Skip to content

Instantly share code, notes, and snippets.

@stsievert

stsievert/2017-09-01.csv

Last active September 1, 2017 20:42
Show Gist options

  • Select an option

    Learn more about clone URLs
  • Save stsievert/31f34ee76d7c31bd68acb96ebe8f11ef to your computer and use it in GitHub Desktop.

Select an option

Learn more about clone URLs
Save stsievert/31f34ee76d7c31bd68acb96ebe8f11ef to your computer and use it in GitHub Desktop.
Download ZIP
Growing batch size timing
Raw
2017-09-01.csv
approx_grad beta chunks d max_iter n pointwise_loss rel_error seed sigma wall_time
0 True [-0.33772545 0.23013211 0.33314611 -0.20971334 -0.30575033 0.37405672 -0.0630507 0.18395188 0.12185236 -0.55615217] 2000 10 1 8000000 404087362.5696646 0.5786328175760411 0 7.0 9.872261762619019
1 True [-0.48261279 0.32924311 0.34738249 -0.15459457 -0.19415953 0.42306455 -0.17252527 0.05592582 0.15623894 -0.43401917] 2000 10 1 8000000 403110631.53552437 0.5542360022939973 1 7.0 7.958230018615723
2 True [-0.38672315 0.18469769 0.28169293 -0.05405629 -0.22192282 0.33552479 -0.11270603 0.00994565 0.1491486 -0.51029152] 2000 10 1 8000000 405328298.33456 0.6073403494210304 2 7.0 7.640789985656738
3 True [-0.38046754 0.23782463 0.18564596 -0.18898523 -0.20634251 0.38498398 -0.27994777 0.05463027 0.1997645 -0.4443485 ] 2000 10 1 8000000 405016186.76993334 0.6002291017741715 3 7.0 9.718455076217651
4 True [-0.50255132 0.36809792 0.30386753 -0.09045098 -0.10998031 0.2425631 -0.18396706 0.0739688 0.19426921 -0.60487433] 2000 10 1 8000000 403628905.38647205 0.5669817044056502 4 7.0 9.272586107254028
5 False [-1.41485996 0.98485814 1.03030351 -0.51697082 -0.86519827 1.50013434 -0.47727208 0.01710804 0.81731256 -1.8128364 ] 2000 10 1 8000000 405024480.2475509 0.5995344874321145 0 7.0 88.51962184906006
6 False [-1.41485996 0.98485814 1.03030351 -0.51697082 -0.86519827 1.50013434 -0.47727208 0.01710804 0.81731256 -1.8128364 ] 2000 10 1 8000000 405024480.2475509 0.5995344874321145 1 7.0 72.5514748096466
7 True [-0.5943318 0.43971841 0.50080667 -0.24237454 -0.44654503 0.65494204 -0.02191296 0.17804982 0.33975226 -0.89895889] 2000 10 2 8000000 395448737.1142796 0.30623972828770346 0 7.0 10.498352766036987
8 True [-0.81280704 0.3892416 0.4144969 -0.19841533 -0.47100421 0.73443749 -0.29300972 0.03410166 0.53298055 -0.94692223] 2000 10 2 8000000 393745566.9921506 0.2138264606829495 1 7.0 10.346577882766724
9 True [-0.67909754 0.54504072 0.53223193 -0.14396416 -0.51022011 0.756577 -0.11488156 -0.03861813 0.30696423 -0.90014167] 2000 10 2 8000000 394154971.6913496 0.23966802233459264 2 7.0 10.392645120620728
10 True [-0.81114016 0.62761384 0.47308587 -0.18381714 -0.40054955 0.75584019 -0.4229697 0.06692303 0.24404778 -0.95903262] 2000 10 2 8000000 393915237.67307943 0.2250417472016826 3 7.0 10.574932098388672
11 True [-0.80950715 0.39580708 0.55266931 -0.30895738 -0.42004591 0.67947341 -0.14089337 -0.04257892 0.37934093 -0.9941112 ] 2000 10 2 8000000 393761112.8971651 0.21548270480883275 4 7.0 11.583020210266113
12 False [-0.88388511 0.61562346 0.6424269 -0.32293019 -0.54162353 0.93730943 -0.29785334 0.01154062 0.51218003 -1.13268237] 2000 10 2 8000000 392098738.7839557 0.0031580153358463157 0 7.0 111.78914093971252
13 False [-0.88388511 0.61562346 0.6424269 -0.32293019 -0.54162353 0.93730943 -0.29785334 0.01154062 0.51218003 -1.13268237] 2000 10 2 8000000 392098738.7839557 0.0031580153358463157 1 7.0 115.92996788024902
14 True [-0.90465601 0.63266138 0.55040893 -0.29349506 -0.6703245 0.99788901 -0.29683437 -0.00451966 0.63901632 -1.13800616] 2000 10 3 8000000 392474274.2668188 0.10178542461115976 0 7.0 16.024066925048828
15 True [-0.90831729 0.72885827 0.58785034 -0.25637456 -0.51163957 0.82166663 -0.29354426 0.00685726 0.36823157 -1.06621979] 2000 10 3 8000000 392580710.7184185 0.11609047177517823 1 7.0 17.255850076675415
16 True [-0.83890115 0.59482835 0.72381753 -0.2997997 -0.65916577 0.82885403 -0.3125268 0.14498232 0.57041606 -1.27209007] 2000 10 3 8000000 392708106.3297756 0.1317116829867297 2 7.0 15.214453935623169
17 True [-0.94140786 0.71792762 0.55559608 -0.49420754 -0.40115296 0.82940781 -0.20421473 -0.15308573 0.38000895 -1.05377066] 2000 10 3 8000000 393229638.7225437 0.17764704001228399 3 7.0 18.07921314239502
18 True [-0.9264768 0.64742598 0.46509358 -0.31031413 -0.29125266 0.71523944 -0.22762164 0.1766211 0.4446097 -1.13484643] 2000 10 3 8000000 393565654.71435046 0.20169795281469685 4 7.0 16.644001722335815
19 False [-0.88407524 0.61558118 0.64313645 -0.32301339 -0.54121421 0.93744053 -0.2980613 0.01114319 0.51154152 -1.13284341] 2000 10 3 8000000 392098732.22278774 0.0031039884258993638 0 7.0 145.35059690475464
20 False [-0.88407524 0.61558118 0.64313645 -0.32301339 -0.54121421 0.93744053 -0.2980613 0.01114319 0.51154152 -1.13284341] 2000 10 3 8000000 392098732.22278774 0.0031039884258993633 1 7.0 147.1575231552124
21 True [-0.82717267 0.49951926 0.70451236 -0.28508626 -0.6863329 0.86017222 -0.50897543 -0.06062527 0.51600433 -1.27950631] 2000 10 4 8000000 393059855.81411487 0.16445766469948245 0 7.0 22.655385971069336
22 True [-0.96167651 0.53437575 0.55646929 -0.20815412 -0.65995117 1.00051923 -0.47053635 0.01242918 0.57915924 -1.29088887] 2000 10 4 8000000 392983624.5057327 0.156090826714074 1 7.0 19.843119859695435
23 True [-0.83111718 0.52141042 0.94626717 -0.34564576 -0.29334977 0.95600683 -0.2213062 -0.00498951 0.86293912 -1.17451326] 2000 10 4 8000000 394477000.17289954 0.25672332004490195 2 7.0 20.78614115715027
24 True [-0.8011392 0.1653294 0.69545393 -0.28936401 -0.72175196 1.09165144 -0.43199373 -0.11066348 0.28553981 -1.3559929 ] 2000 10 4 8000000 395327809.9076261 0.3000645202606274 3 7.0 23.399425983428955
25 True [-0.70535766 0.62419301 0.5853792 -0.25212155 -0.83368049 0.79484887 -0.27825392 0.01808917 0.22836203 -0.97395074] 2000 10 4 8000000 394115898.34814155 0.23886453574779987 4 7.0 22.683278799057007
26 False [-0.88401493 0.61559524 0.64291446 -0.32298714 -0.54134174 0.93739815 -0.29799578 0.01126732 0.51174173 -1.13279222] 2000 10 4 8000000 392098731.9225391 0.003110684443249633 0 7.0 189.257798910141
27 False [-0.88401493 0.61559524 0.64291446 -0.32298714 -0.54134174 0.93739815 -0.29799578 0.01126732 0.51174173 -1.13279222] 2000 10 4 8000000 392098731.9225391 0.003110684443249633 1 7.0 186.81083416938782
28 True [-0.94073745 0.59954698 0.38294218 -0.50008657 -0.57371659 1.11795904 -0.05387964 -0.06221731 -0.00254911 -0.82382329] 2000 10 5 8000000 396586220.0473553 0.3538455628010787 0 7.0 25.987020015716553
29 True [-0.49644144 0.67184472 0.68460269 -0.01046103 -0.34100874 0.9586469 -0.21698443 -0.13017291 0.91255365 -0.60877695] 2000 10 5 8000000 398139813.413357 0.4096640005375313 1 7.0 22.02472710609436
30 True [-0.88296262 1.12126096 0.5684551 -0.65056919 -0.93723157 0.84211933 -0.4219679 -0.08570587 0.08368869 -1.27925514] 2000 10 5 8000000 398210025.845274 0.41423200443526725 2 7.0 22.424404859542847
31 True [-0.82763905 0.62548671 0.32506712 -0.17258921 -0.19899367 0.80237269 -0.29783989 0.15952478 0.76081969 -1.0561405 ] 2000 10 5 8000000 394918666.8340963 0.2788153092040237 3 7.0 23.654550075531006
32 True [-1.06633797 0.51377482 0.751545 -0.86739486 -0.50426436 1.40075141 -0.45931405 0.13862134 0.37183032 -1.25547284] 2000 10 5 8000000 397256096.00919074 0.37940869669671184 4 7.0 30.06166672706604
33 False [-0.88403631 0.61559013 0.64299255 -0.32299642 -0.54129698 0.93741332 -0.29801892 0.0112237 0.51167121 -1.13281038] 2000 10 5 8000000 392098731.7830447 0.003107192716169649 0 7.0 201.68372106552124
34 False [-0.88403631 0.61559013 0.64299255 -0.32299642 -0.54129698 0.93741332 -0.29801892 0.0112237 0.51167121 -1.13281038] 2000 10 5 8000000 392098731.7830447 0.003107192716169649 1 7.0 204.56695413589478
35 True [-0.787126 0.39617331 0.69585277 -0.31420601 -0.63767675 0.96088101 -0.31250841 0.08209674 0.51556585 -1.23084615] 2000 10 6 8000000 392779349.549917 0.13854581491536738 0 7.0 36.94024991989136
36 True [-0.60311584 0.67094872 0.77021633 -0.46545545 -0.77681433 0.90036084 -0.37337881 -0.07810377 0.50542735 -1.11630109] 2000 10 6 8000000 393612977.4438118 0.20778431721208035 1 7.0 28.863558053970337
37 True [-0.86646249 0.43661075 0.7988148 -0.29267853 -0.54280208 0.81562045 -0.1981682 -0.03402006 0.37078323 -1.13776939] 2000 10 6 8000000 392932772.7201239 0.15320684314729843 2 7.0 28.86776614189148
38 True [-0.95081919 0.58030692 0.60816377 -0.43163165 -0.50129482 0.96855226 -0.24577224 0.06299813 0.25903702 -1.25212773] 2000 10 6 8000000 392936926.844349 0.15315234339841183 3 7.0 27.614052772521973
39 True [-0.80910288 0.45004048 0.70615963 -0.32569989 -0.49755334 0.92538758 -0.11361958 0.24122491 0.53651266 -1.12925193] 2000 10 6 8000000 393111480.63514334 0.1684587208741917 4 7.0 26.95098090171814
40 False [-0.88403686 0.61559005 0.64299482 -0.32299667 -0.54129564 0.93741365 -0.29801955 0.01122241 0.51166921 -1.13281084] 2000 10 6 8000000 392098731.78297484 0.00310712691370173 0 7.0 229.23673105239868
41 False [-0.88403686 0.61559005 0.64299482 -0.32299667 -0.54129564 0.93741365 -0.29801955 0.01122241 0.51166921 -1.13281084] 2000 10 6 8000000 392098731.78297484 0.00310712691370173 1 7.0 227.5106761455536
42 True [-0.92891464 0.58920739 0.70124001 -0.4208002 -0.43386872 0.93308967 -0.40059665 0.06483459 0.56730061 -1.22562602] 2000 10 10 8000000 392517421.8296241 0.10793165451244967 0 7.0 52.241666078567505
43 True [-0.97438049 0.54104024 0.77044339 -0.35185475 -0.46356643 0.94816046 -0.40504544 0.06890603 0.43466901 -1.03549392] 2000 10 10 8000000 392636046.76928383 0.12258435738129043 1 7.0 49.099812269210815
44 True [-0.96919768 0.64755943 0.75085001 -0.32858754 -0.54623163 0.89221101 -0.29444048 -0.00778929 0.54280554 -1.05990107] 2000 10 10 8000000 392327909.8895484 0.08030740531555673 2 7.0 53.91449022293091
45 True [-0.91026498 0.54980261 0.73983513 -0.29663315 -0.64841808 0.96840167 -0.10330012 0.05913744 0.74038523 -1.18103853] 2000 10 10 8000000 393077610.3466072 0.16494774361606168 3 7.0 56.24317216873169
46 True [-0.86751863 0.45513845 0.60480757 -0.28315057 -0.58608859 0.86741129 -0.41680482 -0.11345321 0.38210985 -1.0757813 ] 2000 10 10 8000000 392784311.2327362 0.1385175667205294 4 7.0 62.55457830429077
47 False [-0.8840367 0.61559008 0.64299419 -0.3229966 -0.54129601 0.93741355 -0.29801937 0.01122277 0.51166976 -1.13281071] 2000 10 10 8000000 392098731.7829722 0.003107145577421732 0 7.0 301.74830627441406
48 False [-0.8840367 0.61559008 0.64299419 -0.3229966 -0.54129601 0.93741355 -0.29801937 0.01122277 0.51166976 -1.13281071] 2000 10 10 8000000 392098731.7829722 0.003107145577421732 1 7.0 283.68756103515625
49 True [-0.87922602 0.77858377 0.58545987 -0.26521603 -0.50339638 0.9101223 -0.32308699 0.07266678 0.60904047 -1.01549128] 2000 10 15 8000000 392603554.26582336 0.1182117141794517 0 7.0 88.87447690963745
50 True [-0.92646157 0.57933761 0.54597877 -0.34851875 -0.73708875 1.00985498 -0.26970419 -0.15363979 0.38003248 -1.2006135 ] 2000 10 15 8000000 392952025.853789 0.15435494003660002 1 7.0 82.69940185546875
51 True [-0.92060271 0.73987876 0.72700951 -0.38870409 -0.61115147 0.7152245 -0.44118941 0.11140239 0.52156922 -1.07974972] 2000 10 15 8000000 393025962.4803908 0.1625869001127405 2 7.0 92.84537315368652
52 True [-0.78787428 0.67993592 0.5416288 -0.37731163 -0.54935883 0.80540931 -0.41572109 -0.10753865 0.50140264 -1.02760026] 2000 10 15 8000000 392764542.0845498 0.13723331555685436 3 7.0 90.04433107376099
53 True [-0.87445652 0.60443621 0.4699015 -0.2077965 -0.53557315 0.79404823 -0.24735646 -0.13633622 0.56976399 -1.1292764 ] 2000 10 15 8000000 392833172.4833296 0.14196395307157753 4 7.0 80.92051577568054
54 False [-0.8840367 0.61559008 0.64299419 -0.3229966 -0.54129601 0.93741355 -0.29801937 0.01122277 0.51166976 -1.13281071] 2000 10 15 8000000 392098731.7829722 0.003107145577421732 0 7.0 265.0206460952759
55 False [-0.8840367 0.61559008 0.64299419 -0.3229966 -0.54129601 0.93741355 -0.29801937 0.01122277 0.51166976 -1.13281071] 2000 10 15 8000000 392098731.7829722 0.003107145577421732 1 7.0 259.3626790046692
56 True [-0.90420116 0.70294056 0.71100573 -0.37047243 -0.43615661 0.91492584 -0.29503455 0.12055346 0.52098653 -0.92563523] 2000 10 20 8000000 392750046.25591063 0.13527978657110445 0 7.0 108.67117500305176
57 True [-0.97198474 0.56169638 0.74204087 -0.34829215 -0.46395348 1.01752112 -0.51153412 0.10594827 0.46804927 -1.21544019] 2000 10 20 8000000 392873611.206111 0.14661583431505454 1 7.0 117.8687379360199
