| { |
| "best_global_step": null, |
| "best_metric": null, |
| "best_model_checkpoint": null, |
| "epoch": 0.7561436672967864, |
| "eval_steps": 500, |
| "global_step": 1400, |
| "is_hyper_param_search": false, |
| "is_local_process_zero": true, |
| "is_world_process_zero": true, |
| "log_history": [ |
| { |
| "epoch": 0.0054010261949770455, |
| "grad_norm": 1.4275089502334595, |
| "learning_rate": 0.00016, |
| "loss": 3.4877567291259766, |
| "step": 10 |
| }, |
| { |
| "epoch": 0.010802052389954091, |
| "grad_norm": 1.4062877893447876, |
| "learning_rate": 0.00019971150378651282, |
| "loss": 3.177476501464844, |
| "step": 20 |
| }, |
| { |
| "epoch": 0.016203078584931137, |
| "grad_norm": 1.4894630908966064, |
| "learning_rate": 0.0001993508835196538, |
| "loss": 2.949508857727051, |
| "step": 30 |
| }, |
| { |
| "epoch": 0.021604104779908182, |
| "grad_norm": 1.1637152433395386, |
| "learning_rate": 0.00019899026325279482, |
| "loss": 3.2568588256835938, |
| "step": 40 |
| }, |
| { |
| "epoch": 0.027005130974885227, |
| "grad_norm": 1.0710796117782593, |
| "learning_rate": 0.00019862964298593581, |
| "loss": 3.033162307739258, |
| "step": 50 |
| }, |
| { |
| "epoch": 0.032406157169862275, |
| "grad_norm": 1.036720633506775, |
| "learning_rate": 0.0001982690227190768, |
| "loss": 2.9209388732910155, |
| "step": 60 |
| }, |
| { |
| "epoch": 0.03780718336483932, |
| "grad_norm": 1.5094584226608276, |
| "learning_rate": 0.00019790840245221782, |
| "loss": 3.199975776672363, |
| "step": 70 |
| }, |
| { |
| "epoch": 0.043208209559816364, |
| "grad_norm": 1.25065016746521, |
| "learning_rate": 0.00019754778218535884, |
| "loss": 3.1586023330688477, |
| "step": 80 |
| }, |
| { |
| "epoch": 0.04860923575479341, |
| "grad_norm": 1.093263864517212, |
| "learning_rate": 0.00019718716191849983, |
| "loss": 3.1673559188842773, |
| "step": 90 |
| }, |
| { |
| "epoch": 0.054010261949770454, |
| "grad_norm": 1.7025214433670044, |
| "learning_rate": 0.00019682654165164082, |
| "loss": 3.0090349197387694, |
| "step": 100 |
| }, |
| { |
| "epoch": 0.059411288144747505, |
| "grad_norm": 1.0379327535629272, |
| "learning_rate": 0.00019646592138478184, |
| "loss": 3.0862411499023437, |
| "step": 110 |
| }, |
| { |
| "epoch": 0.06481231433972455, |
| "grad_norm": 1.0767595767974854, |
| "learning_rate": 0.00019610530111792285, |
| "loss": 2.9345855712890625, |
| "step": 120 |
| }, |
| { |
| "epoch": 0.07021334053470159, |
| "grad_norm": 1.2473753690719604, |
| "learning_rate": 0.00019574468085106384, |
| "loss": 3.168864440917969, |
| "step": 130 |
| }, |
| { |
| "epoch": 0.07561436672967864, |
| "grad_norm": 1.1580232381820679, |
| "learning_rate": 0.00019538406058420483, |
| "loss": 3.0444286346435545, |
| "step": 140 |
| }, |
| { |
| "epoch": 0.08101539292465569, |
| "grad_norm": 1.1562517881393433, |
| "learning_rate": 0.00019502344031734585, |
| "loss": 3.0681676864624023, |
| "step": 150 |
| }, |
| { |
| "epoch": 0.08641641911963273, |
| "grad_norm": 1.2069370746612549, |
| "learning_rate": 0.00019466282005048684, |
| "loss": 3.147779846191406, |
| "step": 160 |
| }, |
| { |
| "epoch": 0.09181744531460978, |
| "grad_norm": 1.1520271301269531, |
