + { + "epoch": 9.740870172986195, + "grad_norm": 0.3163084089756012, + "learning_rate": 0.0004833951048951048, + "loss": 3.4136, + "step": 33450 + }, + { + "epoch": 9.755431300599918, + "grad_norm": 0.34027642011642456, + "learning_rate": 0.0004832202797202797, + "loss": 3.4073, + "step": 33500 + }, + { + "epoch": 9.76999242821364, + "grad_norm": 0.33346980810165405, + "learning_rate": 0.0004830454545454545, + "loss": 3.3888, + "step": 33550 + }, + { + "epoch": 9.784553555827364, + "grad_norm": 0.3280010521411896, + "learning_rate": 0.00048287062937062933, + "loss": 3.4202, + "step": 33600 + }, + { + "epoch": 9.799114683441086, + "grad_norm": 0.3580598533153534, + "learning_rate": 0.00048269580419580413, + "loss": 3.4014, + "step": 33650 + }, + { + "epoch": 9.813675811054807, + "grad_norm": 0.3334909677505493, + "learning_rate": 0.000482520979020979, + "loss": 3.4149, + "step": 33700 + }, + { + "epoch": 9.82823693866853, + "grad_norm": 0.34729447960853577, + "learning_rate": 0.0004823461538461538, + "loss": 3.4148, + "step": 33750 + }, + { + "epoch": 9.842798066282253, + "grad_norm": 0.3324950635433197, + "learning_rate": 0.00048217132867132864, + "loss": 3.4187, + "step": 33800 + }, + { + "epoch": 9.857359193895975, + "grad_norm": 0.3630949854850769, + "learning_rate": 0.00048199650349650344, + "loss": 3.4116, + "step": 33850 + }, + { + "epoch": 9.871920321509698, + "grad_norm": 0.3364851772785187, + "learning_rate": 0.0004818216783216783, + "loss": 3.3993, + "step": 33900 + }, + { + "epoch": 9.88648144912342, + "grad_norm": 0.32940131425857544, + "learning_rate": 0.0004816468531468531, + "loss": 3.4182, + "step": 33950 + }, + { + "epoch": 9.901042576737142, + "grad_norm": 0.34826645255088806, + "learning_rate": 0.00048147202797202795, + "loss": 3.4047, + "step": 34000 + }, + { + "epoch": 9.901042576737142, + "eval_accuracy": 0.3682340471645316, + "eval_loss": 3.562272310256958, + "eval_runtime": 183.8917, + "eval_samples_per_second": 90.493, + "eval_steps_per_second": 5.661, + "step": 34000 + }, + { + "epoch": 9.915603704350865, + "grad_norm": 0.3615463972091675, + "learning_rate": 0.0004812972027972028, + "loss": 3.4091, + "step": 34050 + }, + { + "epoch": 9.930164831964587, + "grad_norm": 0.3819344639778137, + "learning_rate": 0.0004811223776223776, + "loss": 3.4215, + "step": 34100 + }, + { + "epoch": 9.94472595957831, + "grad_norm": 0.3556666970252991, + "learning_rate": 0.00048094755244755245, + "loss": 3.4186, + "step": 34150 + }, + { + "epoch": 9.959287087192033, + "grad_norm": 0.34192410111427307, + "learning_rate": 0.0004807727272727272, + "loss": 3.4134, + "step": 34200 + }, + { + "epoch": 9.973848214805754, + "grad_norm": 0.3439369201660156, + "learning_rate": 0.00048059790209790205, + "loss": 3.4223, + "step": 34250 + }, + { + "epoch": 9.988409342419477, + "grad_norm": 0.3427305221557617, + "learning_rate": 0.00048042307692307685, + "loss": 3.4118, + "step": 34300 + }, + { + "epoch": 10.002912225522744, + "grad_norm": 0.3675305247306824, + "learning_rate": 0.0004802482517482517, + "loss": 3.3852, + "step": 34350 + }, + { + "epoch": 10.017473353136467, + "grad_norm": 0.34155988693237305, + "learning_rate": 