58 True [-0.72389224 0.48440092 0.54839895 -0.28848806 -0.52919041 0.96645791 -0.42821098 0.09406156 0.5318906 -1.11038011] 2000 10 20 8000000 392728243.39322346 0.1327308431752905 2 7.0 114.08178377151489
59 True [ -9.43806075e-01 5.63991465e-01 5.47209796e-01 -4.29144702e-01 -4.96493694e-01 8.40751725e-01 -2.76133337e-01 -8.97474662e-04 6.53441663e-01 -1.16824473e+00] 2000 10 20 8000000 392578910.63095534 0.11538053946876133 3 7.0 117.60523080825806
60 True [-0.92213781 0.78395028 0.48953743 -0.35216695 -0.67276374 0.93338894 -0.28668526 0.05098494 0.56525422 -1.18400844] 2000 10 20 8000000 392728530.95368534 0.1327486638410105 4 7.0 118.73969268798828
61 False [-0.8840367 0.61559008 0.64299419 -0.3229966 -0.54129601 0.93741355 -0.29801937 0.01122277 0.51166976 -1.13281071] 2000 10 20 8000000 392098731.7829722 0.003107145577421732 0 7.0 267.0480582714081
62 False [-0.8840367 0.61559008 0.64299419 -0.3229966 -0.54129601 0.93741355 -0.29801937 0.01122277 0.51166976 -1.13281071] 2000 10 20 8000000 392098731.7829722 0.003107145577421732 1 7.0 262.2125267982483
63 True [ -8.82913947e-01 6.25917699e-01 5.61003585e-01 -2.92903416e-01 -5.45177322e-01 9.79526520e-01 -2.51585480e-01 3.10356345e-05 5.43200811e-01 -1.10051472e+00] 2000 10 30 8000000 392209513.1652502 0.05404568835450826 0 7.0 195.57359433174133
64 True [-0.86358949 0.62483117 0.63092854 -0.35549949 -0.57298666 0.95078018 -0.3321868 -0.03122305 0.51210296 -1.07047936] 2000 10 30 8000000 392176658.7978555 0.04775019218448381 1 7.0 185.2518310546875
65 True [-0.85402207 0.60317804 0.64227377 -0.36969844 -0.61241052 0.8853708 -0.24456717 0.01286694 0.60608717 -1.11310804] 2000 10 30 8000000 392284167.6319425 0.07326080170913572 2 7.0 195.97774267196655
66 True [-0.88279421 0.61662684 0.58289771 -0.29297874 -0.62129476 0.90724552 -0.27483043 0.06014788 0.54331724 -1.11212129] 2000 10 30 8000000 392228262.4553793 0.06074344122218579 3 7.0 206.70327305793762
67 True [-0.86292853 0.56714302 0.59393631 -0.31363468 -0.48504145 0.90558694 -0.33866759 0.02498044 0.49453811 -1.06576888] 2000 10 30 8000000 392227507.05134 0.059750956746531476 4 7.0 202.11913299560547
68 False [-0.8840367 0.61559008 0.64299419 -0.3229966 -0.54129601 0.93741355 -0.29801937 0.01122277 0.51166976 -1.13281071] 2000 10 30 8000000 392098731.7829722 0.003107145577421732 0 7.0 256.4530289173126
69 False [-0.8840367 0.61559008 0.64299419 -0.3229966 -0.54129601 0.93741355 -0.29801937 0.01122277 0.51166976 -1.13281071] 2000 10 30 8000000 392098731.7829722 0.003107145577421732 1 7.0 271.3524467945099
70 True [-0.91441156 0.6030155 0.57130643 -0.34545575 -0.6424886 0.89319701 -0.23317978 -0.02907522 0.46925774 -1.12194016] 2000 10 40 8000000 392312074.65794194 0.0779374005493023 0 7.0 264.6048061847687
71 True [-0.89738872 0.62562803 0.6978309 -0.32447667 -0.54976181 0.92063601 -0.24527886 -0.02033315 0.49220564 -1.09848166] 2000 10 40 8000000 392170503.4222042 0.04541517197333465 1 7.0 269.0821568965912
72 True [-0.88233932 0.54922246 0.70197134 -0.34350388 -0.58533272 0.98359503 -0.2784477 -0.05592122 0.48873433 -1.11946138] 2000 10 40 8000000 392242635.91384065 0.06373818373096356 2 7.0 256.9263689517975
73 True [-0.82799293 0.57854061 0.68882733 -0.35448065 -0.52789448 0.93418983 -0.28833211 0.12581325 0.47795867 -1.15998478] 2000 10 40 8000000 392281900.7215971 0.0729637904656608 3 7.0 243.45107984542847
74 True [-0.90781194 0.54490536 0.68571411 -0.33039721 -0.54634003 0.95893467 -0.23782348 0.0171961 0.47949845 -1.10483101] 2000 10 40 8000000 392205987.4079259 0.05482705526360315 4 7.0 247.5673270225525
75 False [-0.8840367 0.61559008 0.64299419 -0.3229966 -0.54129601 0.93741355 -0.29801937 0.01122277 0.51166976 -1.13281071] 2000 10 40 8000000 392098731.7829722 0.003107145577421732 0 7.0 261.3412778377533
76 False [-0.8840367 0.61559008 0.64299419 -0.3229966 -0.54129601 0.93741355 -0.29801937 0.01122277 0.51166976 -1.13281071] 2000 10 40 8000000 392098731.7829722 0.003107145577421732 1 7.0 243.74875688552856
77 True [-0.88540731 0.61721786 0.64501459 -0.32189737 -0.54471082 0.93487759 -0.29811606 0.01398124 0.51438395 -1.13709064] 2000 10 80 8000000 392099221.5401707 0.005300432688336602 0 7.0 925.5044710636139
78 True [-0.88188378 0.61344218 0.64198907 -0.32382572 -0.54150351 0.93442856 -0.29847295 0.00894236 0.50935101 -1.12378869] 2000 10 80 8000000 392099628.5426975 0.006657967894667196 1 7.0 1053.5675048828125
79 True [-0.8841524 0.61927902 0.64359241 -0.32249133 -0.54057428 0.93315628 -0.30380469 0.01242947 0.51220316 -1.13304409] 2000 10 80 8000000 392099277.1428425 0.005527006304765571 2 7.0 1059.2264840602875
80 True [-0.88253359 0.60985984 0.64411037 -0.32109031 -0.53909396 0.93314032 -0.29769773 0.01260503 0.51528137 -1.12739244] 2000 10 80 8000000 392099592.01777464 0.005867858009065959 3 7.0 1028.4987251758575
81 True [-0.88717904 0.61963663 0.64313697 -0.32686824 -0.53903717 0.93933682 -0.29481606 0.01014448 0.51201917 -1.13465281] 2000 10 80 8000000 392099251.88777584 0.004409698151563556 4 7.0 1020.5811431407928
82 False [-0.8840367 0.61559008 0.64299419 -0.3229966 -0.54129601 0.93741355 -0.29801937 0.01122277 0.51166976 -1.13281071] 2000 10 80 8000000 392098731.7829722 0.003107145577421732 0 7.0 273.7842438220978
83 False [-0.8840367 0.61559008 0.64299419 -0.3229966 -0.54129601 0.93741355 -0.29801937 0.01122277 0.51166976 -1.13281071] 2000 10 80 8000000 392098731.7829722 0.003107145577421732 1 7.0 267.38317227363586
Display the source blob
Display the rendered blob
Raw
dask-glm.ipynb
{
"cells": [
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Profile step size. How long does that function call take? "
]
},
{
"cell_type": "code",
"execution_count": 1,
"metadata": {
"ExecuteTime": {
"end_time": "2017-09-01T15:11:51.075731Z",
"start_time": "2017-09-01T15:11:49.486603Z"
}
},
"outputs": [],
"source": [
"import dask_glm\n",
"from dask_glm.algorithms import gradient_descent\n",
"from dask_glm.families import Normal\n",
"import numpy as np\n",
"import dask\n",
"import dask.array as da\n",
"import time\n",
"import numpy.linalg as LA\n",
"from pprint import pprint"
]
},
{
"cell_type": "code",
"execution_count": 2,
"metadata": {
"ExecuteTime": {
"end_time": "2017-09-01T15:11:56.202848Z",
"start_time": "2017-09-01T15:11:51.077783Z"
},
"collapsed": true
},
"outputs": [],
"source": [
"n, d = int(8e6), 10\n",
"sigma = d * 0.7\n",
"X = np.random.randn(n, d)\n",
"beta_star = np.random.randn(d)\n",
"y = X @ beta_star\n",
"noise = sigma * np.random.randn(n)\n",
"y += noise"
]
},
{
"cell_type": "code",
"execution_count": 3,
"metadata": {
"ExecuteTime": {
"end_time": "2017-09-01T15:11:57.587105Z",
"start_time": "2017-09-01T15:11:56.212832Z"
},
"collapsed": true
},
"outputs": [],
"source": [
"chunks = 2000\n",
"y = da.from_array(y, chunks)\n",
"X = da.from_array(X, chunks)\n",
"family = Normal\n",
"results = []"
]
},
{
"cell_type": "code",
"execution_count": 4,
"metadata": {
"ExecuteTime": {
"end_time": "2017-09-01T19:50:29.405691Z",
"start_time": "2017-09-01T15:11:57.588712Z"
},
"scrolled": false
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"========================================\n",
"max_iter = 1\n",
"{'approx_grad': True,\n",
" 'rel_error': 0.57863281757604113,\n",
" 'wall_time': 9.872261762619019}\n",
"{'approx_grad': True,\n",
" 'rel_error': 0.55423600229399728,\n",
" 'wall_time': 7.958230018615723}\n",
"{'approx_grad': True,\n",
" 'rel_error': 0.60734034942103043,\n",
" 'wall_time': 7.640789985656738}\n",
"{'approx_grad': True,\n",
" 'rel_error': 0.60022910177417155,\n",
" 'wall_time': 9.718455076217651}\n",
"{'approx_grad': True,\n",
" 'rel_error': 0.56698170440565021,\n",
" 'wall_time': 9.272586107254028}\n",
"{'approx_grad': False,\n",
" 'rel_error': 0.59953448743211446,\n",
" 'wall_time': 88.51962184906006}\n",
"{'approx_grad': False,\n",
" 'rel_error': 0.59953448743211446,\n",
" 'wall_time': 72.5514748096466}\n",
"========================================\n",
"max_iter = 2\n",
"{'approx_grad': True,\n",
" 'rel_error': 0.30623972828770346,\n",
" 'wall_time': 10.498352766036987}\n",
"{'approx_grad': True,\n",
" 'rel_error': 0.2138264606829495,\n",
" 'wall_time': 10.346577882766724}\n",
"{'approx_grad': True,\n",
" 'rel_error': 0.23966802233459264,\n",
" 'wall_time': 10.392645120620728}\n",
"{'approx_grad': True,\n",
" 'rel_error': 0.2250417472016826,\n",
" 'wall_time': 10.574932098388672}\n",
"{'approx_grad': True,\n",
" 'rel_error': 0.21548270480883275,\n",
" 'wall_time': 11.583020210266113}\n",
"{'approx_grad': False,\n",
" 'rel_error': 0.0031580153358463157,\n",
" 'wall_time': 111.78914093971252}\n",
"{'approx_grad': False,\n",
" 'rel_error': 0.0031580153358463157,\n",
" 'wall_time': 115.92996788024902}\n",
"========================================\n",
"max_iter = 3\n",
"{'approx_grad': True,\n",
" 'rel_error': 0.10178542461115976,\n",
" 'wall_time': 16.024066925048828}\n",
"{'approx_grad': True,\n",
" 'rel_error': 0.11609047177517823,\n",
" 'wall_time': 17.255850076675415}\n",
"{'approx_grad': True,\n",
" 'rel_error': 0.13171168298672969,\n",
" 'wall_time': 15.214453935623169}\n",
"{'approx_grad': True,\n",
" 'rel_error': 0.17764704001228399,\n",
" 'wall_time': 18.07921314239502}\n",
"{'approx_grad': True,\n",
" 'rel_error': 0.20169795281469685,\n",
" 'wall_time': 16.644001722335815}\n",
"{'approx_grad': False,\n",
" 'rel_error': 0.0031039884258993638,\n",
" 'wall_time': 145.35059690475464}\n",
"{'approx_grad': False,\n",
" 'rel_error': 0.0031039884258993633,\n",
" 'wall_time': 147.1575231552124}\n",
"========================================\n",
"max_iter = 4\n",
"{'approx_grad': True,\n",
" 'rel_error': 0.16445766469948245,\n",
" 'wall_time': 22.655385971069336}\n",
"{'approx_grad': True,\n",
" 'rel_error': 0.15609082671407401,\n",
" 'wall_time': 19.843119859695435}\n",
"{'approx_grad': True,\n",
" 'rel_error': 0.25672332004490195,\n",
" 'wall_time': 20.78614115715027}\n",
"{'approx_grad': True,\n",
" 'rel_error': 0.30006452026062741,\n",
" 'wall_time': 23.399425983428955}\n",
"{'approx_grad': True,\n",
" 'rel_error': 0.23886453574779987,\n",
" 'wall_time': 22.683278799057007}\n",
"{'approx_grad': False,\n",
" 'rel_error': 0.003110684443249633,\n",
" 'wall_time': 189.257798910141}\n",
"{'approx_grad': False,\n",
" 'rel_error': 0.003110684443249633,\n",
" 'wall_time': 186.81083416938782}\n",
"========================================\n",
"max_iter = 5\n",
"{'approx_grad': True,\n",
" 'rel_error': 0.35384556280107871,\n",
" 'wall_time': 25.987020015716553}\n",
"{'approx_grad': True,\n",
" 'rel_error': 0.40966400053753133,\n",
" 'wall_time': 22.02472710609436}\n",
"{'approx_grad': True,\n",
" 'rel_error': 0.41423200443526725,\n",
" 'wall_time': 22.424404859542847}\n",
"{'approx_grad': True,\n",
" 'rel_error': 0.2788153092040237,\n",
" 'wall_time': 23.654550075531006}\n",
"{'approx_grad': True,\n",
" 'rel_error': 0.37940869669671184,\n",
" 'wall_time': 30.06166672706604}\n",
"{'approx_grad': False,\n",
" 'rel_error': 0.0031071927161696491,\n",
" 'wall_time': 201.68372106552124}\n",
"{'approx_grad': False,\n",
" 'rel_error': 0.0031071927161696491,\n",
" 'wall_time': 204.56695413589478}\n",
"========================================\n",
"max_iter = 6\n",
"{'approx_grad': True,\n",
" 'rel_error': 0.13854581491536738,\n",
" 'wall_time': 36.94024991989136}\n",
"{'approx_grad': True,\n",
" 'rel_error': 0.20778431721208035,\n",
" 'wall_time': 28.863558053970337}\n",
"{'approx_grad': True,\n",
" 'rel_error': 0.15320684314729843,\n",
" 'wall_time': 28.86776614189148}\n",
"{'approx_grad': True,\n",
" 'rel_error': 0.15315234339841183,\n",
" 'wall_time': 27.614052772521973}\n",
"{'approx_grad': True,\n",
" 'rel_error': 0.1684587208741917,\n",
" 'wall_time': 26.95098090171814}\n",
"{'approx_grad': False,\n",
" 'rel_error': 0.0031071269137017301,\n",
" 'wall_time': 229.23673105239868}\n",
"{'approx_grad': False,\n",
" 'rel_error': 0.0031071269137017301,\n",
" 'wall_time': 227.5106761455536}\n",
"========================================\n",
"max_iter = 10\n",
"{'approx_grad': True,\n",
" 'rel_error': 0.10793165451244967,\n",
" 'wall_time': 52.241666078567505}\n",
"{'approx_grad': True,\n",
" 'rel_error': 0.12258435738129043,\n",
" 'wall_time': 49.099812269210815}\n",
"{'approx_grad': True,\n",
" 'rel_error': 0.080307405315556732,\n",
" 'wall_time': 53.91449022293091}\n",
"{'approx_grad': True,\n",
" 'rel_error': 0.16494774361606168,\n",
" 'wall_time': 56.24317216873169}\n",
"{'approx_grad': True,\n",
" 'rel_error': 0.13851756672052939,\n",
" 'wall_time': 62.55457830429077}\n",
"{'approx_grad': False,\n",
" 'rel_error': 0.0031071455774217319,\n",
" 'wall_time': 301.74830627441406}\n",
"{'approx_grad': False,\n",