| "learning_rate": 0.00019430219978362786, |
| "loss": 3.2159664154052736, |
| "step": 170 |
| }, |
| { |
| "epoch": 0.09721847150958682, |
| "grad_norm": 1.3290963172912598, |
| "learning_rate": 0.00019394157951676885, |
| "loss": 3.17901668548584, |
| "step": 180 |
| }, |
| { |
| "epoch": 0.10261949770456387, |
| "grad_norm": 1.4537458419799805, |
| "learning_rate": 0.00019358095924990984, |
| "loss": 3.076709747314453, |
| "step": 190 |
| }, |
| { |
| "epoch": 0.10802052389954091, |
| "grad_norm": 1.6899892091751099, |
| "learning_rate": 0.00019322033898305085, |
| "loss": 2.9362957000732424, |
| "step": 200 |
| }, |
| { |
| "epoch": 0.11342155009451796, |
| "grad_norm": 1.218902349472046, |
| "learning_rate": 0.00019289578074287774, |
| "loss": 3.0862009048461916, |
| "step": 210 |
| }, |
| { |
| "epoch": 0.11882257628949501, |
| "grad_norm": 1.253646969795227, |
| "learning_rate": 0.00019253516047601876, |
| "loss": 3.0416831970214844, |
| "step": 220 |
| }, |
| { |
| "epoch": 0.12422360248447205, |
| "grad_norm": 1.1816236972808838, |
| "learning_rate": 0.00019217454020915977, |
| "loss": 3.09185791015625, |
| "step": 230 |
| }, |
| { |
| "epoch": 0.1296246286794491, |
| "grad_norm": 1.2364989519119263, |
| "learning_rate": 0.00019181391994230076, |
| "loss": 3.1257530212402345, |
| "step": 240 |
| }, |
| { |
| "epoch": 0.13502565487442614, |
| "grad_norm": 1.3117403984069824, |
| "learning_rate": 0.00019145329967544175, |
| "loss": 3.0078147888183593, |
| "step": 250 |
| }, |
| { |
| "epoch": 0.14042668106940318, |
| "grad_norm": 1.060840368270874, |
| "learning_rate": 0.00019109267940858277, |
| "loss": 3.205386734008789, |
| "step": 260 |
| }, |
| { |
| "epoch": 0.14582770726438024, |
| "grad_norm": 1.0836883783340454, |
| "learning_rate": 0.0001907320591417238, |
| "loss": 3.00787410736084, |
| "step": 270 |
| }, |
| { |
| "epoch": 0.15122873345935728, |
| "grad_norm": 1.1705331802368164, |
| "learning_rate": 0.00019037143887486478, |
| "loss": 3.1415082931518556, |
| "step": 280 |
| }, |
| { |
| "epoch": 0.15662975965433432, |
| "grad_norm": 1.1469744443893433, |
| "learning_rate": 0.00019001081860800577, |
| "loss": 3.0711442947387697, |
| "step": 290 |
| }, |
| { |
| "epoch": 0.16203078584931138, |
| "grad_norm": 1.7230859994888306, |
| "learning_rate": 0.00018965019834114679, |
| "loss": 3.0760629653930662, |
| "step": 300 |
| }, |
| { |
| "epoch": 0.16743181204428842, |
| "grad_norm": 1.3305317163467407, |
| "learning_rate": 0.00018928957807428778, |
| "loss": 3.0109764099121095, |
| "step": 310 |
| }, |
| { |
| "epoch": 0.17283283823926546, |
| "grad_norm": 1.1433296203613281, |
| "learning_rate": 0.0001889289578074288, |
| "loss": 2.900168228149414, |
| "step": 320 |
| }, |
| { |
| "epoch": 0.1782338644342425, |
| "grad_norm": 1.4017837047576904, |
| "learning_rate": 0.00018856833754056978, |
| "loss": 2.9257192611694336, |
| "step": 330 |
| }, |
| { |
| "epoch": 0.18363489062921956, |
| "grad_norm": 1.418676495552063, |
| "learning_rate": 0.00018820771727371077, |
| "loss": 3.1230239868164062, |
| "step": 340 |
| }, |
| { |
| "epoch": 0.1890359168241966, |
| "grad_norm": 1.2229913473129272, |