0.0004800734265734265, + "loss": 3.299, + "step": 34400 + }, + { + "epoch": 10.03203448075019, + "grad_norm": 0.3484591245651245, + "learning_rate": 0.00047989860139860136, + "loss": 3.3089, + "step": 34450 + }, + { + "epoch": 10.046595608363912, + "grad_norm": 0.34534206986427307, + "learning_rate": 0.00047972377622377616, + "loss": 3.2998, + "step": 34500 + }, + { + "epoch": 10.061156735977635, + "grad_norm": 0.36416661739349365, + "learning_rate": 0.000479548951048951, + "loss": 3.3052, + "step": 34550 + }, + { + "epoch": 10.075717863591356, + "grad_norm": 0.3576606512069702, + "learning_rate": 0.0004793741258741258, + "loss": 3.3274, + "step": 34600 + }, + { + "epoch": 10.090278991205079, + "grad_norm": 0.3462233245372772, + "learning_rate": 0.00047919930069930067, + "loss": 3.3128, + "step": 34650 + }, + { + "epoch": 10.104840118818801, + "grad_norm": 0.3831905126571655, + "learning_rate": 0.0004790244755244755, + "loss": 3.3314, + "step": 34700 + }, + { + "epoch": 10.119401246432524, + "grad_norm": 0.344808965921402, + "learning_rate": 0.0004788496503496503, + "loss": 3.3312, + "step": 34750 + }, + { + "epoch": 10.133962374046247, + "grad_norm": 0.39229050278663635, + "learning_rate": 0.0004786748251748252, + "loss": 3.323, + "step": 34800 + }, + { + "epoch": 10.148523501659968, + "grad_norm": 0.38846924901008606, + "learning_rate": 0.0004785, + "loss": 3.3259, + "step": 34850 + }, + { + "epoch": 10.16308462927369, + "grad_norm": 0.3591175079345703, + "learning_rate": 0.00047832517482517483, + "loss": 3.339, + "step": 34900 + }, + { + "epoch": 10.177645756887413, + "grad_norm": 0.3358122706413269, + "learning_rate": 0.0004781503496503496, + "loss": 3.3394, + "step": 34950 + }, + { + "epoch": 10.192206884501136, + "grad_norm": 0.35208988189697266, + "learning_rate": 0.00047797552447552443, + "loss": 3.338, + "step": 35000 + }, + { + "epoch": 10.192206884501136, + "eval_accuracy": 0.3676259482656128, + "eval_loss": 3.5794930458068848, + "eval_runtime": 179.1106, + "eval_samples_per_second": 92.909, + "eval_steps_per_second": 5.812, + "step": 35000 + }, + { + "epoch": 10.206768012114859, + "grad_norm": 0.34830009937286377, + "learning_rate": 0.00047780069930069923, + "loss": 3.3351, + "step": 35050 + }, + { + "epoch": 10.221329139728581, + "grad_norm": 0.37505292892456055, + "learning_rate": 0.0004776258741258741, + "loss": 3.3342, + "step": 35100 + }, + { + "epoch": 10.235890267342302, + "grad_norm": 0.3631197512149811, + "learning_rate": 0.0004774510489510489, + "loss": 3.346, + "step": 35150 + }, + { + "epoch": 10.250451394956025, + "grad_norm": 0.3792741596698761, + "learning_rate": 0.00047727622377622374, + "loss": 3.3511, + "step": 35200 + }, + { + "epoch": 10.265012522569748, + "grad_norm": 0.3474135994911194, + "learning_rate": 0.00047710139860139854, + "loss": 3.3472, + "step": 35250 + }, + { + "epoch": 10.27957365018347, + "grad_norm": 0.35298803448677063, + "learning_rate": 0.0004769265734265734, + "loss": 3.3558, + "step": 35300 + }, + { + "epoch": 10.294134777797193, + "grad_norm": 0.35359811782836914, + "learning_rate": 0.0004767517482517482, + "loss": 3.3481, + "step": 35350 + }, + { + "epoch": 