" 'rel_error': 0.0031071455774217319,\n",
" 'wall_time': 283.68756103515625}\n",
"========================================\n",
"max_iter = 15\n",
"{'approx_grad': True,\n",
" 'rel_error': 0.1182117141794517,\n",
" 'wall_time': 88.87447690963745}\n",
"{'approx_grad': True,\n",
" 'rel_error': 0.15435494003660002,\n",
" 'wall_time': 82.69940185546875}\n",
"{'approx_grad': True,\n",
" 'rel_error': 0.16258690011274049,\n",
" 'wall_time': 92.84537315368652}\n",
"{'approx_grad': True,\n",
" 'rel_error': 0.13723331555685436,\n",
" 'wall_time': 90.04433107376099}\n",
"{'approx_grad': True,\n",
" 'rel_error': 0.14196395307157753,\n",
" 'wall_time': 80.92051577568054}\n",
"{'approx_grad': False,\n",
" 'rel_error': 0.0031071455774217319,\n",
" 'wall_time': 265.0206460952759}\n",
"{'approx_grad': False,\n",
" 'rel_error': 0.0031071455774217319,\n",
" 'wall_time': 259.3626790046692}\n",
"========================================\n",
"max_iter = 20\n",
"{'approx_grad': True,\n",
" 'rel_error': 0.13527978657110445,\n",
" 'wall_time': 108.67117500305176}\n",
"{'approx_grad': True,\n",
" 'rel_error': 0.14661583431505454,\n",
" 'wall_time': 117.8687379360199}\n",
"{'approx_grad': True,\n",
" 'rel_error': 0.13273084317529049,\n",
" 'wall_time': 114.08178377151489}\n",
"{'approx_grad': True,\n",
" 'rel_error': 0.11538053946876133,\n",
" 'wall_time': 117.60523080825806}\n",
"{'approx_grad': True,\n",
" 'rel_error': 0.1327486638410105,\n",
" 'wall_time': 118.73969268798828}\n",
"{'approx_grad': False,\n",
" 'rel_error': 0.0031071455774217319,\n",
" 'wall_time': 267.0480582714081}\n",
"{'approx_grad': False,\n",
" 'rel_error': 0.0031071455774217319,\n",
" 'wall_time': 262.2125267982483}\n",
"========================================\n",
"max_iter = 30\n",
"{'approx_grad': True,\n",
" 'rel_error': 0.054045688354508258,\n",
" 'wall_time': 195.57359433174133}\n",
"{'approx_grad': True,\n",
" 'rel_error': 0.047750192184483807,\n",
" 'wall_time': 185.2518310546875}\n",
"{'approx_grad': True,\n",
" 'rel_error': 0.073260801709135723,\n",
" 'wall_time': 195.97774267196655}\n",
"{'approx_grad': True,\n",
" 'rel_error': 0.060743441222185793,\n",
" 'wall_time': 206.70327305793762}\n",
"{'approx_grad': True,\n",
" 'rel_error': 0.059750956746531476,\n",
" 'wall_time': 202.11913299560547}\n",
"{'approx_grad': False,\n",
" 'rel_error': 0.0031071455774217319,\n",
" 'wall_time': 256.4530289173126}\n",
"{'approx_grad': False,\n",
" 'rel_error': 0.0031071455774217319,\n",
" 'wall_time': 271.3524467945099}\n",
"========================================\n",
"max_iter = 40\n",
"{'approx_grad': True,\n",
" 'rel_error': 0.077937400549302302,\n",
" 'wall_time': 264.6048061847687}\n",
"{'approx_grad': True,\n",
" 'rel_error': 0.045415171973334653,\n",
" 'wall_time': 269.0821568965912}\n",
"{'approx_grad': True,\n",
" 'rel_error': 0.063738183730963557,\n",
" 'wall_time': 256.9263689517975}\n",
"{'approx_grad': True,\n",
" 'rel_error': 0.072963790465660802,\n",
" 'wall_time': 243.45107984542847}\n",
"{'approx_grad': True,\n",
" 'rel_error': 0.054827055263603153,\n",
" 'wall_time': 247.5673270225525}\n",
"{'approx_grad': False,\n",
" 'rel_error': 0.0031071455774217319,\n",
" 'wall_time': 261.3412778377533}\n",
"{'approx_grad': False,\n",
" 'rel_error': 0.0031071455774217319,\n",
" 'wall_time': 243.74875688552856}\n",
"========================================\n",
"max_iter = 80\n",
"{'approx_grad': True,\n",
" 'rel_error': 0.005300432688336602,\n",
" 'wall_time': 925.5044710636139}\n",
"{'approx_grad': True,\n",
" 'rel_error': 0.0066579678946671961,\n",
" 'wall_time': 1053.5675048828125}\n",
"{'approx_grad': True,\n",
" 'rel_error': 0.0055270063047655714,\n",
" 'wall_time': 1059.2264840602875}\n",
"{'approx_grad': True,\n",
" 'rel_error': 0.005867858009065959,\n",
" 'wall_time': 1028.4987251758575}\n",
"{'approx_grad': True,\n",
" 'rel_error': 0.0044096981515635561,\n",
" 'wall_time': 1020.5811431407928}\n",
"{'approx_grad': False,\n",
" 'rel_error': 0.0031071455774217319,\n",
" 'wall_time': 273.7842438220978}\n"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"{'approx_grad': False,\n",
" 'rel_error': 0.0031071455774217319,\n",
" 'wall_time': 267.38317227363586}\n",
"========================================\n",
"max_iter = 120\n"
]
},
{
"ename": "KeyboardInterrupt",
"evalue": "",
"output_type": "error",
"traceback": [
"\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
"\u001b[0;31mKeyboardInterrupt\u001b[0m Traceback (most recent call last)",
"\u001b[0;32m<ipython-input-4-9708832f2434>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[1;32m 18\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mrepeat\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mrange\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mrepeats\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mapprox_grad\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m 19\u001b[0m results += [get_stats(seed=repeat, max_iter=max_iter,\n\u001b[0;32m---> 20\u001b[0;31m approx_grad=approx_grad)]\n\u001b[0m\u001b[1;32m 21\u001b[0m pprint({k: v for k, v in results[-1].items()\n\u001b[1;32m 22\u001b[0m if k in ['rel_error', 'wall_time', 'approx_grad']})\n",
"\u001b[0;32m<ipython-input-4-9708832f2434>\u001b[0m in \u001b[0;36mget_stats\u001b[0;34m(seed, max_iter, approx_grad)\u001b[0m\n\u001b[1;32m 3\u001b[0m \u001b[0mstart\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mtime\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mtime\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m 4\u001b[0m beta = gradient_descent(X, y, family=Normal, max_iter=max_iter,\n\u001b[0;32m----> 5\u001b[0;31m approx_grad=approx_grad)\n\u001b[0m\u001b[1;32m 6\u001b[0m return{'beta': beta, 'n': n, 'd': d, 'chunks': chunks,\n\u001b[1;32m 7\u001b[0m \u001b[0;34m'wall_time'\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0mtime\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mtime\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m-\u001b[0m \u001b[0mstart\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m'approx_grad'\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0mapprox_grad\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
"\u001b[0;32m/Users/scott/Work/Open-Source/dask-glm/dask_glm/utils.py\u001b[0m in \u001b[0;36mnormalize_inputs\u001b[0;34m(X, y, *args, **kwargs)\u001b[0m\n\u001b[1;32m 24\u001b[0m \u001b[0mmean\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mmean\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mlen\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mintercept_idx\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32melse\u001b[0m \u001b[0mnp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mzeros\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mmean\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mshape\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m 25\u001b[0m \u001b[0mXn\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0mX\u001b[0m \u001b[0;34m-\u001b[0m \u001b[0mmean\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m/\u001b[0m \u001b[0mstd\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 26\u001b[0;31m \u001b[0mout\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0malgo\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mXn\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0my\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m*\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcopy\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m 27\u001b[0m \u001b[0mi_adj\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mnp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0msum\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mout\u001b[0m \u001b[0;34m*\u001b[0m \u001b[0mmean\u001b[0m \u001b[0;34m/\u001b[0m \u001b[0mstd\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m 28\u001b[0m \u001b[0mout\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mintercept_idx\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m-=\u001b[0m \u001b[0mi_adj\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
"\u001b[0;32m/Users/scott/Work/Open-Source/dask-glm/dask_glm/algorithms.py\u001b[0m in \u001b[0;36mgradient_descent\u001b[0;34m(X, y, max_iter, tol, family, approx_grad, initial_batch_size, **kwargs)\u001b[0m\n\u001b[1;32m 139\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m 140\u001b[0m beta, stepSize, Xbeta, lf, func, grad, Xgradient = persist(\n\u001b[0;32m--> 141\u001b[0;31m beta, stepSize, Xbeta, lf, func, grad, Xgradient)\n\u001b[0m\u001b[1;32m 142\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m 143\u001b[0m \u001b[0mstepSize\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mlf\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfunc\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mgrad\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mcompute\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mstepSize\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mlf\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfunc\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mgrad\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
"\u001b[0;32m/Users/scott/anaconda/lib/python3.6/site-packages/dask/base.py\u001b[0m in \u001b[0;36mpersist\u001b[0;34m(*args, **kwargs)\u001b[0m\n\u001b[1;32m 584\u001b[0m \u001b[0mdsk\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mcollections_to_dsk\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mcollections\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0moptimize_graph\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m 585\u001b[0m \u001b[0mkeys\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mlist\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mflatten\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mvar\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_keys\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mvar\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mcollections\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 586\u001b[0;31m \u001b[0mresults\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mget\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mdsk\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mkeys\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m 587\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m 588\u001b[0m \u001b[0md\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mdict\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mzip\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mkeys\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mresults\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
"\u001b[0;32m/Users/scott/anaconda/lib/python3.6/site-packages/dask/threaded.py\u001b[0m in \u001b[0;36mget\u001b[0;34m(dsk, result, cache, num_workers, **kwargs)\u001b[0m\n\u001b[1;32m 73\u001b[0m results = get_async(pool.apply_async, len(pool._pool), dsk, result,\n\u001b[1;32m 74\u001b[0m \u001b[0mcache\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mcache\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mget_id\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0m_thread_get_id\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 75\u001b[0;31m pack_exception=pack_exception, **kwargs)\n\u001b[0m\u001b[1;32m 76\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m 77\u001b[0m \u001b[0;31m# Cleanup pools associated to dead threads\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
"\u001b[0;32m/Users/scott/anaconda/lib/python3.6/site-packages/dask/local.py\u001b[0m in \u001b[0;36mget_async\u001b[0;34m(apply_async, num_workers, dsk, result, cache, get_id, rerun_exceptions_locally, pack_exception, raise_exception, callbacks, dumps, loads, **kwargs)\u001b[0m\n\u001b[1;32m 510\u001b[0m \u001b[0;31m# Main loop, wait on tasks to finish, insert new ones\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m 511\u001b[0m \u001b[0;32mwhile\u001b[0m \u001b[0mstate\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m'waiting'\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;32mor\u001b[0m \u001b[0mstate\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m'ready'\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;32mor\u001b[0m \u001b[0mstate\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m'running'\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 512\u001b[0;31m \u001b[0mkey\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mres_info\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfailed\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mqueue_get\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mqueue\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m 513\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mfailed\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m 514\u001b[0m \u001b[0mexc\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtb\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mloads\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mres_info\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