| "learning_rate": 0.0001878470970068518, |
| "loss": 2.8634975433349608, |
| "step": 350 |
| }, |
| { |
| "epoch": 0.19443694301917364, |
| "grad_norm": 1.2936865091323853, |
| "learning_rate": 0.0001874864767399928, |
| "loss": 3.0415794372558596, |
| "step": 360 |
| }, |
| { |
| "epoch": 0.1998379692141507, |
| "grad_norm": 1.3978519439697266, |
| "learning_rate": 0.0001871258564731338, |
| "loss": 2.9328636169433593, |
| "step": 370 |
| }, |
| { |
| "epoch": 0.20523899540912774, |
| "grad_norm": 1.0732481479644775, |
| "learning_rate": 0.0001867652362062748, |
| "loss": 3.147154426574707, |
| "step": 380 |
| }, |
| { |
| "epoch": 0.21064002160410478, |
| "grad_norm": 1.304332971572876, |
| "learning_rate": 0.0001864046159394158, |
| "loss": 2.98757266998291, |
| "step": 390 |
| }, |
| { |
| "epoch": 0.21604104779908181, |
| "grad_norm": 1.2670519351959229, |
| "learning_rate": 0.00018604399567255682, |
| "loss": 3.0830259323120117, |
| "step": 400 |
| }, |
| { |
| "epoch": 0.22144207399405888, |
| "grad_norm": 1.3472830057144165, |
| "learning_rate": 0.0001856833754056978, |
| "loss": 3.0725728988647463, |
| "step": 410 |
| }, |
| { |
| "epoch": 0.22684310018903592, |
| "grad_norm": 1.5913994312286377, |
| "learning_rate": 0.0001853227551388388, |
| "loss": 2.872740936279297, |
| "step": 420 |
| }, |
| { |
| "epoch": 0.23224412638401296, |
| "grad_norm": 1.4995367527008057, |
| "learning_rate": 0.00018496213487197982, |
| "loss": 3.1985055923461916, |
| "step": 430 |
| }, |
| { |
| "epoch": 0.23764515257899002, |
| "grad_norm": 1.3097314834594727, |
| "learning_rate": 0.0001846015146051208, |
| "loss": 2.9131542205810548, |
| "step": 440 |
| }, |
| { |
| "epoch": 0.24304617877396706, |
| "grad_norm": 1.2168841361999512, |
| "learning_rate": 0.00018424089433826182, |
| "loss": 3.0022405624389648, |
| "step": 450 |
| }, |
| { |
| "epoch": 0.2484472049689441, |
| "grad_norm": 1.2428247928619385, |
| "learning_rate": 0.00018388027407140281, |
| "loss": 3.13458194732666, |
| "step": 460 |
| }, |
| { |
| "epoch": 0.25384823116392113, |
| "grad_norm": 1.0732594728469849, |
| "learning_rate": 0.00018351965380454383, |
| "loss": 2.995112991333008, |
| "step": 470 |
| }, |
| { |
| "epoch": 0.2592492573588982, |
| "grad_norm": 1.0187745094299316, |
| "learning_rate": 0.00018315903353768482, |
| "loss": 2.9637735366821287, |
| "step": 480 |
| }, |
| { |
| "epoch": 0.2646502835538752, |
| "grad_norm": 1.2797402143478394, |
| "learning_rate": 0.00018279841327082584, |
| "loss": 2.9933282852172853, |
| "step": 490 |
| }, |
| { |
| "epoch": 0.2700513097488523, |
| "grad_norm": 0.9948688745498657, |
| "learning_rate": 0.00018243779300396683, |
| "loss": 3.078018569946289, |
| "step": 500 |
| }, |
| { |
| "epoch": 0.27545233594382934, |
| "grad_norm": 1.0406428575515747, |
| "learning_rate": 0.00018207717273710782, |
| "loss": 2.9701290130615234, |
| "step": 510 |
| }, |
| { |
| "epoch": 0.28085336213880635, |
| "grad_norm": 1.184054970741272, |
| "learning_rate": 0.00018171655247024884, |
| "loss": 3.094232368469238, |
| "step": 520 |
| }, |
| { |
| "epoch": 0.2862543883337834, |
| "grad_norm": 1.2439318895339966, |