10.308695905410914, + "grad_norm": 0.33651646971702576, + "learning_rate": 0.00047657692307692304, + "loss": 3.3529, + "step": 35400 + }, + { + "epoch": 10.323257033024637, + "grad_norm": 0.34708303213119507, + "learning_rate": 0.0004764020979020979, + "loss": 3.3632, + "step": 35450 + }, + { + "epoch": 10.33781816063836, + "grad_norm": 0.32987862825393677, + "learning_rate": 0.0004762272727272727, + "loss": 3.3639, + "step": 35500 + }, + { + "epoch": 10.352379288252083, + "grad_norm": 0.359037309885025, + "learning_rate": 0.00047605244755244755, + "loss": 3.3647, + "step": 35550 + }, + { + "epoch": 10.366940415865805, + "grad_norm": 0.3546748757362366, + "learning_rate": 0.00047587762237762235, + "loss": 3.3731, + "step": 35600 + }, + { + "epoch": 10.381501543479526, + "grad_norm": 0.3893897831439972, + "learning_rate": 0.0004757027972027972, + "loss": 3.3714, + "step": 35650 + }, + { + "epoch": 10.396062671093249, + "grad_norm": 0.3714803457260132, + "learning_rate": 0.00047552797202797195, + "loss": 3.3465, + "step": 35700 + }, + { + "epoch": 10.410623798706972, + "grad_norm": 0.3687228560447693, + "learning_rate": 0.0004753531468531468, + "loss": 3.3651, + "step": 35750 + }, + { + "epoch": 10.425184926320695, + "grad_norm": 0.37224873900413513, + "learning_rate": 0.0004751783216783216, + "loss": 3.3577, + "step": 35800 + }, + { + "epoch": 10.439746053934417, + "grad_norm": 0.3495529294013977, + "learning_rate": 0.00047500349650349646, + "loss": 3.3744, + "step": 35850 + }, + { + "epoch": 10.454307181548138, + "grad_norm": 0.34006187319755554, + "learning_rate": 0.00047482867132867126, + "loss": 3.3791, + "step": 35900 + }, + { + "epoch": 10.468868309161861, + "grad_norm": 0.34275439381599426, + "learning_rate": 0.0004746538461538461, + "loss": 3.3593, + "step": 35950 + }, + { + "epoch": 10.483429436775584, + "grad_norm": 0.34504151344299316, + "learning_rate": 0.0004744790209790209, + "loss": 3.3659, + "step": 36000 + }, + { + "epoch": 10.483429436775584, + "eval_accuracy": 0.3682414558341568, + "eval_loss": 3.570232629776001, + "eval_runtime": 179.0044, + "eval_samples_per_second": 92.964, + "eval_steps_per_second": 5.816, + "step": 36000 + }, + { + "epoch": 10.497990564389307, + "grad_norm": 0.3609980046749115, + "learning_rate": 0.00047430419580419576, + "loss": 3.3822, + "step": 36050 + }, + { + "epoch": 10.51255169200303, + "grad_norm": 0.36238977313041687, + "learning_rate": 0.0004741293706293706, + "loss": 3.3634, + "step": 36100 + }, + { + "epoch": 10.52711281961675, + "grad_norm": 0.3405495584011078, + "learning_rate": 0.0004739545454545454, + "loss": 3.3682, + "step": 36150 + }, + { + "epoch": 10.541673947230473, + "grad_norm": 0.33177825808525085, + "learning_rate": 0.00047377972027972027, + "loss": 3.3693, + "step": 36200 + }, + { + "epoch": 10.556235074844196, + "grad_norm": 0.3571736812591553, + "learning_rate": 0.00047360489510489507, + "loss": 3.3694, + "step": 36250 + }, + { + "epoch": 10.570796202457919, + "grad_norm": 0.3262953758239746, + "learning_rate": 0.0004734300699300699, + "loss": 3.3744, + "step": 36300 + }, + { + "epoch": 10.585357330071641, + "grad_norm": 0.37936583161354065, + "learning_rate": 