"\u001b[0;32m/Users/scott/anaconda/lib/python3.6/site-packages/dask/local.py\u001b[0m in \u001b[0;36mqueue_get\u001b[0;34m(q)\u001b[0m\n\u001b[1;32m 149\u001b[0m \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m 150\u001b[0m \u001b[0;32mdef\u001b[0m \u001b[0mqueue_get\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mq\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 151\u001b[0;31m \u001b[0;32mreturn\u001b[0m \u001b[0mq\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mget\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m 152\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m 153\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
"\u001b[0;32m/Users/scott/anaconda/lib/python3.6/queue.py\u001b[0m in \u001b[0;36mget\u001b[0;34m(self, block, timeout)\u001b[0m\n\u001b[1;32m 162\u001b[0m \u001b[0;32melif\u001b[0m \u001b[0mtimeout\u001b[0m \u001b[0;32mis\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m 163\u001b[0m \u001b[0;32mwhile\u001b[0m \u001b[0;32mnot\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_qsize\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 164\u001b[0;31m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mnot_empty\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mwait\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m 165\u001b[0m \u001b[0;32melif\u001b[0m \u001b[0mtimeout\u001b[0m \u001b[0;34m<\u001b[0m \u001b[0;36m0\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m 166\u001b[0m \u001b[0;32mraise\u001b[0m \u001b[0mValueError\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m\"'timeout' must be a non-negative number\"\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
"\u001b[0;32m/Users/scott/anaconda/lib/python3.6/threading.py\u001b[0m in \u001b[0;36mwait\u001b[0;34m(self, timeout)\u001b[0m\n\u001b[1;32m 293\u001b[0m \u001b[0;32mtry\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0;31m# restore state no matter what (e.g., KeyboardInterrupt)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m 294\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mtimeout\u001b[0m \u001b[0;32mis\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 295\u001b[0;31m \u001b[0mwaiter\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0macquire\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m 296\u001b[0m \u001b[0mgotit\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;32mTrue\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m 297\u001b[0m \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
"\u001b[0;31mKeyboardInterrupt\u001b[0m: "
]
}
],
"source": [
"def get_stats(seed=42, max_iter=100, approx_grad=True):\n",
" np.random.seed(seed)\n",
" start = time.time()\n",
" beta = gradient_descent(X, y, family=Normal, max_iter=max_iter,\n",
" approx_grad=approx_grad)\n",
" return{'beta': beta, 'n': n, 'd': d, 'chunks': chunks,\n",
" 'wall_time': time.time() - start, 'approx_grad': approx_grad,\n",
" 'max_iter': max_iter, 'sigma': sigma,\n",
" 'seed': seed,\n",
" 'rel_error': LA.norm(beta - beta_star) / LA.norm(beta_star),\n",
" 'pointwise_loss': family.pointwise_loss(beta, X, y).compute()}\n",
" \n",
"repeats = {True: 5, False: 2}\n",
"for max_iter in [1, 2, 3, 4, 5, 6, 10, 15, 20, 30, 40, 80, 120, 160]:\n",
" print('=' * 40)\n",
" print(f'max_iter = {max_iter}')\n",
" for approx_grad in [True, False]:\n",
" for repeat in range(repeats[approx_grad]):\n",
" results += [get_stats(seed=repeat, max_iter=max_iter,\n",
" approx_grad=approx_grad)]\n",
" pprint({k: v for k, v in results[-1].items()\n",
" if k in ['rel_error', 'wall_time', 'approx_grad']})"
]
},
{
"cell_type": "code",
"execution_count": 5,
"metadata": {
"ExecuteTime": {
"end_time": "2017-09-01T19:50:56.560763Z",
"start_time": "2017-09-01T19:50:55.837910Z"
}
},
"outputs": [
{
"data": {
"text/html": [
"<div>\n",
"<style>\n",
" .dataframe thead tr:only-child th {\n",
" text-align: right;\n",
" }\n",
"\n",
" .dataframe thead th {\n",
" text-align: left;\n",
" }\n",
"\n",
" .dataframe tbody tr th {\n",
" vertical-align: top;\n",
" }\n",
"</style>\n",
"<table border=\"1\" class=\"dataframe\">\n",
" <thead>\n",
" <tr style=\"text-align: right;\">\n",
" <th></th>\n",
" <th>approx_grad</th>\n",
" <th>beta</th>\n",
" <th>chunks</th>\n",
" <th>d</th>\n",
" <th>max_iter</th>\n",
" <th>n</th>\n",
" <th>pointwise_loss</th>\n",
" <th>rel_error</th>\n",
" <th>seed</th>\n",
" <th>sigma</th>\n",
" <th>wall_time</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <th>0</th>\n",
" <td>True</td>\n",
" <td>[-0.337725448124, 0.230132106768, 0.3331461140...</td>\n",
" <td>2000</td>\n",
" <td>10</td>\n",
" <td>1</td>\n",
" <td>8000000</td>\n",
" <td>4.040874e+08</td>\n",
" <td>0.578633</td>\n",
" <td>0</td>\n",
" <td>7.0</td>\n",
" <td>9.872262</td>\n",
" </tr>\n",
" <tr>\n",
" <th>1</th>\n",
" <td>True</td>\n",
" <td>[-0.482612792982, 0.329243111397, 0.3473824920...</td>\n",
" <td>2000</td>\n",
" <td>10</td>\n",
" <td>1</td>\n",
" <td>8000000</td>\n",
" <td>4.031106e+08</td>\n",
" <td>0.554236</td>\n",
" <td>1</td>\n",
" <td>7.0</td>\n",
" <td>7.958230</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2</th>\n",
" <td>True</td>\n",
" <td>[-0.386723154211, 0.184697688514, 0.2816929331...</td>\n",
" <td>2000</td>\n",
" <td>10</td>\n",
" <td>1</td>\n",
" <td>8000000</td>\n",
" <td>4.053283e+08</td>\n",
" <td>0.607340</td>\n",
" <td>2</td>\n",
" <td>7.0</td>\n",
" <td>7.640790</td>\n",
" </tr>\n",
" <tr>\n",
" <th>3</th>\n",
" <td>True</td>\n",
" <td>[-0.380467537489, 0.237824632003, 0.1856459593...</td>\n",
" <td>2000</td>\n",
" <td>10</td>\n",
" <td>1</td>\n",
" <td>8000000</td>\n",
" <td>4.050162e+08</td>\n",
" <td>0.600229</td>\n",
" <td>3</td>\n",
" <td>7.0</td>\n",
" <td>9.718455</td>\n",
" </tr>\n",
" <tr>\n",
" <th>4</th>\n",
" <td>True</td>\n",
" <td>[-0.502551315934, 0.368097916724, 0.3038675340...</td>\n",
" <td>2000</td>\n",
" <td>10</td>\n",
" <td>1</td>\n",
" <td>8000000</td>\n",
" <td>4.036289e+08</td>\n",
" <td>0.566982</td>\n",
" <td>4</td>\n",
" <td>7.0</td>\n",
" <td>9.272586</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>\n",
"</div>"
],
"text/plain": [
" approx_grad beta chunks d \\\n",
"0 True [-0.337725448124, 0.230132106768, 0.3331461140... 2000 10 \n",
"1 True [-0.482612792982, 0.329243111397, 0.3473824920... 2000 10 \n",
"2 True [-0.386723154211, 0.184697688514, 0.2816929331... 2000 10 \n",
"3 True [-0.380467537489, 0.237824632003, 0.1856459593... 2000 10 \n",
"4 True [-0.502551315934, 0.368097916724, 0.3038675340... 2000 10 \n",
"\n",
" max_iter n pointwise_loss rel_error seed sigma wall_time \n",
"0 1 8000000 4.040874e+08 0.578633 0 7.0 9.872262 \n",
"1 1 8000000 4.031106e+08 0.554236 1 7.0 7.958230 \n",
"2 1 8000000 4.053283e+08 0.607340 2 7.0 7.640790 \n",
"3 1 8000000 4.050162e+08 0.600229 3 7.0 9.718455 \n",
"4 1 8000000 4.036289e+08 0.566982 4 7.0 9.272586 "
]
},
"execution_count": 5,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"import pandas as pd\n",
"df = pd.DataFrame(results)\n",
"# df.to_csv('2017-09-01.csv')\n",
"df.head()"
]
},
{
"cell_type": "code",
"execution_count": 7,
"metadata": {
"ExecuteTime": {
"end_time": "2017-09-01T19:51:03.880589Z",
"start_time": "2017-09-01T19:51:03.855323Z"
}
},
"outputs": [
{
"data": {
"text/html": [
"<div class=\"vega-embed\" id=\"adad4346-4c33-4fd9-8d7c-22b6667ed6ac\"></div>\n",
"\n",
"<style>\n",
".vega-embed svg, .vega-embed canvas {\n",
" border: 1px dotted gray;\n",
"}\n",
"\n",
".vega-embed .vega-actions a {\n",
" margin-right: 6px;\n",
"}\n",
"</style>\n"
]
},
"metadata": {
"jupyter-vega": "#adad4346-4c33-4fd9-8d7c-22b6667ed6ac"
},
"output_type": "display_data"
},
{
"data": {
"application/javascript": [
"var spec = {\"config\": {\"cell\": {\"width\": 500, \"height\": 350}}, \"encoding\": {\"color\": {\"field\": \"approx_grad\", \"type\": \"nominal\"}, \"x\": {\"field\": \"max_iter\", \"type\": \"quantitative\"}, \"y\": {\"aggregate\": \"min\", \"field\": \"rel_error\", \"type\": \"quantitative\"}}, \"mark\": \"line\", \"data\": {\"values\": [{\"approx_grad\": true, \"beta\": [-0.3377254481236899, 0.2301321067680867, 0.333146114043381, -0.20971334143549528, -0.30575032752005804, 0.3740567200034455, -0.06305069711196563, 0.18395187863790194, 0.12185236088423221, -0.5561521713524271], \"chunks\": 2000, \"d\": 10, \"max_iter\": 1, \"n\": 8000000, \"pointwise_loss\": 404087362.5696646, \"rel_error\": 0.5786328175760411, \"seed\": 0, \"sigma\": 7.0, \"wall_time\": 9.872261762619019}, {\"approx_grad\": true, \"beta\": [-0.4826127929824557, 0.3292431113973369, 0.3473824920275945, -0.15459457027121878, -0.19415953367371733, 0.42306454754726186, -0.17252527019265015, 0.05592581965829817, 0.15623894385874282, -0.43401916546021496], \"chunks\": 2000, \"d\": 10, \"max_iter\": 1, \"n\": 8000000, \"pointwise_loss\": 403110631.53552437, \"rel_error\": 0.5542360022939973, \"seed\": 1, \"sigma\": 7.0, \"wall_time\": 7.958230018615723}, {\"approx_grad\": true, \"beta\": [-0.3867231542110521, 0.18469768851413176, 0.2816929331607336, -0.054056285360848305, -0.22192282415811285, 0.33552478632731875, -0.11270602798687447, 0.009945650699710791, 0.1491485983684059, -0.5102915167728846], \"chunks\": 2000, \"d\": 10, \"max_iter\": 1, \"n\": 8000000, \"pointwise_loss\": 405328298.33456, \"rel_error\": 0.6073403494210304, \"seed\": 2, \"sigma\": 7.0, \"wall_time\": 7.640789985656738}, {\"approx_grad\": true, \"beta\": [-0.38046753748948453, 0.23782463200335902, 0.18564595937362763, -0.1889852272595575, -0.20634251350348562, 0.38498398172180837, -0.2799477671469581, 0.05463026794689662, 0.1997645016076342, -0.4443484968294167], \"chunks\": 2000, \"d\": 10, \"max_iter\": 1, \"n\": 8000000, \"pointwise_loss\": 405016186.76993334, \"rel_error\": 0.6002291017741715, \"seed\": 3, \"sigma\": 7.0, \"wall_time\": 9.718455076217651}, {\"approx_grad\": true, \"beta\": [-0.5025513159335334, 0.3680979167244689, 0.3038675340941493, -0.09045097996902372, -0.10998030666990562, 0.24256309845221122, -0.1839670635653993, 0.07396880196168064, 0.1942692091907179, -0.6048743288747124], \"chunks\": 2000, \"d\": 10, \"max_iter\": 1, \"n\": 8000000, \"pointwise_loss\": 403628905.38647205, \"rel_error\": 0.5669817044056502, \"seed\": 4, \"sigma\": 7.0, \"wall_time\": 9.272586107254028}, {\"approx_grad\": false, \"beta\": [-1.4148599575143694, 0.9848581358758047, 1.0303035090932562, -0.5169708152326576, -0.8651982685784295, 1.5001343392240916, -0.47727207669491456, 0.0171080426595288, 0.8173125569729787, -1.8128364003611475], \"chunks\": 2000, \"d\": 10, \"max_iter\": 1, \"n\": 8000000, \"pointwise_loss\": 405024480.2475509, \"rel_error\": 0.5995344874321145, \"seed\": 0, \"sigma\": 7.0, \"wall_time\": 88.51962184906006}, {\"approx_grad\": false, \"beta\": [-1.4148599575143694, 0.9848581358758047, 1.0303035090932562, -0.5169708152326576, -0.8651982685784295, 1.5001343392240916, -0.47727207669491456, 0.0171080426595288, 0.8173125569729787, -1.8128364003611475], \"chunks\": 2000, \"d\": 10, \"max_iter\": 1, \"n\": 8000000, \"pointwise_loss\": 405024480.2475509, \"rel_error\": 0.5995344874321145, \"seed\": 1, \"sigma\": 7.0, \"wall_time\": 72.5514748096466}, {\"approx_grad\": true, \"beta\": [-0.594331803280871, 0.43971841399477063, 0.5008066717190522, -0.24237453836988399, -0.4465450331221631, 0.654942040865184, -0.02191295874530111, 0.17804981679861503, 0.3397522624591079, -0.8989588947159972], \"chunks\": 2000, \"d\": 10, \"max_iter\": 2, \"n\": 8000000, \"pointwise_loss\": 395448737.1142796, \"rel_error\": 0.30623972828770346, \"seed\": 0, \"sigma\": 7.0, \"wall_time\": 10.498352766036987}, {\"approx_grad\": true, \"beta\": [-0.8128070379894679, 0.38924159720017654, 0.4144969044449924, -0.19841532859246383, -0.4710042101673059, 0.734437493524168, -0.2930097249427355, 0.034101664921965966, 0.5329805523482001, -0.9469222291091087], \"chunks\": 2000, \"d\": 10, \"max_iter\": 2, \"n\": 8000000, \"pointwise_loss\": 393745566.9921506, \"rel_error\": 0.2138264606829495, \"seed\": 1, \"sigma\": 7.0, \"wall_time\": 10.346577882766724}, {\"approx_grad\": true, \"beta\": [-0.6790975376476712, 0.5450407181605879, 