| "learning_rate": 0.00018135593220338985, |
| "loss": 3.0941631317138674, |
| "step": 530 |
| }, |
| { |
| "epoch": 0.2916554145287605, |
| "grad_norm": 1.3002263307571411, |
| "learning_rate": 0.00018099531193653084, |
| "loss": 2.9608373641967773, |
| "step": 540 |
| }, |
| { |
| "epoch": 0.2970564407237375, |
| "grad_norm": 1.5705729722976685, |
| "learning_rate": 0.00018063469166967183, |
| "loss": 3.026919364929199, |
| "step": 550 |
| }, |
| { |
| "epoch": 0.30245746691871456, |
| "grad_norm": 1.1658886671066284, |
| "learning_rate": 0.00018027407140281285, |
| "loss": 2.9997627258300783, |
| "step": 560 |
| }, |
| { |
| "epoch": 0.3078584931136916, |
| "grad_norm": 1.0013272762298584, |
| "learning_rate": 0.00017991345113595387, |
| "loss": 2.919036865234375, |
| "step": 570 |
| }, |
| { |
| "epoch": 0.31325951930866863, |
| "grad_norm": 1.1810517311096191, |
| "learning_rate": 0.00017955283086909486, |
| "loss": 3.0280284881591797, |
| "step": 580 |
| }, |
| { |
| "epoch": 0.3186605455036457, |
| "grad_norm": 1.4247483015060425, |
| "learning_rate": 0.00017919221060223585, |
| "loss": 2.944463920593262, |
| "step": 590 |
| }, |
| { |
| "epoch": 0.32406157169862276, |
| "grad_norm": 1.1780109405517578, |
| "learning_rate": 0.00017883159033537686, |
| "loss": 3.165457534790039, |
| "step": 600 |
| }, |
| { |
| "epoch": 0.3294625978935998, |
| "grad_norm": 1.082470417022705, |
| "learning_rate": 0.00017847097006851785, |
| "loss": 2.7378583908081056, |
| "step": 610 |
| }, |
| { |
| "epoch": 0.33486362408857684, |
| "grad_norm": 1.1170358657836914, |
| "learning_rate": 0.00017811034980165887, |
| "loss": 3.1393898010253904, |
| "step": 620 |
| }, |
| { |
| "epoch": 0.34026465028355385, |
| "grad_norm": 1.0197583436965942, |
| "learning_rate": 0.00017774972953479986, |
| "loss": 2.8996389389038084, |
| "step": 630 |
| }, |
| { |
| "epoch": 0.3456656764785309, |
| "grad_norm": 0.9194260835647583, |
| "learning_rate": 0.00017738910926794085, |
| "loss": 3.0526611328125, |
| "step": 640 |
| }, |
| { |
| "epoch": 0.351066702673508, |
| "grad_norm": 1.3681086301803589, |
| "learning_rate": 0.00017702848900108187, |
| "loss": 2.85855770111084, |
| "step": 650 |
| }, |
| { |
| "epoch": 0.356467728868485, |
| "grad_norm": 1.3217997550964355, |
| "learning_rate": 0.00017666786873422289, |
| "loss": 2.9244100570678713, |
| "step": 660 |
| }, |
| { |
| "epoch": 0.36186875506346206, |
| "grad_norm": 1.0755029916763306, |
| "learning_rate": 0.00017630724846736388, |
| "loss": 2.989073944091797, |
| "step": 670 |
| }, |
| { |
| "epoch": 0.3672697812584391, |
| "grad_norm": 1.2188295125961304, |
| "learning_rate": 0.00017594662820050487, |
| "loss": 3.005179023742676, |
| "step": 680 |
| }, |
| { |
| "epoch": 0.37267080745341613, |
| "grad_norm": 1.0459349155426025, |
| "learning_rate": 0.00017558600793364588, |
| "loss": 2.9304704666137695, |
| "step": 690 |
| }, |
| { |
| "epoch": 0.3780718336483932, |
| "grad_norm": 1.1236655712127686, |
| "learning_rate": 0.0001752253876667869, |
| "loss": 3.048674964904785, |
| "step": 700 |
| }, |
| { |
| "epoch": 0.38347285984337026, |
| "grad_norm": 1.0007507801055908, |
| "learning_rate": 0.00017486476739992786, |