0.0004732552447552447, + "loss": 3.3759, + "step": 36350 + }, + { + "epoch": 10.599918457685362, + "grad_norm": 0.33367085456848145, + "learning_rate": 0.0004730804195804196, + "loss": 3.3774, + "step": 36400 + }, + { + "epoch": 10.614479585299085, + "grad_norm": 0.35648494958877563, + "learning_rate": 0.0004729055944055943, + "loss": 3.3808, + "step": 36450 + }, + { + "epoch": 10.629040712912808, + "grad_norm": 0.36746612191200256, + "learning_rate": 0.0004727307692307692, + "loss": 3.3872, + "step": 36500 + }, + { + "epoch": 10.64360184052653, + "grad_norm": 0.39090093970298767, + "learning_rate": 0.000472555944055944, + "loss": 3.3782, + "step": 36550 + }, + { + "epoch": 10.658162968140253, + "grad_norm": 0.35592901706695557, + "learning_rate": 0.00047238111888111883, + "loss": 3.3916, + "step": 36600 + }, + { + "epoch": 10.672724095753976, + "grad_norm": 0.34853339195251465, + "learning_rate": 0.00047220629370629363, + "loss": 3.3814, + "step": 36650 + }, + { + "epoch": 10.687285223367697, + "grad_norm": 0.33484116196632385, + "learning_rate": 0.0004720314685314685, + "loss": 3.3762, + "step": 36700 + }, + { + "epoch": 10.70184635098142, + "grad_norm": 0.3227470815181732, + "learning_rate": 0.0004718566433566433, + "loss": 3.3825, + "step": 36750 + }, + { + "epoch": 10.716407478595142, + "grad_norm": 0.3507768511772156, + "learning_rate": 0.00047168181818181814, + "loss": 3.3757, + "step": 36800 + }, + { + "epoch": 10.730968606208865, + "grad_norm": 0.32565152645111084, + "learning_rate": 0.000471506993006993, + "loss": 3.3774, + "step": 36850 + }, + { + "epoch": 10.745529733822588, + "grad_norm": 0.35553401708602905, + "learning_rate": 0.0004713321678321678, + "loss": 3.3875, + "step": 36900 + }, + { + "epoch": 10.760090861436309, + "grad_norm": 0.36727797985076904, + "learning_rate": 0.00047115734265734265, + "loss": 3.3742, + "step": 36950 + }, + { + "epoch": 10.774651989050032, + "grad_norm": 0.3588033616542816, + "learning_rate": 0.00047098251748251745, + "loss": 3.3898, + "step": 37000 + }, + { + "epoch": 10.774651989050032, + "eval_accuracy": 0.36879945801465763, + "eval_loss": 3.562309503555298, + "eval_runtime": 179.1101, + "eval_samples_per_second": 92.909, + "eval_steps_per_second": 5.812, + "step": 37000 + }, + { + "epoch": 10.789213116663754, + "grad_norm": 0.35610097646713257, + "learning_rate": 0.0004708076923076923, + "loss": 3.3857, + "step": 37050 + }, + { + "epoch": 10.803774244277477, + "grad_norm": 0.3569492995738983, + "learning_rate": 0.0004706328671328671, + "loss": 3.3838, + "step": 37100 + }, + { + "epoch": 10.8183353718912, + "grad_norm": 0.33020931482315063, + "learning_rate": 0.00047045804195804195, + "loss": 3.3869, + "step": 37150 + }, + { + "epoch": 10.83289649950492, + "grad_norm": 0.33165302872657776, + "learning_rate": 0.0004702832167832167, + "loss": 3.3891, + "step": 37200 + }, + { + "epoch": 10.847457627118644, + "grad_norm": 0.34875956177711487, + "learning_rate": 0.00047010839160839155, + "loss": 3.3986, + "step": 37250 + }, + { + "epoch": 10.862018754732366, + "grad_norm": 0.34301435947418213, + "learning_rate": 0.00046993356643356635, + "loss": 3.3834, + "step": 