0.5322319335883364, -0.14396416217709798, -0.5102201147023271, 0.7565770006251672, -0.11488155620503593, -0.03861813027138468, 0.3069642283910206, -0.9001416737560651], \"chunks\": 2000, \"d\": 10, \"max_iter\": 2, \"n\": 8000000, \"pointwise_loss\": 394154971.6913496, \"rel_error\": 0.23966802233459264, \"seed\": 2, \"sigma\": 7.0, \"wall_time\": 10.392645120620728}, {\"approx_grad\": true, \"beta\": [-0.811140159228611, 0.6276138384907568, 0.473085872816493, -0.1838171446679991, -0.40054954722887764, 0.7558401907634041, -0.42296970180408683, 0.06692302612485297, 0.24404777524362825, -0.9590326212179863], \"chunks\": 2000, \"d\": 10, \"max_iter\": 2, \"n\": 8000000, \"pointwise_loss\": 393915237.67307943, \"rel_error\": 0.2250417472016826, \"seed\": 3, \"sigma\": 7.0, \"wall_time\": 10.574932098388672}, {\"approx_grad\": true, \"beta\": [-0.8095071525177628, 0.39580708437907686, 0.5526693122538857, -0.30895738289937397, -0.4200459055558351, 0.6794734103735167, -0.14089336983588296, -0.042578922137064476, 0.37934092991144963, -0.994111196953386], \"chunks\": 2000, \"d\": 10, \"max_iter\": 2, \"n\": 8000000, \"pointwise_loss\": 393761112.8971651, \"rel_error\": 0.21548270480883275, \"seed\": 4, \"sigma\": 7.0, \"wall_time\": 11.583020210266113}, {\"approx_grad\": false, \"beta\": [-0.8838851099696825, 0.6156234649452773, 0.642426897118965, -0.32293019293604086, -0.541623528391047, 0.9373094310885044, -0.2978533442281764, 0.011540620405336708, 0.5121800257134298, -1.1326823717915948], \"chunks\": 2000, \"d\": 10, \"max_iter\": 2, \"n\": 8000000, \"pointwise_loss\": 392098738.7839557, \"rel_error\": 0.0031580153358463157, \"seed\": 0, \"sigma\": 7.0, \"wall_time\": 111.78914093971252}, {\"approx_grad\": false, \"beta\": [-0.8838851099696825, 0.6156234649452773, 0.642426897118965, -0.32293019293604086, -0.541623528391047, 0.9373094310885044, -0.2978533442281764, 0.011540620405336708, 0.5121800257134298, -1.1326823717915948], \"chunks\": 2000, \"d\": 10, \"max_iter\": 2, \"n\": 8000000, \"pointwise_loss\": 392098738.7839557, \"rel_error\": 0.0031580153358463157, \"seed\": 1, \"sigma\": 7.0, \"wall_time\": 115.92996788024902}, {\"approx_grad\": true, \"beta\": [-0.9046560108081807, 0.6326613755118812, 0.5504089312086718, -0.29349505832691863, -0.6703244988401252, 0.997889006713227, -0.2968343677535923, -0.00451966296877129, 0.6390163185945504, -1.1380061562065575], \"chunks\": 2000, \"d\": 10, \"max_iter\": 3, \"n\": 8000000, \"pointwise_loss\": 392474274.2668188, \"rel_error\": 0.10178542461115976, \"seed\": 0, \"sigma\": 7.0, \"wall_time\": 16.024066925048828}, {\"approx_grad\": true, \"beta\": [-0.9083172935353033, 0.7288582671124717, 0.5878503446288098, -0.2563745573662875, -0.5116395655415267, 0.8216666253332333, -0.2935442589357671, 0.006857256737447226, 0.36823156748019614, -1.066219790184447], \"chunks\": 2000, \"d\": 10, \"max_iter\": 3, \"n\": 8000000, \"pointwise_loss\": 392580710.7184185, \"rel_error\": 0.11609047177517823, \"seed\": 1, \"sigma\": 7.0, \"wall_time\": 17.255850076675415}, {\"approx_grad\": true, \"beta\": [-0.8389011499255122, 0.5948283477142385, 0.7238175294286159, -0.2997997038890462, -0.6591657691232473, 0.8288540320772136, -0.3125267990645646, 0.1449823186819636, 0.5704160565098788, -1.2720900725201265], \"chunks\": 2000, \"d\": 10, \"max_iter\": 3, \"n\": 8000000, \"pointwise_loss\": 392708106.3297756, \"rel_error\": 0.1317116829867297, \"seed\": 2, \"sigma\": 7.0, \"wall_time\": 15.214453935623169}, {\"approx_grad\": true, \"beta\": [-0.9414078608218889, 0.7179276231792979, 0.5555960772034156, -0.49420754272047746, -0.40115295600901557, 0.8294078054465174, -0.2042147327239715, -0.15308573285461974, 0.38000894740318536, -1.05377066222109], \"chunks\": 2000, \"d\": 10, \"max_iter\": 3, \"n\": 8000000, \"pointwise_loss\": 393229638.7225437, \"rel_error\": 0.17764704001228399, \"seed\": 3, \"sigma\": 7.0, \"wall_time\": 18.07921314239502}, {\"approx_grad\": true, \"beta\": [-0.926476804277389, 0.6474259797624344, 0.46509357607425766, -0.31031412639931977, -0.2912526589717621, 0.7152394416847999, -0.22762164094017698, 0.17662110468719983, 0.4446096993058229, -1.1348464288890303], \"chunks\": 2000, \"d\": 10, \"max_iter\": 3, \"n\": 8000000, \"pointwise_loss\": 393565654.71435046, \"rel_error\": 0.20169795281469685, \"seed\": 4, \"sigma\": 7.0, \"wall_time\": 16.644001722335815}, {\"approx_grad\": false, \"beta\": [-0.8840752372471408, 0.6155811769804287, 0.6431364460452087, -0.32301339020313813, -0.5412142113891235, 0.9374405298154759, -0.2980612954990996, 0.011143191960726204, 0.5115415216339614, -1.132843412924105], \"chunks\": 2000, \"d\": 10, \"max_iter\": 3, \"n\": 8000000, \"pointwise_loss\": 392098732.22278774, \"rel_error\": 0.0031039884258993638, \"seed\": 0, \"sigma\": 7.0, \"wall_time\": 145.35059690475464}, {\"approx_grad\": false, \"beta\": [-0.8840752372471408, 0.6155811769804287, 0.6431364460452087, -0.32301339020313813, -0.5412142113891235, 0.9374405298154759, -0.2980612954990996, 0.011143191960726204, 0.5115415216339614, -1.132843412924105], \"chunks\": 2000, \"d\": 10, \"max_iter\": 3, \"n\": 8000000, \"pointwise_loss\": 392098732.22278774, \"rel_error\": 0.0031039884258993633, \"seed\": 1, \"sigma\": 7.0, \"wall_time\": 147.1575231552124}, {\"approx_grad\": true, \"beta\": [-0.8271726726227692, 0.49951926049589246, 0.704512362421464, -0.2850862602450924, -0.6863328958031757, 0.8601722207246763, -0.508975427468857, -0.06062527242346647, 0.5160043265879157, -1.2795063115296572], \"chunks\": 2000, \"d\": 10, \"max_iter\": 4, \"n\": 8000000, \"pointwise_loss\": 393059855.81411487, \"rel_error\": 0.16445766469948245, \"seed\": 0, \"sigma\": 7.0, \"wall_time\": 22.655385971069336}, {\"approx_grad\": true, \"beta\": [-0.9616765072338722, 0.5343757451320267, 0.5564692898481249, -0.20815412080423262, -0.6599511653081934, 1.00051922660928, -0.47053635410121947, 0.01242918279858982, 0.5791592404773261, -1.2908888698588328], \"chunks\": 2000, \"d\": 10, \"max_iter\": 4, \"n\": 8000000, \"pointwise_loss\": 392983624.5057327, \"rel_error\": 0.156090826714074, \"seed\": 1, \"sigma\": 7.0, \"wall_time\": 19.843119859695435}, {\"approx_grad\": true, \"beta\": [-0.8311171789400179, 0.521410421564028, 0.9462671660740188, -0.3456457641986597, -0.29334977323977884, 0.9560068287397671, -0.22130619909174298, -0.0049895114959127105, 0.8629391239905513, -1.1745132638244187], \"chunks\": 2000, \"d\": 10, \"max_iter\": 4, \"n\": 8000000, \"pointwise_loss\": 394477000.17289954, \"rel_error\": 0.25672332004490195, \"seed\": 2, \"sigma\": 7.0, \"wall_time\": 20.78614115715027}, {\"approx_grad\": true, \"beta\": [-0.801139199063629, 0.16532940342659558, 0.6954539332675801, -0.28936401002679035, -0.7217519603155139, 1.091651439217816, -0.43199373306063593, -0.11066348276909367, 0.2855398064677888, -1.355992899941729], \"chunks\": 2000, \"d\": 10, \"max_iter\": 4, \"n\": 8000000, \"pointwise_loss\": 395327809.9076261, \"rel_error\": 0.3000645202606274, \"seed\": 3, \"sigma\": 7.0, \"wall_time\": 23.399425983428955}, {\"approx_grad\": true, \"beta\": [-0.7053576581148432, 0.6241930091984895, 0.5853791964035107, -0.2521215481878079, -0.8336804942073975, 0.7948488717641425, -0.278253920773729, 0.018089168528746924, 0.22836203225370016, -0.9739507392001161], \"chunks\": 2000, \"d\": 10, \"max_iter\": 4, \"n\": 8000000, \"pointwise_loss\": 394115898.34814155, \"rel_error\": 0.23886453574779987, \"seed\": 4, \"sigma\": 7.0, \"wall_time\": 22.683278799057007}, {\"approx_grad\": false, \"beta\": [-0.8840149290810264, 0.6155952392872667, 0.6429144576863511, -0.32298713855219047, -0.5413417425959787, 0.9373981467172503, -0.2979957802729954, 0.01126732427321426, 0.5117417330680154, -1.1327922152299434], \"chunks\": 2000, \"d\": 10, \"max_iter\": 4, \"n\": 8000000, \"pointwise_loss\": 392098731.9225391, \"rel_error\": 0.003110684443249633, \"seed\": 0, \"sigma\": 7.0, \"wall_time\": 189.257798910141}, {\"approx_grad\": false, \"beta\": [-0.8840149290810264, 0.6155952392872667, 0.6429144576863511, -0.32298713855219047, -0.5413417425959787, 0.9373981467172503, -0.2979957802729954, 0.01126732427321426, 0.5117417330680154, -1.1327922152299434], \"chunks\": 2000, \"d\": 10, \"max_iter\": 4, \"n\": 8000000, \"pointwise_loss\": 392098731.9225391, \"rel_error\": 0.003110684443249633, \"seed\": 1, \"sigma\": 7.0, \"wall_time\": 186.81083416938782}, {\"approx_grad\": true, \"beta\": [-0.9407374451959918, 0.5995469772485784, 0.3829421849180021, -0.5000865670359426, -0.5737165925062941, 1.1179590375704331, -0.05387964227593586, -0.062217309307136766, -0.002549105064196923, -0.8238232877423541], \"chunks\": 2000, \"d\": 10, \"max_iter\": 5, \"n\": 8000000, \"pointwise_loss\": 396586220.0473553, \"rel_error\": 0.3538455628010787, \"seed\": 0, \"sigma\": 7.0, \"wall_time\": 25.987020015716553}, {\"approx_grad\": true, \"beta\": [-0.4964414415005384, 0.6718447193669304, 0.6846026851661291, -0.010461025842725123, -0.34100874172096585, 0.9586468990831749, -0.21698443097406914, -0.13017291237255, 0.9125536530331779, -0.6087769509034419], \"chunks\": 2000, \"d\": 10, \"max_iter\": 5, \"n\": 8000000, \"pointwise_loss\": 398139813.413357, \"rel_error\": 0.4096640005375313, \"seed\": 1, \"sigma\": 7.0, \"wall_time\": 22.02472710609436}, {\"approx_grad\": true, \"beta\": [-0.8829626205083301, 1.1212609577680581, 0.5684551025155554, -0.6505691911598741, -0.9372315683339582, 0.842119326116806, -0.42196789947305235, -0.08570587191028647, 0.08368869163034408, -1.2792551410778739], \"chunks\": 2000, \"d\": 10, \"max_iter\": 5, \"n\": 8000000, \"pointwise_loss\": 398210025.845274, \"rel_error\": 0.41423200443526725, \"seed\": 2, \"sigma\": 7.0, \"wall_time\": 22.424404859542847}, {\"approx_grad\": true, \"beta\": [-0.8276390460755914, 0.6254867138609531, 0.3250671199876772, -0.17258921033454458, -0.19899367174226415, 0.802372686454787, -0.29783989269792505, 0.15952477773340476, 0.7608196944681033, -1.0561404959330822], \"chunks\": 2000, \"d\": 10, \"max_iter\": 5, \"n\": 8000000, \"pointwise_loss\": 394918666.8340963, \"rel_error\": 0.2788153092040237, \"seed\": 3, \"sigma\": 7.0, \"wall_time\": 23.654550075531006}, {\"approx_grad\": true, \"beta\": [-1.0663379690958243, 0.5137748168705915, 0.7515449969486085, -0.8673948569046709, -0.5042643587430429, 1.400751405530614, -0.45931404840107765, 0.13862134013387145, 0.3718303171564792, -1.255472840503655], \"chunks\": 2000, \"d\": 10, \"max_iter\": 5, \"n\": 8000000, \"pointwise_loss\": 397256096.00919074, \"rel_error\": 0.37940869669671184, \"seed\": 4, \"sigma\": 7.0, \"wall_time\": 30.06166672706604}, {\"approx_grad\": false, \"beta\": [-0.8840363068073565, 0.6155901299411422, 0.6429925519141578, -0.32299641680865165, -0.5412969800789815, 0.937413323810281, -0.298018917144369, 0.011223695205511832, 0.5116712128875367, -1.132810384813737], \"chunks\": 2000, \"d\": 10, \"max_iter\": 5, \"n\": 8000000, \"pointwise_loss\": 392098731.7830447, \"rel_error\": 0.003107192716169649, \"seed\": 0, \"sigma\": 7.0, \"wall_time\": 201.68372106552124}, {\"approx_grad\": false, \"beta\": [-0.8840363068073565, 0.6155901299411422, 0.6429925519141578, -0.32299641680865165, -0.5412969800789815, 0.937413323810281, -0.298018917144369, 0.011223695205511832, 0.5116712128875367, -1.132810384813737], \"chunks\": 2000, \"d\": 10, \"max_iter\": 5, \"n\": 8000000, \"pointwise_loss\": 392098731.7830447, \"rel_error\": 0.003107192716169649, \"seed\": 1, \"sigma\": 7.0, \"wall_time\": 204.56695413589478}, {\"approx_grad\": true, \"beta\": [-0.7871260042083935, 0.3961733101780085, 0.6958527709901753, -0.3142060076706436, -0.6376767465113462, 0.960881014904799, -0.31250840955083714, 0.08209674189561741, 0.515565851142813, -1.2308461475839017], \"chunks\": 2000, \"d\": 10, \"max_iter\": 6, \"n\": 8000000, \"pointwise_loss\": 392779349.549917, \"rel_error\": 0.13854581491536738, \"seed\": 0, \"sigma\": 7.0, \"wall_time\": 36.94024991989136}, {\"approx_grad\": true, \"beta\": [-0.6031158370785182, 0.6709487169693519, 0.7702163302750972, -0.4654554489878465, -0.7768143279733025, 0.9003608368757029, -0.37337880686709085, -0.07810376917934554, 0.5054273482702586, -1.1163010862552663], \"chunks\": 2000, \"d\": 10, \"max_iter\": 6, \"n\": 8000000, \"pointwise_loss\": 393612977.4438118, \"rel_error\": 0.20778431721208035, \"seed\": 1, \"sigma\": 7.0, \"wall_time\": 28.863558053970337}, {\"approx_grad\": true, \"beta\": [-0.8664624880186267, 0.4366107494098207, 0.7988147992751405, -0.2926785272602952, -0.5428020755237238, 0.8156204506426754, -0.19816820095111792, -0.034020062784066066, 