| "loss": 3.1241172790527343, |
| "step": 710 |
| }, |
| { |
| "epoch": 0.38887388603834727, |
| "grad_norm": 1.2311853170394897, |
| "learning_rate": 0.00017450414713306888, |
| "loss": 2.8489208221435547, |
| "step": 720 |
| }, |
| { |
| "epoch": 0.39427491223332434, |
| "grad_norm": 1.1284152269363403, |
| "learning_rate": 0.0001741435268662099, |
| "loss": 2.889890670776367, |
| "step": 730 |
| }, |
| { |
| "epoch": 0.3996759384283014, |
| "grad_norm": 1.0716314315795898, |
| "learning_rate": 0.0001737829065993509, |
| "loss": 3.019309234619141, |
| "step": 740 |
| }, |
| { |
| "epoch": 0.4050769646232784, |
| "grad_norm": 1.3626065254211426, |
| "learning_rate": 0.00017342228633249188, |
| "loss": 2.8685939788818358, |
| "step": 750 |
| }, |
| { |
| "epoch": 0.4104779908182555, |
| "grad_norm": 1.133290410041809, |
| "learning_rate": 0.0001730616660656329, |
| "loss": 2.813917541503906, |
| "step": 760 |
| }, |
| { |
| "epoch": 0.4158790170132325, |
| "grad_norm": 1.2982553243637085, |
| "learning_rate": 0.0001727010457987739, |
| "loss": 3.066880798339844, |
| "step": 770 |
| }, |
| { |
| "epoch": 0.42128004320820955, |
| "grad_norm": 1.0441523790359497, |
| "learning_rate": 0.0001723404255319149, |
| "loss": 2.8972396850585938, |
| "step": 780 |
| }, |
| { |
| "epoch": 0.4266810694031866, |
| "grad_norm": 1.1574914455413818, |
| "learning_rate": 0.0001719798052650559, |
| "loss": 3.0566661834716795, |
| "step": 790 |
| }, |
| { |
| "epoch": 0.43208209559816363, |
| "grad_norm": 1.146910309791565, |
| "learning_rate": 0.0001716191849981969, |
| "loss": 2.9340593338012697, |
| "step": 800 |
| }, |
| { |
| "epoch": 0.4374831217931407, |
| "grad_norm": 1.1724562644958496, |
| "learning_rate": 0.0001712585647313379, |
| "loss": 2.9801464080810547, |
| "step": 810 |
| }, |
| { |
| "epoch": 0.44288414798811776, |
| "grad_norm": 1.3650254011154175, |
| "learning_rate": 0.00017089794446447892, |
| "loss": 2.9490041732788086, |
| "step": 820 |
| }, |
| { |
| "epoch": 0.44828517418309477, |
| "grad_norm": 1.0346980094909668, |
| "learning_rate": 0.00017053732419761993, |
| "loss": 2.971526336669922, |
| "step": 830 |
| }, |
| { |
| "epoch": 0.45368620037807184, |
| "grad_norm": 0.9564275741577148, |
| "learning_rate": 0.00017017670393076092, |
| "loss": 3.0509788513183596, |
| "step": 840 |
| }, |
| { |
| "epoch": 0.4590872265730489, |
| "grad_norm": 1.4447107315063477, |
| "learning_rate": 0.0001698160836639019, |
| "loss": 2.9528224945068358, |
| "step": 850 |
| }, |
| { |
| "epoch": 0.4644882527680259, |
| "grad_norm": 1.0819096565246582, |
| "learning_rate": 0.00016945546339704293, |
| "loss": 3.0397647857666015, |
| "step": 860 |
| }, |
| { |
| "epoch": 0.469889278963003, |
| "grad_norm": 1.3749688863754272, |
| "learning_rate": 0.00016909484313018395, |
| "loss": 2.9335342407226563, |
| "step": 870 |
| }, |
| { |
| "epoch": 0.47529030515798004, |
| "grad_norm": 0.9845815300941467, |
| "learning_rate": 0.0001687342228633249, |
| "loss": 3.1886215209960938, |
| "step": 880 |
| }, |
| { |
| "epoch": 0.48069133135295705, |
| "grad_norm": 1.0727243423461914, |
| "learning_rate": 0.00016837360259646593, |