37300 + }, + { + "epoch": 10.876579882346089, + "grad_norm": 0.3363425135612488, + "learning_rate": 0.0004697587412587412, + "loss": 3.3956, + "step": 37350 + }, + { + "epoch": 10.891141009959812, + "grad_norm": 0.3664510250091553, + "learning_rate": 0.000469583916083916, + "loss": 3.3815, + "step": 37400 + }, + { + "epoch": 10.905702137573535, + "grad_norm": 0.3228503167629242, + "learning_rate": 0.00046940909090909086, + "loss": 3.3895, + "step": 37450 + }, + { + "epoch": 10.920263265187256, + "grad_norm": 0.3490771949291229, + "learning_rate": 0.0004692342657342657, + "loss": 3.3977, + "step": 37500 + }, + { + "epoch": 10.934824392800978, + "grad_norm": 0.33811721205711365, + "learning_rate": 0.0004690594405594405, + "loss": 3.3961, + "step": 37550 + }, + { + "epoch": 10.949385520414701, + "grad_norm": 0.3598197102546692, + "learning_rate": 0.00046888461538461537, + "loss": 3.3979, + "step": 37600 + }, + { + "epoch": 10.963946648028424, + "grad_norm": 0.34854692220687866, + "learning_rate": 0.00046870979020979017, + "loss": 3.3942, + "step": 37650 + }, + { + "epoch": 10.978507775642147, + "grad_norm": 0.3387480080127716, + "learning_rate": 0.000468534965034965, + "loss": 3.4129, + "step": 37700 + }, + { + "epoch": 10.993068903255867, + "grad_norm": 0.3624948561191559, + "learning_rate": 0.0004683601398601398, + "loss": 3.3932, + "step": 37750 + }, + { + "epoch": 11.007571786359136, + "grad_norm": 0.34303465485572815, + "learning_rate": 0.0004681853146853147, + "loss": 3.3435, + "step": 37800 + }, + { + "epoch": 11.022132913972857, + "grad_norm": 0.3491910696029663, + "learning_rate": 0.0004680104895104895, + "loss": 3.2876, + "step": 37850 + }, + { + "epoch": 11.03669404158658, + "grad_norm": 0.35964518785476685, + "learning_rate": 0.00046783566433566433, + "loss": 3.2751, + "step": 37900 + }, + { + "epoch": 11.051255169200303, + "grad_norm": 0.369924932718277, + "learning_rate": 0.0004676608391608391, + "loss": 3.2885, + "step": 37950 + }, + { + "epoch": 11.065816296814026, + "grad_norm": 0.39939865469932556, + "learning_rate": 0.00046748601398601393, + "loss": 3.2971, + "step": 38000 + }, + { + "epoch": 11.065816296814026, + "eval_accuracy": 0.3682685033581853, + "eval_loss": 3.574512481689453, + "eval_runtime": 179.1624, + "eval_samples_per_second": 92.882, + "eval_steps_per_second": 5.81, + "step": 38000 + }, + { + "epoch": 11.080377424427748, + "grad_norm": 0.3350479006767273, + "learning_rate": 0.00046731118881118873, + "loss": 3.298, + "step": 38050 + }, + { + "epoch": 11.09493855204147, + "grad_norm": 0.3533541262149811, + "learning_rate": 0.0004671363636363636, + "loss": 3.2985, + "step": 38100 + }, + { + "epoch": 11.109499679655192, + "grad_norm": 0.3974376618862152, + "learning_rate": 0.00046696153846153844, + "loss": 3.3088, + "step": 38150 + }, + { + "epoch": 11.124060807268915, + "grad_norm": 0.347258597612381, + "learning_rate": 0.00046678671328671324, + "loss": 3.3106, + "step": 38200 + }, + { + "epoch": 11.138621934882638, + "grad_norm": 0.3746992349624634, + "learning_rate": 0.0004666118881118881, + "loss": 3.3245, + "step": 38250 + }, + { + "epoch": 11.15318306249636, + "grad_norm": 