0.37078322939922836, -1.137769387721782], \"chunks\": 2000, \"d\": 10, \"max_iter\": 6, \"n\": 8000000, \"pointwise_loss\": 392932772.7201239, \"rel_error\": 0.15320684314729843, \"seed\": 2, \"sigma\": 7.0, \"wall_time\": 28.86776614189148}, {\"approx_grad\": true, \"beta\": [-0.9508191916738734, 0.5803069168850657, 0.6081637705377964, -0.4316316463999406, -0.5012948159484267, 0.968552261019256, -0.24577223705663653, 0.06299812960488932, 0.2590370207057818, -1.252127732722464], \"chunks\": 2000, \"d\": 10, \"max_iter\": 6, \"n\": 8000000, \"pointwise_loss\": 392936926.844349, \"rel_error\": 0.15315234339841183, \"seed\": 3, \"sigma\": 7.0, \"wall_time\": 27.614052772521973}, {\"approx_grad\": true, \"beta\": [-0.8091028799637264, 0.45004047894098886, 0.7061596317906195, -0.3256998918927045, -0.49755334068387674, 0.9253875802460394, -0.11361957804608215, 0.24122491463255788, 0.5365126582799239, -1.1292519285384686], \"chunks\": 2000, \"d\": 10, \"max_iter\": 6, \"n\": 8000000, \"pointwise_loss\": 393111480.63514334, \"rel_error\": 0.1684587208741917, \"seed\": 4, \"sigma\": 7.0, \"wall_time\": 26.95098090171814}, {\"approx_grad\": false, \"beta\": [-0.8840368556424033, 0.6155900530326254, 0.6429948167902779, -0.3229966670664642, -0.541295637133796, 0.937413646700167, -0.2980195490623621, 0.011222412409941856, 0.5116692056516307, -1.132810842239684], \"chunks\": 2000, \"d\": 10, \"max_iter\": 6, \"n\": 8000000, \"pointwise_loss\": 392098731.78297484, \"rel_error\": 0.00310712691370173, \"seed\": 0, \"sigma\": 7.0, \"wall_time\": 229.23673105239868}, {\"approx_grad\": false, \"beta\": [-0.8840368556424033, 0.6155900530326254, 0.6429948167902779, -0.3229966670664642, -0.541295637133796, 0.937413646700167, -0.2980195490623621, 0.011222412409941856, 0.5116692056516307, -1.132810842239684], \"chunks\": 2000, \"d\": 10, \"max_iter\": 6, \"n\": 8000000, \"pointwise_loss\": 392098731.78297484, \"rel_error\": 0.00310712691370173, \"seed\": 1, \"sigma\": 7.0, \"wall_time\": 227.5106761455536}, {\"approx_grad\": true, \"beta\": [-0.9289146412636535, 0.589207394081429, 0.7012400145186803, -0.4208002014380348, -0.43386871861124676, 0.933089665396426, -0.4005966465091005, 0.06483459323652249, 0.5673006082227218, -1.225626019604088], \"chunks\": 2000, \"d\": 10, \"max_iter\": 10, \"n\": 8000000, \"pointwise_loss\": 392517421.8296241, \"rel_error\": 0.10793165451244967, \"seed\": 0, \"sigma\": 7.0, \"wall_time\": 52.241666078567505}, {\"approx_grad\": true, \"beta\": [-0.9743804887017291, 0.5410402380568015, 0.7704433890334813, -0.35185475181030945, -0.46356642753222593, 0.9481604622933278, -0.40504543584553954, 0.06890603470985716, 0.4346690052333042, -1.0354939209957497], \"chunks\": 2000, \"d\": 10, \"max_iter\": 10, \"n\": 8000000, \"pointwise_loss\": 392636046.76928383, \"rel_error\": 0.12258435738129043, \"seed\": 1, \"sigma\": 7.0, \"wall_time\": 49.099812269210815}, {\"approx_grad\": true, \"beta\": [-0.9691976834229626, 0.6475594312600076, 0.7508500065046348, -0.32858753685093967, -0.5462316324992845, 0.8922110086481335, -0.29444047752987906, -0.007789285206273814, 0.5428055414686452, -1.0599010679159921], \"chunks\": 2000, \"d\": 10, \"max_iter\": 10, \"n\": 8000000, \"pointwise_loss\": 392327909.8895484, \"rel_error\": 0.08030740531555673, \"seed\": 2, \"sigma\": 7.0, \"wall_time\": 53.91449022293091}, {\"approx_grad\": true, \"beta\": [-0.910264983109954, 0.5498026119249138, 0.7398351291370956, -0.29663314818807934, -0.6484180771824207, 0.9684016736573872, -0.10330011565812211, 0.05913743859741222, 0.7403852268345604, -1.1810385316788785], \"chunks\": 2000, \"d\": 10, \"max_iter\": 10, \"n\": 8000000, \"pointwise_loss\": 393077610.3466072, \"rel_error\": 0.16494774361606168, \"seed\": 3, \"sigma\": 7.0, \"wall_time\": 56.24317216873169}, {\"approx_grad\": true, \"beta\": [-0.8675186275412994, 0.4551384495672032, 0.604807568567115, -0.2831505725214636, -0.5860885866345346, 0.8674112928558778, -0.416804818232409, -0.1134532111679631, 0.38210984739072146, -1.075781297833849], \"chunks\": 2000, \"d\": 10, \"max_iter\": 10, \"n\": 8000000, \"pointwise_loss\": 392784311.2327362, \"rel_error\": 0.1385175667205294, \"seed\": 4, \"sigma\": 7.0, \"wall_time\": 62.55457830429077}, {\"approx_grad\": false, \"beta\": [-0.8840367015486226, 0.6155900769031977, 0.6429941913527001, -0.3229965972154067, -0.5412960062534135, 0.9374135532285174, -0.2980193731251264, 0.011222765961551154, 0.511669761456385, -1.1328107133068281], \"chunks\": 2000, \"d\": 10, \"max_iter\": 10, \"n\": 8000000, \"pointwise_loss\": 392098731.7829722, \"rel_error\": 0.003107145577421732, \"seed\": 0, \"sigma\": 7.0, \"wall_time\": 301.74830627441406}, {\"approx_grad\": false, \"beta\": [-0.8840367015486226, 0.6155900769031977, 0.6429941913527001, -0.3229965972154067, -0.5412960062534135, 0.9374135532285174, -0.2980193731251264, 0.011222765961551154, 0.511669761456385, -1.1328107133068281], \"chunks\": 2000, \"d\": 10, \"max_iter\": 10, \"n\": 8000000, \"pointwise_loss\": 392098731.7829722, \"rel_error\": 0.003107145577421732, \"seed\": 1, \"sigma\": 7.0, \"wall_time\": 283.68756103515625}, {\"approx_grad\": true, \"beta\": [-0.879226020902641, 0.7785837679404635, 0.5854598664790497, -0.2652160310122725, -0.5033963847997842, 0.9101222961481134, -0.3230869875520656, 0.0726667810267524, 0.6090404727113902, -1.015491282425678], \"chunks\": 2000, \"d\": 10, \"max_iter\": 15, \"n\": 8000000, \"pointwise_loss\": 392603554.26582336, \"rel_error\": 0.1182117141794517, \"seed\": 0, \"sigma\": 7.0, \"wall_time\": 88.87447690963745}, {\"approx_grad\": true, \"beta\": [-0.9264615710300912, 0.5793376114442538, 0.5459787720950554, -0.3485187459810788, -0.737088753955497, 1.0098549802053038, -0.2697041886515856, -0.15363978801505496, 0.38003248055376626, -1.2006135031985268], \"chunks\": 2000, \"d\": 10, \"max_iter\": 15, \"n\": 8000000, \"pointwise_loss\": 392952025.853789, \"rel_error\": 0.15435494003660002, \"seed\": 1, \"sigma\": 7.0, \"wall_time\": 82.69940185546875}, {\"approx_grad\": true, \"beta\": [-0.9206027132484186, 0.739878760663106, 0.7270095070520408, -0.38870409165038816, -0.611151472242097, 0.715224504550175, -0.441189414563162, 0.11140238970762492, 0.5215692165038381, -1.0797497203817892], \"chunks\": 2000, \"d\": 10, \"max_iter\": 15, \"n\": 8000000, \"pointwise_loss\": 393025962.4803908, \"rel_error\": 0.1625869001127405, \"seed\": 2, \"sigma\": 7.0, \"wall_time\": 92.84537315368652}, {\"approx_grad\": true, \"beta\": [-0.787874281047773, 0.6799359185856565, 0.5416287983104555, -0.37731162844352173, -0.5493588312272013, 0.8054093051255071, -0.41572109145020375, -0.10753865241218799, 0.5014026377711878, -1.027600255568214], \"chunks\": 2000, \"d\": 10, \"max_iter\": 15, \"n\": 8000000, \"pointwise_loss\": 392764542.0845498, \"rel_error\": 0.13723331555685436, \"seed\": 3, \"sigma\": 7.0, \"wall_time\": 90.04433107376099}, {\"approx_grad\": true, \"beta\": [-0.8744565169107531, 0.604436206417696, 0.46990150193534835, -0.2077964962204974, -0.5355731471077425, 0.7940482312493866, -0.2473564612549025, -0.13633621899376802, 0.5697639944472284, -1.1292764001824989], \"chunks\": 2000, \"d\": 10, \"max_iter\": 15, \"n\": 8000000, \"pointwise_loss\": 392833172.4833296, \"rel_error\": 0.14196395307157753, \"seed\": 4, \"sigma\": 7.0, \"wall_time\": 80.92051577568054}, {\"approx_grad\": false, \"beta\": [-0.8840367015486226, 0.6155900769031977, 0.6429941913527001, -0.3229965972154067, -0.5412960062534135, 0.9374135532285174, -0.2980193731251264, 0.011222765961551154, 0.511669761456385, -1.1328107133068281], \"chunks\": 2000, \"d\": 10, \"max_iter\": 15, \"n\": 8000000, \"pointwise_loss\": 392098731.7829722, \"rel_error\": 0.003107145577421732, \"seed\": 0, \"sigma\": 7.0, \"wall_time\": 265.0206460952759}, {\"approx_grad\": false, \"beta\": [-0.8840367015486226, 0.6155900769031977, 0.6429941913527001, -0.3229965972154067, -0.5412960062534135, 0.9374135532285174, -0.2980193731251264, 0.011222765961551154, 0.511669761456385, -1.1328107133068281], \"chunks\": 2000, \"d\": 10, \"max_iter\": 15, \"n\": 8000000, \"pointwise_loss\": 392098731.7829722, \"rel_error\": 0.003107145577421732, \"seed\": 1, \"sigma\": 7.0, \"wall_time\": 259.3626790046692}, {\"approx_grad\": true, \"beta\": [-0.9042011637589843, 0.7029405553155955, 0.7110057308292758, -0.3704724339708527, -0.43615660607115325, 0.9149258394399113, -0.2950345455604007, 0.12055345649464859, 0.5209865340313132, -0.9256352343235037], \"chunks\": 2000, \"d\": 10, \"max_iter\": 20, \"n\": 8000000, \"pointwise_loss\": 392750046.25591063, \"rel_error\": 0.13527978657110445, \"seed\": 0, \"sigma\": 7.0, \"wall_time\": 108.67117500305176}, {\"approx_grad\": true, \"beta\": [-0.9719847430972252, 0.561696376565175, 0.7420408719990494, -0.34829215067329033, -0.46395348346167564, 1.0175211166551112, -0.5115341156758816, 0.1059482707651871, 0.46804927489393644, -1.2154401929156522], \"chunks\": 2000, \"d\": 10, \"max_iter\": 20, \"n\": 8000000, \"pointwise_loss\": 392873611.206111, \"rel_error\": 0.14661583431505454, \"seed\": 1, \"sigma\": 7.0, \"wall_time\": 117.8687379360199}, {\"approx_grad\": true, \"beta\": [-0.7238922364851894, 0.48440091636092486, 0.5483989511032223, -0.28848806460103726, -0.5291904133328238, 0.966457907630273, -0.4282109825815653, 0.09406156210969836, 0.5318905979292694, -1.1103801072529718], \"chunks\": 2000, \"d\": 10, \"max_iter\": 20, \"n\": 8000000, \"pointwise_loss\": 392728243.39322346, \"rel_error\": 0.1327308431752905, \"seed\": 2, \"sigma\": 7.0, \"wall_time\": 114.08178377151489}, {\"approx_grad\": true, \"beta\": [-0.9438060751341817, 0.5639914647299185, 0.5472097964866391, -0.42914470178028863, -0.4964936943584284, 0.8407517247734084, -0.276133337484286, -0.0008974746619392729, 0.6534416631567372, -1.1682447312995101], \"chunks\": 2000, \"d\": 10, \"max_iter\": 20, \"n\": 8000000, \"pointwise_loss\": 392578910.63095534, \"rel_error\": 0.11538053946876133, \"seed\": 3, \"sigma\": 7.0, \"wall_time\": 117.60523080825806}, {\"approx_grad\": true, \"beta\": [-0.9221378142827324, 0.7839502847556458, 0.48953742742167167, -0.3521669540471849, -0.6727637426480056, 0.9333889365145457, -0.2866852634483545, 0.05098493792076283, 0.5652542241558753, -1.1840084441981549], \"chunks\": 2000, \"d\": 10, \"max_iter\": 20, \"n\": 8000000, \"pointwise_loss\": 392728530.95368534, \"rel_error\": 0.1327486638410105, \"seed\": 4, \"sigma\": 7.0, \"wall_time\": 118.73969268798828}, {\"approx_grad\": false, \"beta\": [-0.8840367015486226, 0.6155900769031977, 0.6429941913527001, -0.3229965972154067, -0.5412960062534135, 0.9374135532285174, -0.2980193731251264, 0.011222765961551154, 0.511669761456385, -1.1328107133068281], \"chunks\": 2000, \"d\": 10, \"max_iter\": 20, \"n\": 8000000, \"pointwise_loss\": 392098731.7829722, \"rel_error\": 0.003107145577421732, \"seed\": 0, \"sigma\": 7.0, \"wall_time\": 267.0480582714081}, {\"approx_grad\": false, \"beta\": [-0.8840367015486226, 0.6155900769031977, 0.6429941913527001, -0.3229965972154067, -0.5412960062534135, 0.9374135532285174, -0.2980193731251264, 0.011222765961551154, 0.511669761456385, -1.1328107133068281], \"chunks\": 2000, \"d\": 10, \"max_iter\": 20, \"n\": 8000000, \"pointwise_loss\": 392098731.7829722, \"rel_error\": 0.003107145577421732, \"seed\": 1, \"sigma\": 7.0, \"wall_time\": 262.2125267982483}, {\"approx_grad\": true, \"beta\": [-0.8829139465236511, 0.625917698868711, 0.5610035845124403, -0.29290341592114466, -0.5451773218871103, 0.9795265198140615, -0.2515854796694451, 3.103563452222811e-05, 0.5432008109787038, -1.100514723843423], \"chunks\": 2000, \"d\": 10, \"max_iter\": 30, \"n\": 8000000, \"pointwise_loss\": 392209513.1652502, \"rel_error\": 0.05404568835450826, \"seed\": 0, \"sigma\": 7.0, \"wall_time\": 195.57359433174133}, {\"approx_grad\": true, \"beta\": [-0.8635894856734007, 0.6248311725057915, 0.630928540527987, -0.3554994869075913, -0.5729866601651461, 0.950780176643394, -0.33218680175151266, -0.031223046549164638, 0.5121029562335224, -1.070479358507842], \"chunks\": 2000, \"d\": 10, \"max_iter\": 30, \"n\": 8000000, \"pointwise_loss\": 392176658.7978555, \"rel_error\": 0.04775019218448381, \"seed\": 1, \"sigma\": 7.0, \"wall_time\": 185.2518310546875}, {\"approx_grad\": true, \"beta\": [-0.854022071521578, 0.6031780361923385, 0.6422737722072802, -0.36969844012775666, -0.6124105215693041, 0.8853708036980525, -0.2445671723441642, 0.012866940525705072, 0.6060871681055311, -1.1131080448188513], \"chunks\": 2000, \"d\": 10, \"max_iter\": 30, \"n\": 8000000, \"pointwise_loss\": 