| "loss": 3.071747970581055, |
| "step": 890 |
| }, |
| { |
| "epoch": 0.4860923575479341, |
| "grad_norm": 1.0900219678878784, |
| "learning_rate": 0.00016801298232960694, |
| "loss": 2.9294233322143555, |
| "step": 900 |
| }, |
| { |
| "epoch": 0.4914933837429111, |
| "grad_norm": 1.5818015336990356, |
| "learning_rate": 0.00016765236206274793, |
| "loss": 2.7862874984741213, |
| "step": 910 |
| }, |
| { |
| "epoch": 0.4968944099378882, |
| "grad_norm": 1.1314663887023926, |
| "learning_rate": 0.00016729174179588892, |
| "loss": 2.8420650482177736, |
| "step": 920 |
| }, |
| { |
| "epoch": 0.5022954361328652, |
| "grad_norm": 1.2435640096664429, |
| "learning_rate": 0.00016693112152902994, |
| "loss": 2.8855379104614256, |
| "step": 930 |
| }, |
| { |
| "epoch": 0.5076964623278423, |
| "grad_norm": 1.0684075355529785, |
| "learning_rate": 0.00016657050126217096, |
| "loss": 2.908370018005371, |
| "step": 940 |
| }, |
| { |
| "epoch": 0.5130974885228193, |
| "grad_norm": 1.4429422616958618, |
| "learning_rate": 0.00016620988099531195, |
| "loss": 3.0200515747070313, |
| "step": 950 |
| }, |
| { |
| "epoch": 0.5184985147177964, |
| "grad_norm": 1.3540408611297607, |
| "learning_rate": 0.00016584926072845294, |
| "loss": 2.653054428100586, |
| "step": 960 |
| }, |
| { |
| "epoch": 0.5238995409127735, |
| "grad_norm": 1.1349395513534546, |
| "learning_rate": 0.00016548864046159395, |
| "loss": 3.068928909301758, |
| "step": 970 |
| }, |
| { |
| "epoch": 0.5293005671077504, |
| "grad_norm": 1.0107704401016235, |
| "learning_rate": 0.00016512802019473494, |
| "loss": 2.912160110473633, |
| "step": 980 |
| }, |
| { |
| "epoch": 0.5347015933027275, |
| "grad_norm": 1.402188777923584, |
| "learning_rate": 0.00016476739992787596, |
| "loss": 3.3307727813720702, |
| "step": 990 |
| }, |
| { |
| "epoch": 0.5401026194977046, |
| "grad_norm": 1.200325608253479, |
| "learning_rate": 0.00016440677966101695, |
| "loss": 3.0100019454956053, |
| "step": 1000 |
| }, |
| { |
| "epoch": 0.5455036456926816, |
| "grad_norm": 1.1472864151000977, |
| "learning_rate": 0.00016404615939415794, |
| "loss": 2.938519096374512, |
| "step": 1010 |
| }, |
| { |
| "epoch": 0.5509046718876587, |
| "grad_norm": 1.184414267539978, |
| "learning_rate": 0.00016368553912729896, |
| "loss": 2.9431873321533204, |
| "step": 1020 |
| }, |
| { |
| "epoch": 0.5563056980826357, |
| "grad_norm": 1.0453943014144897, |
| "learning_rate": 0.00016332491886043998, |
| "loss": 2.98077335357666, |
| "step": 1030 |
| }, |
| { |
| "epoch": 0.5617067242776127, |
| "grad_norm": 1.1441236734390259, |
| "learning_rate": 0.00016296429859358097, |
| "loss": 3.0176897048950195, |
| "step": 1040 |
| }, |
| { |
| "epoch": 0.5671077504725898, |
| "grad_norm": 0.9775217771530151, |
| "learning_rate": 0.00016260367832672196, |
| "loss": 3.0514453887939452, |
| "step": 1050 |
| }, |
| { |
| "epoch": 0.5725087766675668, |
| "grad_norm": 1.6667460203170776, |
| "learning_rate": 0.00016224305805986297, |
| "loss": 2.9403846740722654, |
| "step": 1060 |
| }, |
| { |
| "epoch": 0.5779098028625439, |
| "grad_norm": 1.2570966482162476, |
| "learning_rate": 0.000161882437793004, |