0.36625468730926514, + "learning_rate": 0.0004664370629370629, + "loss": 3.2991, + "step": 38300 + }, + { + "epoch": 11.167744190110081, + "grad_norm": 0.33680254220962524, + "learning_rate": 0.00046626223776223774, + "loss": 3.327, + "step": 38350 + }, + { + "epoch": 11.182305317723804, + "grad_norm": 0.37545084953308105, + "learning_rate": 0.00046608741258741254, + "loss": 3.3212, + "step": 38400 + }, + { + "epoch": 11.196866445337527, + "grad_norm": 0.3423392176628113, + "learning_rate": 0.0004659125874125874, + "loss": 3.3324, + "step": 38450 + }, + { + "epoch": 11.21142757295125, + "grad_norm": 0.3637043833732605, + "learning_rate": 0.0004657377622377622, + "loss": 3.3325, + "step": 38500 + }, + { + "epoch": 11.225988700564972, + "grad_norm": 0.35049334168434143, + "learning_rate": 0.00046556293706293705, + "loss": 3.3216, + "step": 38550 + }, + { + "epoch": 11.240549828178693, + "grad_norm": 0.33961722254753113, + "learning_rate": 0.00046538811188811185, + "loss": 3.3102, + "step": 38600 + }, + { + "epoch": 11.255110955792416, + "grad_norm": 0.35656216740608215, + "learning_rate": 0.0004652132867132867, + "loss": 3.3324, + "step": 38650 + }, + { + "epoch": 11.269672083406139, + "grad_norm": 0.34844666719436646, + "learning_rate": 0.00046503846153846145, + "loss": 3.3222, + "step": 38700 + }, + { + "epoch": 11.284233211019862, + "grad_norm": 0.3558867871761322, + "learning_rate": 0.0004648636363636363, + "loss": 3.3319, + "step": 38750 + }, + { + "epoch": 11.298794338633584, + "grad_norm": 0.3627782166004181, + "learning_rate": 0.0004646888111888111, + "loss": 3.349, + "step": 38800 + }, + { + "epoch": 11.313355466247307, + "grad_norm": 0.37425482273101807, + "learning_rate": 0.00046451398601398596, + "loss": 3.3335, + "step": 38850 + }, + { + "epoch": 11.327916593861028, + "grad_norm": 0.3499017059803009, + "learning_rate": 0.0004643391608391608, + "loss": 3.3334, + "step": 38900 + }, + { + "epoch": 11.34247772147475, + "grad_norm": 0.34610697627067566, + "learning_rate": 0.0004641643356643356, + "loss": 3.3416, + "step": 38950 + }, + { + "epoch": 11.357038849088473, + "grad_norm": 0.3500326871871948, + "learning_rate": 0.00046398951048951046, + "loss": 3.3319, + "step": 39000 + }, + { + "epoch": 11.357038849088473, + "eval_accuracy": 0.3688927131735907, + "eval_loss": 3.5684919357299805, + "eval_runtime": 179.217, + "eval_samples_per_second": 92.854, + "eval_steps_per_second": 5.809, + "step": 39000 + }, + { + "epoch": 11.371599976702196, + "grad_norm": 0.36741727590560913, + "learning_rate": 0.00046381468531468526, + "loss": 3.3346, + "step": 39050 + }, + { + "epoch": 11.386161104315919, + "grad_norm": 0.32896384596824646, + "learning_rate": 0.0004636398601398601, + "loss": 3.3451, + "step": 39100 + }, + { + "epoch": 11.40072223192964, + "grad_norm": 0.3439570665359497, + "learning_rate": 0.0004634650349650349, + "loss": 3.3411, + "step": 39150 + }, + { + "epoch": 11.415283359543363, + "grad_norm": 0.35606417059898376, + "learning_rate": 0.00046329020979020977, + "loss": 3.339, + "step": 39200 + }, + { + "epoch": 11.429844487157085, + "grad_norm": 0.3697795569896698, + "learning_rate": 0.00046311538461538457, + "loss": 