| "loss": 2.7937477111816404, |
| "step": 1070 |
| }, |
| { |
| "epoch": 0.583310829057521, |
| "grad_norm": 1.3722922801971436, |
| "learning_rate": 0.00016152181752614498, |
| "loss": 2.8590538024902346, |
| "step": 1080 |
| }, |
| { |
| "epoch": 0.588711855252498, |
| "grad_norm": 1.0699330568313599, |
| "learning_rate": 0.00016116119725928597, |
| "loss": 2.8490627288818358, |
| "step": 1090 |
| }, |
| { |
| "epoch": 0.594112881447475, |
| "grad_norm": 1.2045018672943115, |
| "learning_rate": 0.000160800576992427, |
| "loss": 2.920571708679199, |
| "step": 1100 |
| }, |
| { |
| "epoch": 0.599513907642452, |
| "grad_norm": 1.2119923830032349, |
| "learning_rate": 0.00016043995672556798, |
| "loss": 2.8805025100708006, |
| "step": 1110 |
| }, |
| { |
| "epoch": 0.6049149338374291, |
| "grad_norm": 1.1207152605056763, |
| "learning_rate": 0.000160079336458709, |
| "loss": 2.937060737609863, |
| "step": 1120 |
| }, |
| { |
| "epoch": 0.6103159600324062, |
| "grad_norm": 1.3675578832626343, |
| "learning_rate": 0.00015971871619184998, |
| "loss": 2.9657745361328125, |
| "step": 1130 |
| }, |
| { |
| "epoch": 0.6157169862273832, |
| "grad_norm": 1.3562731742858887, |
| "learning_rate": 0.000159358095924991, |
| "loss": 3.0127033233642577, |
| "step": 1140 |
| }, |
| { |
| "epoch": 0.6211180124223602, |
| "grad_norm": 1.0387409925460815, |
| "learning_rate": 0.000158997475658132, |
| "loss": 2.876498794555664, |
| "step": 1150 |
| }, |
| { |
| "epoch": 0.6265190386173373, |
| "grad_norm": 1.2561172246932983, |
| "learning_rate": 0.000158636855391273, |
| "loss": 2.8820684432983397, |
| "step": 1160 |
| }, |
| { |
| "epoch": 0.6319200648123143, |
| "grad_norm": 1.293705701828003, |
| "learning_rate": 0.000158276235124414, |
| "loss": 2.9766061782836912, |
| "step": 1170 |
| }, |
| { |
| "epoch": 0.6373210910072914, |
| "grad_norm": 1.2604966163635254, |
| "learning_rate": 0.000157915614857555, |
| "loss": 3.119212341308594, |
| "step": 1180 |
| }, |
| { |
| "epoch": 0.6427221172022685, |
| "grad_norm": 1.094990611076355, |
| "learning_rate": 0.000157554994590696, |
| "loss": 3.071938133239746, |
| "step": 1190 |
| }, |
| { |
| "epoch": 0.6481231433972455, |
| "grad_norm": 1.1782667636871338, |
| "learning_rate": 0.00015719437432383702, |
| "loss": 3.0528400421142576, |
| "step": 1200 |
| }, |
| { |
| "epoch": 0.6535241695922225, |
| "grad_norm": 1.467441201210022, |
| "learning_rate": 0.000156833754056978, |
| "loss": 2.967136764526367, |
| "step": 1210 |
| }, |
| { |
| "epoch": 0.6589251957871995, |
| "grad_norm": 1.3159500360488892, |
| "learning_rate": 0.000156473133790119, |
| "loss": 2.9709142684936523, |
| "step": 1220 |
| }, |
| { |
| "epoch": 0.6643262219821766, |
| "grad_norm": 1.1980047225952148, |
| "learning_rate": 0.00015611251352326002, |
| "loss": 3.0129560470581054, |
| "step": 1230 |
| }, |
| { |
| "epoch": 0.6697272481771537, |
| "grad_norm": 1.125895619392395, |
| "learning_rate": 0.00015575189325640104, |
| "loss": 2.962136077880859, |
| "step": 1240 |
| }, |
| { |
| "epoch": 0.6751282743721307, |
| "grad_norm": 1.5294994115829468, |
| "learning_rate": 0.00015539127298954203, |
| "loss": 3.139348793029785, |
| "step": 1250 |
| }, |
| { |
| "epoch": 0.6805293005671077, |
| "grad_norm": 1.2642886638641357, |
| "learning_rate": 0.00015503065272268302, |
| "loss": 3.018353080749512, |
| "step": 1260 |
| }, |
| { |
| "epoch": 0.6859303267620848, |
| "grad_norm": 1.0708301067352295, |
| "learning_rate": 0.00015467003245582403, |
| "loss": 2.9507617950439453, |
| "step": 1270 |
| }, |
| { |
| "epoch": 0.6913313529570618, |
| "grad_norm": 1.2057912349700928, |
| "learning_rate": 0.00015430941218896502, |
| "loss": 2.8525314331054688, |
| "step": 1280 |
| }, |
| { |
| "epoch": 0.6967323791520389, |
| "grad_norm": 1.003139615058899, |
| "learning_rate": 0.00015394879192210604, |
| "loss": 2.952615737915039, |
| "step": 1290 |
| }, |
| { |
| "epoch": 0.702133405347016, |
| "grad_norm": 1.3539371490478516, |
| "learning_rate": 0.00015358817165524703, |
| "loss": 2.994022560119629, |
| "step": 1300 |
| }, |
| { |
| "epoch": 0.707534431541993, |
| "grad_norm": 0.9936187863349915, |
| "learning_rate": 0.00015322755138838802, |
| "loss": 3.0847434997558594, |
| "step": 1310 |
| }, |
| { |
| "epoch": 0.71293545773697, |
| "grad_norm": 1.077620267868042, |
| "learning_rate": 0.00015286693112152904, |
| "loss": 3.066029930114746, |
| "step": 1320 |
| }, |
| { |
| "epoch": 0.718336483931947, |
| "grad_norm": 1.4198359251022339, |
| "learning_rate": 0.00015250631085467005, |
| "loss": 2.7741952896118165, |
| "step": 1330 |
| }, |
| { |
| "epoch": 0.7237375101269241, |
| "grad_norm": 1.1443026065826416, |
| "learning_rate": 0.00015214569058781104, |
| "loss": 2.9916961669921873, |
| "step": 1340 |
| }, |
| { |
| "epoch": 0.7291385363219012, |
| "grad_norm": 1.142872929573059, |
| "learning_rate": 0.00015178507032095203, |
| "loss": 2.9340869903564455, |
| "step": 1350 |
| }, |
| { |
| "epoch": 0.7345395625168782, |
| "grad_norm": 1.061112403869629, |
| "learning_rate": 0.00015142445005409305, |
| "loss": 2.909784126281738, |
| "step": 1360 |
| }, |
| { |
| "epoch": 0.7399405887118553, |
| "grad_norm": 1.5785682201385498, |
| "learning_rate": 0.00015106382978723407, |
| "loss": 2.6537490844726563, |
| "step": 1370 |
| }, |
| { |
| "epoch": 0.7453416149068323, |
| "grad_norm": 1.0441385507583618, |
| "learning_rate": 0.00015070320952037503, |
| "loss": 3.0098087310791017, |
| "step": 1380 |
| }, |
| { |
| "epoch": 0.7507426411018093, |
| "grad_norm": 1.1422723531723022, |
| "learning_rate": 0.00015034258925351605, |
| "loss": 3.039506721496582, |
| "step": 1390 |
| }, |
| { |
| "epoch": 0.7561436672967864, |
| "grad_norm": 1.1552327871322632, |
| "learning_rate": 0.00014998196898665707, |
| "loss": 2.951033020019531, |
| "step": 1400 |
| } |
| ], |
| "logging_steps": 10, |
| "max_steps": 5556, |
| "num_input_tokens_seen": 0, |
| "num_train_epochs": 3, |
| "save_steps": 200, |
| "stateful_callbacks": { |
| "TrainerControl": { |
| "args": { |
| "should_epoch_stop": false, |
| "should_evaluate": false, |
| "should_log": false, |
| "should_save": true, |
| "should_training_stop": false |
| }, |
| "attributes": {} |
| } |
| }, |
| "total_flos": 5.089704563699712e+16, |
| "train_batch_size": 2, |
| "trial_name": null, |
| "trial_params": null |
| } |
|
|