3.3365, + "step": 39250 + }, + { + "epoch": 11.444405614770808, + "grad_norm": 0.3536551594734192, + "learning_rate": 0.0004629405594405594, + "loss": 3.3459, + "step": 39300 + }, + { + "epoch": 11.458966742384531, + "grad_norm": 0.3512251079082489, + "learning_rate": 0.0004627657342657342, + "loss": 3.3642, + "step": 39350 + }, + { + "epoch": 11.473527869998252, + "grad_norm": 0.3524966835975647, + "learning_rate": 0.0004625909090909091, + "loss": 3.3449, + "step": 39400 + }, + { + "epoch": 11.488088997611975, + "grad_norm": 0.36877357959747314, + "learning_rate": 0.0004624160839160838, + "loss": 3.3546, + "step": 39450 + }, + { + "epoch": 11.502650125225697, + "grad_norm": 0.3593060374259949, + "learning_rate": 0.0004622412587412587, + "loss": 3.3467, + "step": 39500 + }, + { + "epoch": 11.51721125283942, + "grad_norm": 0.37755638360977173, + "learning_rate": 0.00046206643356643353, + "loss": 3.355, + "step": 39550 + }, + { + "epoch": 11.531772380453143, + "grad_norm": 0.3538103401660919, + "learning_rate": 0.00046189160839160833, + "loss": 3.3575, + "step": 39600 + }, + { + "epoch": 11.546333508066864, + "grad_norm": 0.37226471304893494, + "learning_rate": 0.0004617167832167832, + "loss": 3.3392, + "step": 39650 + }, + { + "epoch": 11.560894635680587, + "grad_norm": 0.3464292287826538, + "learning_rate": 0.000461541958041958, + "loss": 3.3524, + "step": 39700 + }, + { + "epoch": 11.57545576329431, + "grad_norm": 0.356367826461792, + "learning_rate": 0.00046136713286713284, + "loss": 3.3629, + "step": 39750 + }, + { + "epoch": 11.590016890908032, + "grad_norm": 0.3554668724536896, + "learning_rate": 0.00046119230769230764, + "loss": 3.3625, + "step": 39800 + }, + { + "epoch": 11.604578018521755, + "grad_norm": 0.36473312973976135, + "learning_rate": 0.0004610174825174825, + "loss": 3.3541, + "step": 39850 + }, + { + "epoch": 11.619139146135478, + "grad_norm": 0.36857956647872925, + "learning_rate": 0.0004608426573426573, + "loss": 3.3553, + "step": 39900 + }, + { + "epoch": 11.633700273749199, + "grad_norm": 0.3510534167289734, + "learning_rate": 0.00046066783216783215, + "loss": 3.3699, + "step": 39950 + }, + { + "epoch": 11.648261401362921, + "grad_norm": 0.34622034430503845, + "learning_rate": 0.00046049300699300695, + "loss": 3.3606, + "step": 40000 + }, + { + "epoch": 11.648261401362921, + "eval_accuracy": 0.36932476797046315, + "eval_loss": 3.5598602294921875, + "eval_runtime": 179.1262, + "eval_samples_per_second": 92.901, + "eval_steps_per_second": 5.812, + "step": 40000 + } + ], + "logging_steps": 50, + "max_steps": 171700, + "num_input_tokens_seen": 0, + "num_train_epochs": 50, + "save_steps": 10000, + "stateful_callbacks": { + "EarlyStoppingCallback": { + "args": { + "early_stopping_patience": 20, + "early_stopping_threshold": 0.0 + }, + "attributes": { + "early_stopping_patience_counter": 0 + } + }, + "TrainerControl": { + "args": { + "should_epoch_stop": false, + "should_evaluate": false, + "should_log": false, + "should_save": true, + "should_training_stop": false + }, + "attributes": {} + } + }, + "total_flos": 8.36056898666496e+17, + "train_batch_size": 16, + "trial_name": null, + "trial_params": null +}