392284167.6319425, \"rel_error\": 0.07326080170913572, \"seed\": 2, \"sigma\": 7.0, \"wall_time\": 195.97774267196655}, {\"approx_grad\": true, \"beta\": [-0.8827942112640698, 0.6166268364806528, 0.5828977133369241, -0.2929787420823192, -0.6212947554007182, 0.9072455211672876, -0.2748304348371266, 0.060147878868150505, 0.5433172421187823, -1.1121212867854486], \"chunks\": 2000, \"d\": 10, \"max_iter\": 30, \"n\": 8000000, \"pointwise_loss\": 392228262.4553793, \"rel_error\": 0.06074344122218579, \"seed\": 3, \"sigma\": 7.0, \"wall_time\": 206.70327305793762}, {\"approx_grad\": true, \"beta\": [-0.8629285267307552, 0.5671430205245978, 0.5939363055152163, -0.3136346831101807, -0.48504145478632693, 0.9055869404266276, -0.3386675928486884, 0.024980442735857315, 0.49453810816927624, -1.0657688750768044], \"chunks\": 2000, \"d\": 10, \"max_iter\": 30, \"n\": 8000000, \"pointwise_loss\": 392227507.05134, \"rel_error\": 0.059750956746531476, \"seed\": 4, \"sigma\": 7.0, \"wall_time\": 202.11913299560547}, {\"approx_grad\": false, \"beta\": [-0.8840367015486226, 0.6155900769031977, 0.6429941913527001, -0.3229965972154067, -0.5412960062534135, 0.9374135532285174, -0.2980193731251264, 0.011222765961551154, 0.511669761456385, -1.1328107133068281], \"chunks\": 2000, \"d\": 10, \"max_iter\": 30, \"n\": 8000000, \"pointwise_loss\": 392098731.7829722, \"rel_error\": 0.003107145577421732, \"seed\": 0, \"sigma\": 7.0, \"wall_time\": 256.4530289173126}, {\"approx_grad\": false, \"beta\": [-0.8840367015486226, 0.6155900769031977, 0.6429941913527001, -0.3229965972154067, -0.5412960062534135, 0.9374135532285174, -0.2980193731251264, 0.011222765961551154, 0.511669761456385, -1.1328107133068281], \"chunks\": 2000, \"d\": 10, \"max_iter\": 30, \"n\": 8000000, \"pointwise_loss\": 392098731.7829722, \"rel_error\": 0.003107145577421732, \"seed\": 1, \"sigma\": 7.0, \"wall_time\": 271.3524467945099}, {\"approx_grad\": true, \"beta\": [-0.9144115552180049, 0.603015498404459, 0.5713064287707974, -0.34545575275615004, -0.6424885972652146, 0.8931970146541598, -0.2331797791375983, -0.02907522461562066, 0.46925773932688525, -1.1219401556826756], \"chunks\": 2000, \"d\": 10, \"max_iter\": 40, \"n\": 8000000, \"pointwise_loss\": 392312074.65794194, \"rel_error\": 0.0779374005493023, \"seed\": 0, \"sigma\": 7.0, \"wall_time\": 264.6048061847687}, {\"approx_grad\": true, \"beta\": [-0.8973887214484578, 0.625628029190906, 0.6978309024390227, -0.32447667279242576, -0.5497618115288487, 0.9206360070908075, -0.2452788563571229, -0.02033314724967066, 0.4922056424940327, -1.0984816585576087], \"chunks\": 2000, \"d\": 10, \"max_iter\": 40, \"n\": 8000000, \"pointwise_loss\": 392170503.4222042, \"rel_error\": 0.04541517197333465, \"seed\": 1, \"sigma\": 7.0, \"wall_time\": 269.0821568965912}, {\"approx_grad\": true, \"beta\": [-0.8823393235027647, 0.5492224550551256, 0.7019713370961441, -0.3435038812509715, -0.5853327207643598, 0.9835950273865081, -0.27844770423239806, -0.055921215247640724, 0.4887343295786643, -1.1194613758252137], \"chunks\": 2000, \"d\": 10, \"max_iter\": 40, \"n\": 8000000, \"pointwise_loss\": 392242635.91384065, \"rel_error\": 0.06373818373096356, \"seed\": 2, \"sigma\": 7.0, \"wall_time\": 256.9263689517975}, {\"approx_grad\": true, \"beta\": [-0.8279929314291558, 0.5785406092292175, 0.6888273335787916, -0.354480652670551, -0.5278944818518453, 0.9341898266525525, -0.28833210613646065, 0.12581324637761054, 0.47795866600785925, -1.159984780697032], \"chunks\": 2000, \"d\": 10, \"max_iter\": 40, \"n\": 8000000, \"pointwise_loss\": 392281900.7215971, \"rel_error\": 0.0729637904656608, \"seed\": 3, \"sigma\": 7.0, \"wall_time\": 243.45107984542847}, {\"approx_grad\": true, \"beta\": [-0.9078119382540301, 0.5449053559924391, 0.6857141065778634, -0.33039721408797773, -0.546340031592623, 0.9589346749302237, -0.2378234765841817, 0.01719609554537357, 0.4794984490149854, -1.1048310139934197], \"chunks\": 2000, \"d\": 10, \"max_iter\": 40, \"n\": 8000000, \"pointwise_loss\": 392205987.4079259, \"rel_error\": 0.05482705526360315, \"seed\": 4, \"sigma\": 7.0, \"wall_time\": 247.5673270225525}, {\"approx_grad\": false, \"beta\": [-0.8840367015486226, 0.6155900769031977, 0.6429941913527001, -0.3229965972154067, -0.5412960062534135, 0.9374135532285174, -0.2980193731251264, 0.011222765961551154, 0.511669761456385, -1.1328107133068281], \"chunks\": 2000, \"d\": 10, \"max_iter\": 40, \"n\": 8000000, \"pointwise_loss\": 392098731.7829722, \"rel_error\": 0.003107145577421732, \"seed\": 0, \"sigma\": 7.0, \"wall_time\": 261.3412778377533}, {\"approx_grad\": false, \"beta\": [-0.8840367015486226, 0.6155900769031977, 0.6429941913527001, -0.3229965972154067, -0.5412960062534135, 0.9374135532285174, -0.2980193731251264, 0.011222765961551154, 0.511669761456385, -1.1328107133068281], \"chunks\": 2000, \"d\": 10, \"max_iter\": 40, \"n\": 8000000, \"pointwise_loss\": 392098731.7829722, \"rel_error\": 0.003107145577421732, \"seed\": 1, \"sigma\": 7.0, \"wall_time\": 243.74875688552856}, {\"approx_grad\": true, \"beta\": [-0.8854073087268456, 0.6172178631773021, 0.6450145855969107, -0.32189737380866645, -0.5447108173125104, 0.9348775941459327, -0.29811605878941355, 0.013981237885834294, 0.5143839539544327, -1.1370906424701517], \"chunks\": 2000, \"d\": 10, \"max_iter\": 80, \"n\": 8000000, \"pointwise_loss\": 392099221.5401707, \"rel_error\": 0.005300432688336602, \"seed\": 0, \"sigma\": 7.0, \"wall_time\": 925.5044710636139}, {\"approx_grad\": true, \"beta\": [-0.8818837818078422, 0.613442176739526, 0.6419890655824715, -0.32382572208301896, -0.5415035086126159, 0.9344285595975396, -0.29847295104227245, 0.00894236421481264, 0.5093510115674734, -1.1237886908468033], \"chunks\": 2000, \"d\": 10, \"max_iter\": 80, \"n\": 8000000, \"pointwise_loss\": 392099628.5426975, \"rel_error\": 0.006657967894667196, \"seed\": 1, \"sigma\": 7.0, \"wall_time\": 1053.5675048828125}, {\"approx_grad\": true, \"beta\": [-0.8841524006885965, 0.6192790187917091, 0.6435924079837052, -0.3224913282377281, -0.5405742823931305, 0.9331562752015514, -0.303804688437872, 0.012429471065645424, 0.5122031575325743, -1.1330440943007851], \"chunks\": 2000, \"d\": 10, \"max_iter\": 80, \"n\": 8000000, \"pointwise_loss\": 392099277.1428425, \"rel_error\": 0.005527006304765571, \"seed\": 2, \"sigma\": 7.0, \"wall_time\": 1059.2264840602875}, {\"approx_grad\": true, \"beta\": [-0.8825335911850323, 0.6098598438138241, 0.6441103657737104, -0.3210903115661359, -0.5390939638365305, 0.933140324813863, -0.2976977255708058, 0.012605030831765042, 0.5152813665905337, -1.1273924354863225], \"chunks\": 2000, \"d\": 10, \"max_iter\": 80, \"n\": 8000000, \"pointwise_loss\": 392099592.01777464, \"rel_error\": 0.005867858009065959, \"seed\": 3, \"sigma\": 7.0, \"wall_time\": 1028.4987251758575}, {\"approx_grad\": true, \"beta\": [-0.8871790409057009, 0.6196366303695788, 0.6431369705966894, -0.3268682376868445, -0.5390371654462093, 0.9393368181590982, -0.294816062643927, 0.010144475140230992, 0.5120191693819195, -1.1346528130151012], \"chunks\": 2000, \"d\": 10, \"max_iter\": 80, \"n\": 8000000, \"pointwise_loss\": 392099251.88777584, \"rel_error\": 0.004409698151563556, \"seed\": 4, \"sigma\": 7.0, \"wall_time\": 1020.5811431407928}, {\"approx_grad\": false, \"beta\": [-0.8840367015486226, 0.6155900769031977, 0.6429941913527001, -0.3229965972154067, -0.5412960062534135, 0.9374135532285174, -0.2980193731251264, 0.011222765961551154, 0.511669761456385, -1.1328107133068281], \"chunks\": 2000, \"d\": 10, \"max_iter\": 80, \"n\": 8000000, \"pointwise_loss\": 392098731.7829722, \"rel_error\": 0.003107145577421732, \"seed\": 0, \"sigma\": 7.0, \"wall_time\": 273.7842438220978}, {\"approx_grad\": false, \"beta\": [-0.8840367015486226, 0.6155900769031977, 0.6429941913527001, -0.3229965972154067, -0.5412960062534135, 0.9374135532285174, -0.2980193731251264, 0.011222765961551154, 0.511669761456385, -1.1328107133068281], \"chunks\": 2000, \"d\": 10, \"max_iter\": 80, \"n\": 8000000, \"pointwise_loss\": 392098731.7829722, \"rel_error\": 0.003107145577421732, \"seed\": 1, \"sigma\": 7.0, \"wall_time\": 267.38317227363586}]}};\n",
"var selector = \"#adad4346-4c33-4fd9-8d7c-22b6667ed6ac\";\n",
"var type = \"vega-lite\";\n",
"\n",
"var output_area = this;\n",
"require(['nbextensions/jupyter-vega/index'], function(vega) {\n",
" vega.render(selector, spec, type, output_area);\n",
"}, function (err) {\n",
" if (err.requireType !== 'scripterror') {\n",
" throw(err);\n",
" }\n",
"});\n"
]
},
"metadata": {
"jupyter-vega": "#adad4346-4c33-4fd9-8d7c-22b6667ed6ac"
},
"output_type": "display_data"
},
{
"data": {
"image/png": "iVBORw0KGgoAAAANSUhEUgAAAqYAAAGVCAYAAAAsfTMBAAAAAXNSR0IArs4c6QAAQABJREFUeAHsvQu8XMV951l1+j70vJIAgRF2JBGM7NjCODF21rAm3pAlXlszznixA5uNYOM4mcEOZDHY8dio/YiT8IjJYDLByeyITWLGMNlJDNhJTDbBI/yZxCYYMEmQGXOFYwESoPfjPvrU/H7nVPU9t2+f0933dve93f2rz+ffVaeqTp063xLip3/VqTJGQQREQAREQAREQAREQAREQAREQAREQAREQAREQAREQAREQAREQAREQAREQAREQAREQAREQAREQAREQAREQAREQAREQAREQAREQAREQAREQAREQAREQAREQAREQAREoD6BUv3studatHghbAvsJOwILIT1SKzM2CjSrBPCViTOgx2AZfNDuWIREAEREAEREAEREAERaIrAEGq5Gnuvv5OCtbaM1yFQjGbLKVIVREAEREAEREAEREAERGBeBG7BXRSX9IwyjMMoOBmYxzIKzk3eNiBm2AZjGT2tDOMwXlPMKoiACIiACIiACIiACIhASwSCR/SSzF3rkL7eX29CHESqz6pG40h9sXo1I2IpXNnGY7DtMIrVIGCvzlyfi7SCCIiACIiACIiACIiACCQEwjQ+vaYUkhSQQZSyQhCW40iz7CHYahhDbd0gcjehjOKU5TQKUApYpsdh9L7eD+O1vKuAoCACIiACIiACIiACIpBOzwcBuR1AwvT8uIdDkcpyrjl9EyysKV3u87Oe1iBMmReEaRCxYUlA7XWEugoiIAIiIAIiIAIiIAI9QIAezW6EC/CQb/kHMf1NGEXjzbBbYTGMgQJzCraVFwhbYA8mqZmfp2eS5qhPD/u49jpTNUlSBOcFPkthcAhswquOD87r6k1BYBNsHKYwOAQ24VXHB+d19aaewG6R6F0CnRamL3g0L2YQ7c2k6UX9Cmy/z6v4+ATig7DN/prRKp+ezOQFL2rIqr0O+SFmeb1Awao/yPXI9Heexry/x7fe22nM61Hp7zyNeX+Pb+3b6fuSWiK6nkOA0/M0Ts+PwsZh4YOnbBk9qF+AUSSyHkUr0/xDRgEd1qgiWZ3KD1P1G5DHunnXvKco8F6FwSKgv7wGa7z5thpzjfngERi8N9Z/5z0+5p32mBLPK2DPw47zwocVPuYfoH2wbNkWXE/A7oJdCnsKFkK4L1wrFgEREAEREAEREAEREIGWCdBjSqsX6CHNK2M+t4fKm4av116refKYtkqs9+vrX9W9P4atvoHGvFVivV9fY977Y9jqG2jMWyW2xOqHqe9udIvrRmn1Aj2keWXM55S/xGM9csoTAREQAREQAREQgaVH4CF0idqNyy2bDt0Upk13ShVFQAREQAREQAREQAR6msBa3/uRVt5CwrQVWqorAiIgAiIgAiIwyAS4vJCH+HAmdxzGj7bDUsRLkKaXkHuzMx6H8YChsBSxUTnrfhG2DZb1NHKf93Gfx+fylEuGTTA+h/2hnuN3Q0wzj0sgGwV6Mvlh+TiMz+a78F7uCU9jGd+lDONz+YzQF16zPPQFycQzyjyWsb3zYArzJKBlAvME18O3aR1SDw/ePLuuMZ8nuB6+TWPew4M3z653esypF2gUcBRhTI/DGCjSQjnFWUiPI83QqDy0F+6jcKSYDdcUnSH9RaQZxmHMo6gM5aEMWbmBwjO0le0r8zZ5C+UhPjtzD58R7tuOdKP2UEWhFQKErjBYBDr9l9dg0eyNt9WY98Y4tbOXGvN20uyNtjo95hSX9AaOwoJoHEeaIQhPljOsgwVRR69qo/IgTFnPwhiC+KOnkoHPDW0yTS9puGbM+s2E0BeKWYZsu5twTVEc2uU1A/PKMHpR+VyKc9a5Hhba+yLSDLXtpbn6bZoAwSoMFoFO/+U1WDR742015r0xTu3spca8nTR7o61OjzlFZxBsIaagZAjijKdYhhDEJvvVqDzUDaKUU+fhGSGP7YZ64V0pFEO9razQRAh9ydYPQnMT7g/CNLwbm6TYDM8Oz2NMYUqPLdPZ9kLdTchvOvClFURABERABERABERABIoJbELxdbA9MB6FPgarF37cZ1JMhnWW389UbFQeRGg4rp23nunvz7b5HPKo477kyxh9PZMuSoZ94X/NV2K7b6tzw5pM3u1I830+CGN9xiF81yd+OmQg/qFMuumkhGnTqFRRBERABERABERABBICr8XvLs8iK96Y9WUYp8if4YUP2S0xG5WHexjf4C9+gJheztDmHqSPwP4axvAAjHn8Ej4sJUAyN/yxL7kcMT2bWRGcd9MqX0CP8CWwz/vr44j/zKdvQszns83wVb4vUtQKAbqfFQaLQJgCGay3Huy31ZgP3vhrzDXm7SRAL+E4jJqBRvEZ0uuQDtPj45l8loev9huVU8yxfq3TkEIvPIcx67EOxWHI53WYfmdedjkBLusG/vcxDmP97Lvw3tAWnxVCyAvP/CIKmA7rWrf761AeYt6n0CIBwlMYLAL6H9ZgjTffVmOuMR88AoP3xt3475xrLfnxT20IwpNCjCKW9bKhUXm2bm2az+OX70Hk1pbXXrP+m2Bc81lrZLQJxv5QTDJQ2FJgUg/VCmNkzQrsA9+vXuBzm+1jvfvrgq1bUZkiIAIiIAIiIAIiIAKGp1UWhREUUuDl1WtUXq/taWRy6r7ZQHH4AdjhnBs+h/wwFb8zU2cP0o2m9bPLEjK3Jkn2k6awQALymC4QYA/ezn8xKgwWAY35YI0331ZjrjHvJoEL8TBOiXNav15oVF7vnk7mbULjX4Rxun4cFrynSCosNgEJ08Uege4/X//D6j7zxX6ixnyxR6D7z9eYd5/5Yj9xqY553tT3YvNacs+vt0ZiyXVSHRIBERABERABERCBpULghz56/yXGuR9N+mPt3z/7m+96sKBv9JTugt0K+3BOPQpXTqGfBdubU2cgsiVMB2KY9ZIiIAIiIAIiIAILJfCqj93/+qhi7sQK0rdWv//BnOvGj9z/jbhkfun7n33Xd+o843eRVyRKeUvwqI7UuX+gshp9eTVQMPSyIiACIiACIiACIlCPwIbyfadFcbKGFKLUfA/fN/12akybt7KMdWru3YZrbkrPjfkv8WXZ7Z+YZqhdUlj2eczfDguBWzmNw5jP9ax952DsljDlvwToyuag1NvPilsZsKzeguGiMtyiIAIiIAIiIAIiIAKdJTB0Ivo45OBGaMKv7vne8XP3/Na26xJDmnksS+rM7sYuf/kA4kdg3KKJInUzbItPvwlxNlAv7YBxWv8C2E4YtRC3n9oH+8+w02Gvh+2H9VXohjClmue6CQ7O12A8veC9sBC4b9bjMJa9DCP8EIrKQh3FIiACIiACIiACItBRAta6d/MBsYtuNPe+t1J9GNJJHjJCnWpZujfoHlx/FUZN819gFKXjMHpaD8LeBcuGczIX30KaAnUc9n4Yw4dhFKTUSzxdaTWsb0I3hOlvelpU9/SccoDu9HnbEBPqRbBQRpHKdFEZihVEQAREQAREQAREoEsEnHkln/T9Z449WvvEap6vU1uO63A2PYt4P6fip2DUQMdg2fCHuKAWoiOP9b4Am4SFwDzaYZ9xaijoh7jTwpQCky7rn4IFd/Mbkf4sjOF22N2wh3mBQJc1w5mwojJO+XPfre2wMEAXIk0XebiGa11BBERABERABERABNpAwJp/Ziuv2rySOmZWqOb5OrMKZ188ict7YcOw4JCbXcOYlch4M6wE4/rUd8J+DUZxSw8r72MZ26BjbxzWN6HTwpTgGH4aRiFJ0UhX9M0who2w7L88XkxyjeFXaUVlPNGAg7UTtgVGcbsLdj2M+Q/AnoJx8BREQAREQAREQAREYEEEnLN/ygYiG3/KXHZP0DeG6SQPZaFOgwfRGTcN49pSah0K0WzYgYvnYVwG+QTs6zCGP4fRw8pvdVj2qzBqn74KQx1+m8TtjWfQa3oljGtIvwy7GnY2jCErTNOcmdM66pWdg0r/4CuOIT4CuwZ2OYzrLXh9FYwLhClMHWx+gX/wsutI5teK7hIBERABERABEehxAtPL488Mn7TvNs6+Y+PZK3abj9yXCFVIjXfj1c6G4tgzvSz+TJ3X3JPJuxbpnd4OIqZjjUL007AQPokEdVNWv9DBx6NAWfYDWAibQ6JfYtvhF9mE9p+BcYqeC3gZ+C+Eb8KGYVxf8UHYHTAG9of/Cvhh2H+H1SvjIHCtBQeGHl8OHP/1UO+a/6JheyFkBznkhZie12r4oY/c96Q1dmjP5372R8zk0ZlFztUaSvQ4gU3o/3iPv4O63xqBTag+3totqt3jBDah/+M9/g7qfusEdrd+S3N3VPcxxfZQNXcU7WNaUzXZ4on6ZMIXUMtktUqoz9lhhtqz6elUzN6fVOqXn057TF/woMIUPS/3+jwOAv+1kFX7q3zZyYIyitIQKGSzYrP2OtQLMcvrBbYx6w8yRCnd7EObPnbfM+Plt7M/Cv1HYNaY99/r6Y3qENCY14HS51ka8z4f4JrX6+j3JX4D/QtbPPmppovJND41Rgj1RCnLagVpqM97s/eHfMVNEjiAejQq/1HYOIzXDNthFIX8g0SRHNahIllYtgHlvI//ymBodJ3Wyv9lW7MCTnE4BHPnlL/C5QIK/Uego3959R+uvngjjXlfDGNLL6ExbwlXX1TWmPf4MHbaY0o8r4BxEe9xXvgQtk24C9eXwvihUgjNlIW6nYy5zMAcP1IZ6eRD1LYIiIAIiIAIiIAIiEBKIHgcO8mDayjWwSg4aZxOz7qnr/D5pyBmf5op24t6bCe4vxtdo2rLYZJ3DI8Ocy2sggiIgAiIgAiIgAiIQIcJdMNjGl4hKzhDXohZlldeVBbub39s8YEVJvijCXlM2w9XLYqACIiACIiACIjAXALdFKZzn76Uc1yyY4CZGp6Wx3Qpj5P6JgIiIAIiIAJdJuDKay7BpO2Ppo+N/t6WDz1Y0IVNKHsOFr7CL6iqIgnT/D8DyVR+ZIe1xjSfkUpEQAREQAREYGAIuPKq1xsX3Ymd9N+arijkqzvjdox9w9j4l2z56HfqwHgGeTfAwuFCdaooKxCQMA0k5sbJx0+luCKP6Vw2yhEBERABERCBgSLgyqtPgxi9Hy+9EfY92J96AO9GDKEa3Y86b7LlI9ktMt/r69yE+Bswfk/zY7C3w9jWHhjDPWlkQn1e81uaP4ZdDuP2mm+B7Yb1dZAwzRle7GM6iX8DmWnr5DHNYaRsERABERABERgcAvbjkAUbIRe/ap48vM3ea5LDd9xl8Ia+buw+lL3DWNQx5toMk6/5NI9K/wfYFbAdsK/D/gz2L2HZI0kpesM1Pa1rYGfBPgDjDkbhxEsk+zN046v8niQHUZp4TCMXyWPakyOoTouACIiACIhAWwnQM4pgbwyiNLlKBKq9MSkyxtfxV+m+7fSKfhUW9nDn9cWwIu/nepRTpJ4N2wsrwxh+Po3691fCNH9sJ1lk5THNJ6QSERABERABERgcAq9MXvXJQ4/OeeWZvLTOnArJdpnM5baZO5nICa/2+cEp9jKunTcW8f6+DhKmOcOLhR2JxxSxpvJzGClbBERABERABAaIwD8n7/q6NW+c884zeWmdORUKM45lSi9DmtdBe1CIlmDUa+fBfg/W10HCNGd48c+TxGMaOxv+1ZJTU9kiIAIiIAIiIAIDQOBP03d0n8K6UorFJKRp9yl/6ev4qzTiOtEzZuXMvvgYLqnH3gRb64ue9fEHEfMwIU7rPw47E9bXQR8/5Q9vIkyNpvLzCalEBERABERABAaGgPsM1ve9O/nI6UfGdrsdma/yHdaCWn5hjzpzAz+I+jzshblF5g+QdxMs+ZAK8UFfh2J0C4wfPLGcgVtO7U5SffwjYZozuJzKTxZ16OOnHELKFgEREAEREIHBIcBtoLCP6bvg3LwTb/1W2P+deftvwLHJfUyzW0WF4juQ+F0YZUVt4AdRkBxmFDZRU0gRSk/qMl9Gsdr3QcI0Z4jDVD7+HIV1Hjk1lS0CIiACIiACIjAIBPwG+he2ePIT0dQTpVlktaI0lPE+Hs0+MEHCNHeoLT5+ciZyTmtMcxmpQAREQAREQAQGj4A/gvTBwXvzzr8xXcQKdQhgH1O/xjSSx7QOH2WJgAiIgAiIgAiIQLsJSJjmELU23S7KuVge0xxGyhYBERABERABERCBdhKQMM2j6az3mFb3EsurqXwREAEREAEREAEREIE2EOjGGlN+bXZaTV+P4jos5l1fUzaN63BsF4u2wrj/1yOwbD4uOxesc1MOPbdGU/mdo6yWRUAEREAEREAERGCGQDeEKbdU2DXzyCR1K34/DKNo3ZfkzP5hPgOFaNhsltc89eAJJjodXIQN9vEtnNPHT51GrfZFQAREQAREQAREICHQDWF6Pp7E0wreAQsfEr2UPH3Gk0rBecTnpVPoxmzDNUXpRbCHYeMwtsPlB422XUCVBYZkKh+P0Qb7CwSp20VABERABERABESgOQLdWGNKQfoF2F7YuLcgQlfimqcc0As67o31GG6H3Q2jKGW4II2S47jWIf0YbDuMIpV2IezqzPW5SM87OOuwXRR1qY4knTdE3SgCIiACIiACIiACLRDohjB9FfrDo7g4LU8BWYaFcDES9IqOw1j2EGw1jGEj7NEklf6E0xTodV0Oo5d1J2wLjAKWywWuhzH/ARiP8QpLApBsLXx86I8/xTsiW+GJCwoiIAIiIAIiIAIiIAIdJtBpYUphSKFIr+jbYO+D7YDR28lwehol57/SI8q6h2EUngxZYZrmGHNOSCAeg+2GXePztiKm9/Uqfz1vYVoycXL014idDn3xTSoSAREQAREQAREQARHoBIFOrzGlF7QEC+e7UjTya/wvwyiKb4bdCgvl65HmFDoFJgO9obUnKzydlKQ//LqfIew1Wnudls78sj95YdbUv3UV7g5gomMHXoFoVlleA8rvKQKbeqq36mw7CGjM20Gxt9rQmPfWeKm3ImA6LUw5Lf/LMArQEMZDAvF22Fdg+31exccUr/SybvbXjFb5dPg4ipf0iGbFZu21v6Ua5XlQ2QY9r9UwZBz7sGp45SoK5Vll1UpK9DoBjWuvj2Dr/deYt86s1+/QmPf6CLbWfzmSWuO15Gp3eip/Em98E6wMoyjkR0v8sn4PjF7S22D8S4PT5ezLnTAG5l0Luw7GP2QU0LtgDHvTqLO/JRtP8AlDblprTDuLWq2LgAiIgAiIgAiIQEKg0x5TijuuK/0SbEfyxFSUnuPTFJ3cx/S4v2bE6XvedxfsUthTsBBWhESn45KpJMIUalnCtNOw1b4IiIAIiIAIiIAIgECnPaaEfI9/DkUlbRMsWb+JmFP49KRS/LGM6ey0yxU+/xTE7Cun1xnoNWXdsDa10TXvaSlExp3kDda60ZZuVGUREAEREAEREAEREIF5Eei0xzR0ims4g6gMedk48U5mMzJp3ld0b6Zq+5LwmIZnSpi2D6taEgEREAEREAEREIFcAt3wmOY+fCkXYI1pIkzhlh1Zyv1U30RABERABERABESgXwhImOaMZBTHyVQ+PvqXxzSHkbJFQAREQAREQAREoJ0EJExzaJaiOPkgy9nqHqk5NZUtAiIgAiIgAiIgAiLQDgISpjkUSy4VpsZFmsrPYaRsERABERABERABEWgnAQnTHJpDNj7GInz2H06VyqmpbBEQAREQAREQAREQgXYQkDDNoThk/FS+s93auSCnJ8oWAREQAREQAREQgcEgIGGaM87D0fRRFsVGwjQHkbJFQAREQAREQAREoK0EJExzcEbxdDqVbyN5THMYKVsEREAEREAEREAE2klAwjSH5qidOsKi2EWlnCrKFgEREAEREAEREAERaCMBCdMcmFhjmkzlV4yVMM1hpGwREAEREAEREAERaCcBCdMcmqvik6nH1JbEKIeRskVABERABERABESgnQQkunJoLosqiTCdcpEY5TBStgiIgAiIgAiIgAi0k4BEVw7Nte7lwyyqGHlMcxApWwREQAREQAREQATaSkDCNAfncjNxdAiy1BlrTPmv9WV+Didli4AIiIAIiIAIiEC7CEiY5pEslU5QmDK88tB+nf6Ux0n5IiACIiACIiACItAmAt3wBMLlaE6r6S+/eD+RyduK9BmwR2AHMvlMFpXVVG3n5fTxYTNtTpoRMzKybgQtZ/vbzgepLREQAREQAREQAREQARDohjB9K56zq4b2rbj+sM+jEF2bKT8P6SeaKMvc0onkyPFhC4+pM+bk9El5TDuBWG2KgAiIgAiIgAiIQIZAN4Tp+Xje47B3wOh5ZHgpjcw2xBSlF8Eeho3DWJdLDN4FyyuDXOx0OJh4TPmUUikO/e70Q9W+CIiACIiACIiACAwsgW6sMaUg/QJsL2zcW7IVE9K3w+6GUZQyXJBG5kzERWXrUP4YbDuMIpV2IezqzPW5SM872LKZDGtMT3cHls+7Id0oAiIgAiIgAiIgAiLQFIFuCNNXoSefh3HKngKyDAthIxKPhgvEL/o0PZRFZRSKnPLfCdsCo7jlcoHrYcx/APYUjOtb5x0gTGPefLY5sHrejehGERABERABERABERCBpgh0WphSGFIoHoS9DfY+2A4YvZ3h2VlhiuwkBG9nvbJzQiXEY7DdsGt83lbEXJ96lb9ekDAt2VSYjkWH+BwFERABERABERABERCBDhLo9BpTekh51nzieURM0civ278MC8KUHs8HYdnwtL8oKmMVft3PMJxGc659djVif/JCEMPV8lJcSeof3PNPr0Hm89UCJfqBwKZ+eAm9Q0sENOYt4eqLyhrzvhhGvcQgEei0MOUU+C/Dbs5AHfdpilV6Ujf7a0arfPok4ryySV+HET2iWbFZe52pmiTzPKhsg57XWWHYxtNofXjz2a/kKVBzymdV1kUvEtCY9uKoLazPGvOF8evFuzXmvThq8+/zHCfT/JvSnYtBIHgtO/XsSTR8E6wMoyjkR0v86n4PjML0Wth1MP5BokjeBWPYCysqSyp1+ieyLtlhfziqrOz0s9S+CIiACIiACIiACAw6gU57TCcAmOtKvwTb4WFTlJ7j03chvhTGD5VCWOETRWWhbkfjyMBjioAp/eDJ7ejz1LgIiIAIiIAIiIAIDDKBTntMyfYeGJ9DwUnbBEsEH2KGK2DMPwXGetkTlvLK6FG1sLB2tdE1qrYeIpcK0yHr5DFtHZ/uEAEREAEREAEREIGWCHTaYxo6wzWcWcEZ8kPMsrzyorJwf0fikveYGufkMe0IYTUqAiIgAiIgAiIgAjMEuuExnXlar6WsmUq6bK022O+1sVN/RUAEREAEREAEeo6AhGnBkGGNaSpMjabyCzCpSAREQAREQAREQATaQkDCtAAjFrFyVwGsZHVcA6sgAiIgAiIgAiIgAiLQQQISpgVwI+NSYRrZZQXVVCQCIiACIiACIiACItAGAhKmhRAdt7vCHvuR1pgWclKhCIiACIiACIiACCycgIRpAUPrbCJMYyOPaQEmFYmACIiACIiACIhAWwhImBZgdNbwaFSeeaqp/AJOKhIBERABERABERCBdhCQMC2gaK1Jp/KdhGkBJhWJgAiIgAiIgAiIQFsISJgWYIydTTb9d8aOFFRTkQiIgAiIgAiIgAiIQBsISJgWQIQwTabyK7Y0WlBNRSIgAiIgAiIgAiIgAm0gIGFaANFF9jiLnTPymBZwUpEIiIAIiIAIiIAItIOAhGkBxThOp/JjU5IwLeCkIhEQAREQAREQARFoBwEJ0wKKU2Yo8ZjG1gwXVFORCIiACIiACIiACIhAGwhImBZArJjoGItjF0mYFnBSkQiIgAiIgAiIgAi0g4CEaQHF495jCoEqYVrASUUiIAIiIAIiIAIi0A4CQ+1opIU2VqPuebCHM/esz6SZnIYdyORtRfoM2CM1+ZkqnUlOmdHUY2qibnPqzAupVREQAREQAREQARFYwgS6LbgOg8VB2DrPxCLe59PZiPkMFKhrk1T6Q1H7ROa6o8kjbsVRPqDiolJHH6TGRUAEREAEREAEREAETDen8m/xvJ/NcD/Npyk4N3s7y+dtQ0xRehGMQnUP7HGfRtT5cMSsTDym06ZUcpcZidPOI9cTREAEREAEREAEBphAt4TpuWB8XR3OK5FHDyq9oOPe9iJmuB12NyxM+1/ATIQzYfS4PgbbDsNR9oldiPjqzDWfuaDgoniKDUxRk75u/fIFNaabRUAEREAEREAEREAECgl0Q5jyGU/BboC9D/ZDsBAuRoJe0XEYBeZDMK5DZdgIezRJpT8v+jT3FKVIpJd1J2wLjAJ2F+x6GPMfgPGZ9LTOOzhnJ3nztOGKhxMr5t2QbhQBERABERABERABEWhIoBvC9MvoBafhb4atr+nR6f6aopUeUYpKrkMN3smsMEV2Es4JCcRjsN2wa3weP5Si9/Uqf70gYVrywnSSwnRqKPTJN61IBERABERABERABESgnQToCuxk2IDG3wm7ErYJ9gYYAwUkBSXF6q2wGMZA4crpc5Yz0Bv6YJKa+Xl6JmmSj5NwHbZzqr3OVE2S9MrmhTlT/0f/6W/OXPGai+ExLZnbvnniR3DjaN7Nyu85Apt6rsfq8EIJaMwXSrD37teY996YqccDTqDTwpQik2tIb/OcOW3PwI+YToH9LOwrsP0whkoamROIed9mf81olU8n0+s+TY9oVmzWXvtq1Yjl9QLboFCeFZa/9uLVbB0nQJlr37Js36/+5cScOrNu0EWvEdB49tqILby/GvOFM+y1FjTmvTZiC+vvHCfTwprT3d0m0OmpfApOfqgU7INIU3BSIB6AUbDyLw1Ok7Mvd8IYmHctjB9M8Q8ZBfQuGMPeNOr8b2m6kn785PhBfqw1pp1HrieIgAiIgAiIgAgMMIFOe0xr0T5bk0HRyX1Mj2fyOX0/AbsLdimMHzGF0FVx6IajSS4yoMfUlCKtMQ2joFgEREAEREAEREAEOkCg0x7T2i7fhwx6T0OgR5Xe02Uwik6ms9MuV/h8Tvuzr5ziZ6DXlHXD2tRG17yn5VCZLvntoqjf5TFtGaBuEAEREAEREAEREIEWCHTbY5rXNXpI8wLFaBCkeXU6kj8yMjk5PT2U7mNqbFe9tR15ITUqAiIgAiIgAiIgAkuYQLc9pksYxdyuxX67qGQq3zgJ07mIlCMCIiACIiACIiACbSMgYVqAcurEaDKVz+2iTGy1xrSAlYpEQAREQAREQAREYKEEJEwLCE6sOTzJ4inHFQ+ayi9ApSIREAEREAEREAERWDABCdMChGteWuU/foLH1GoqvwCVikRABERABERABERgwQQkTAsQPn37O1KPabKNqjymBahUJAIiIAIiIAIiIAILJiBhWojQOmvi2GFnqgk7tLKwqgpFQAREQAREQAREQAQWREDCtAG+knHTrDJhRsKRqA3uULEIiIAIiIAIiIAIiMB8CEiYNqJmTSpMnTymjVCpXAREQAREQAREQAQWQkDCtAG9yMXpllESpg1IqVgEREAEREAEREAEFkZAwrQBP5x7mgjTCTOkDfYbsFKxCIiACIiACIiACCyEgIRpI3o2FaYVU5IwbcRK5SIgAiIgAiIgAiKwAAISpg3gwWOabBk17SRMG6BSsQiIgAiIgAiIgAgsiICEaQN82Csqmcqv2JKOJG3ASsUiIAIiIAIiIAIisBACEqYN6dkJVplykYRpQ1aqIAIiIAIiIAIiIALzJyBh2oBdbFJhijWmyxpUVbEIiIAIiIAIiIAIiMACCHRbmK5GXy+s09+tyLsEtq7FsjrV25vlTHSSLU5LmLYXrFoTAREQAREQAREQgRoCQzXXnb48jAcchGUF6AFcr808+Dykn/DXRWWZWzqXrHiPKYTpiCubyJZN3LmnqWUREAEREAEREAERGFwC3fSY3uIxP5vBvQ1pitKLYPgA3uyBPe7TRWWo0p3gwlf5poT+bdB0fnew6ykiIAIiIAIiIAIDSKAZYcrp9/fCvgC738e8Zn6z4VxUvK5O5duRdzfsYV92gY/PRFxURo/rY7DtMGjHxLhE4OrMNZ+54BAZm2wXNWnoXD6ivUwXTFQNiIAIiIAIiIAIiEB9AkVT+etxy4MwTq3Xhl/0GfRw/gRsHJYXKH6fgt0AY/07YSFsROKOcIH4RZ8eQVxUxmrs107YFlgZtgvG9pn/GzA+k8+mcJ13cMYl20VhH1O0odOf5g1SN4qACIiACIiACIhAAwJ5HtNLcN8+GEUexd4nYf8Cxil3xrxmPsXjM7AyLC98GQWsezOMYjeE8OxHQ0YmDt7OemXnZOqNIb0bdo3P24qY61Ov8tfWxwuJEo/pVOIxnZbHdCEkda8IiIAIiIAIiIAIFBAYyil7I/IfgF0OO1Knzn3IK8O4t+fvwCgMy7DasAEZ74RdCdsEewOMgQLy6SSVejzpmc2GZspY/6i/adjHtdc+uxoVeU+DGK5WZiI+cWh5tHwNdtkfMv/hkcprmDWrgi56lcCmXu24+j1vAhrzeaPr2Rs15j07dOr4oBLIE6b0btLoceRazq/BPgyrDSeQ8QFvtWW85jQ4v8K/jRcI4et7fuB0Coxlm2EhrPIJbtGUV5Z4MH099i8rNmuvfbVqxPJ6gW3Q8zonQJQmywumTMn8wvml/e+/r369OTcqoxcI1B3zXui4+jhvAhrzeaPr2Rs15j07dPPqeF0n07xa0k2LQiBMpxc9nNP51xVVKCjbj7J1Gfsg0hScFIgHYNfC2Db/IFEk74Ix7IUVlSWVuvHjnE3WmCZT+TbSVH43oOsZIiACIiACIiACA0kgz2MaYNCTSCG5FjYOy36ohEtzCyzrsWReUXi2pvAuXF8K44dKIQTxV1QW6nY8jqJ4EuI0mcrHLH7oW8efqweIgAiIgAiIgAiIwKARaCRM6VGlKGXYCLspSc383IpkK8KUa1PpQc2GK3DxCzDuEUoRnG0vr4we1ey0fKNrVJ9fgCid5J1TDqgiyzW1CiIgAiIgAiIgAiIgAh0g0EiY8kOf98HqCTKuB82KyIV0j2tVafVCUVm9+m3Nc9ZOWed4JCnadfKYtpWuGhMBERABERABERCBGQKNhClr3gOjd/L1sE2w52B/DxuIr9Mjg6l8vH66wb6VMMXAK4iACIiACIiACIhAJwg0I0z5YVJ2DWjoxwVIfCtc9GvsYnz8BFkuj2m/jrDeSwREQAREQAREYKkQaOar/KwovRsd5zpQhm/CmhG2SeWe/Yn8GlO+aqw1pj07juq4CIiACIiACIjAkifQSJiGD5W42T6n8/kxEvP40RPD2WnUv782To8k5T6mxmoqv39HWm8mAiIgAiIgAiKw2AQaCdPw0dNDNR0N18kX6zVlfXWJr7tmvsrXx099NbZ6GREQAREQAREQgaVFoNFUPD90YuA2UXtgPAHqx2BfhjG8lEb9++usm7JwFicb7Bt5TPt3pPVmIiACIiACIiACi02gkceU20F90nfyS4hfhlGcMvBY0SNJqo9/orCPaTKV74IHuY/fWK8mAiIgAiIgAiIgAotDoJEwZa/KsItgFKL88Ime038BewOs/0P24ydjtF1U/4+43lAEREAEREAERGCRCDSayrfo17dh9JIOhhCtGYg4jqestX67KAnTGjy6FAEREAEREAEREIG2EWjGY3oennZd257YYw1hc/3Mx08Spj02fOquCIiACIiACIhADxFo5DHlGlNO36+FjcPugGXDLbho17Gk2XaXTDqy0RQPuUq2i3J1j2ZdMn1VR0RABERABERABESglwk0Eqb0qFKUMmyE8ev8bOB+pn0tTJ2rTHIqP/0qXx7T7OArLQIiIAIiIAIiIALtJNBImMZ42Ptg9b5GX4X8vhalBF2CxzTGa0qYkoaCCIiACIiACIiACHSOQCNhyo+ffh72EOzmznVj6bbsuF2UlTBduiOknomACIiACIiACPQLgWY+fnonXrZ2Cr9f3r/he8RRjDWmJl1jaut6jhu2oQoiIAIiIAIiIAIiIAKNCTTymHKqPnz89BjSv55pktP7/y+smen8DajHE6N4ktS3YdOwENaHhI9ZdiCTtxXpM2CP1ORnqnQuGcNjWuJUvgMqfPyEl8WK06beuXOdUssiIAIiIAIiIAIi0IcEGgnT7MdP3DaKpz9lwx/iopEw3YY64QjTcO9ZSOyFcanAvpCZiZnPQIEaPr7iNfvwBBPdCsMumoxtxUyaYb5nZH7VLDOfMye69Xw9RwREQAREQAREQAQGhUAjYcqPnz4Iq3fi0UrkNxKlFJgUpV+H/QTsNBiFaBn2ARivGSg4w/Gmyb6huKagpSjlqVMPw8ZhPH2KYrnRc1GlPaESVaawxJQfP5FFySxbDRZHJEzbg1etiIAIiIAIiIAIiECVQCNhyorcu5T1zoedDbsXRkG5H9YoLPMV3oWYYpL33A27DEZhSnF7EFbPC3o78lmXopThAhhF7ZkwCsO/gf02bCeMgQKWffw8LxC2wHYnqQX8VOJocshiH1NO5TOMDtfboSAt068IiIAIiIAIiIAIiMC8CdD72ChwjSc/APomjFP5FJsUiNfDGgUKSHgYE28oBR3buhx2LYzhYtha2DiMwvUh2GoYw0bYo0kq/XnRp0cQsy16WXfCKEApYHfB2CfmPwB7ChaWBCA5vzAyNDXz8VPSxBTfR0EEREAEREAEREAERKDNBJoRppw+rxduQmYQkfXKQx7FKQXicVho62lfeLqPb0BMjyhF5WFY8EpmhSmyk3BOSCAeg9Ereo3Po/Cl9/Uqf71gYVqJS8nSAr+PKZotSZh6uIpEQAREQAREQAREoJ0E/Px0bpNBIH4SNX4H9jKMQpMikh7UU2FHYI0CvaEUibQ7YfRuUhTfDLsVxvWbDOth9FBSYDLQG/pgkpr5CaKWOUd99rCPa699djViP/LCufUKnvtPH1mx4X2/hY+fSonIvffJCfaJIluhtwls6u3uq/fzIKAxnwe0Hr9FY97jA6juDx6BRsI0CMYNQBNEKim9pUlUFHucUqcIpSik3QL7RRiF3s/DvgLbD2OopFEifrn2dLO/ZrTKpxMPpk+zjazYrL321aoRy+sFtkHP65xw7pU7Ro5Cilfw3RPDZa8dwZKCE3XrzrlZGUudgMZxqY9Q+/unMW8/06XeosZ8qY9Qe/tX18nU3keotU4SaDSVP+EfTiH5A58eRxw+MHrW5+VFz/mCHYj5rHWwv/R5FL23wfiXBkUvy+lNZWAe16FeB+MfMgpoelkZ9qZRd36fLF+WrDGtmChyiZaONZXfHfR6igiIgAiIgAiIwIARaCRMieOUGiYb/TWntINHtaZK9ZLT/O+DUZjSG8qlAGtgoU2KTn78xKlxllMAs10K4rtgd8PocaU45PrTRRCF3CwKH+VDlEKcIhktQh/wWAUREAEREAEREAER6HMCzQjTA2DAKfCzYBSNNN4XpkeuRroo3INCzoPzQyWKOnpN2SYDp/DZNr/0ZxnToV0kzRUw5lPI8plc38pArynrBmHc6Jr3LCRM8ubJxHErj+lCQOpeERABERABERABEcgj0IwwDfdS/FE00uhFZOD9YVo/ycj5oYCk9zQIy9pq9JDmlTGfQjY8s/beblwn0/nTDvo6stm1tt14tp4hAiIgAiIgAiIgAgNBoBVhmgeEHyn1eXCJxzTdMsrRg6sgAiIgAiIgAiIgAiLQZgLtEKZt7tJSbM4mHtNUmFoJ06U4ROqTCIiACIiACIhAzxOQMG1uCL3HlEtl5TFtDplqiYAIiIAIiIAIiEBrBCRMm+M14zF1WmPaHDLVEgEREAEREAEREIHWCEiYNsPL4YN8BE3lNwNLdURABERABERABERgfgQWKkz5pfyz83t0D91lvTB13OdfU/k9NHLqqgiIgAiIgAiIQA8RyDuSdBPe4c2wvC2c+IrcNuke2Bt40ech3S4q2Y5VHz/1+Vjr9URABERABERABBaJQJ4wvRL92dFEn/4z6oRN7puo3ptVLDymDr7hZIN967SPaW8Oo3otAiIgAiIgAiKwxAnkCdM70O9VsKMF/Wf5Ym56X9C19hY55+AxtWY68ZguxrGo7X0ftSYCIiACIiACIiACS5FAnjDlUaEfznSYx3++GrYe9g0YRSlPchqM4OwkD0BNP36SMB2MQddbioAIiIAIiIAIdJtAMx8/bUKnOF3/FGwXjOfaH4Zthw1EsDbdYD+ZyjcSpgMx6HpJERABERABERCBrhNoRpg+k9OrncgfiPWWzsTJdlHTjhvsS5gSgoIIiIAIiIAIiIAItJtAI2E66h94K+LTffok4p/y6TN83N8Rp/IRkql8NxhivL8HVG8nAiIgAiIgAiKwFAk0Eqahz9mPoPjB01m+YCRU6Oc4TOVP6eOnfh5mvZsIiIAIiIAIiMAiE2gkTCd8/3Yg3ufTjyHe6dPf83GjaAMqbIO9CVbvg6utyL8Etg5WG4rKaut25Nq5dCo/8ZhaeUw7AlmNioAIiIAIiIAIDDyBeiKxFgq9oz/IZJ7n0xcgns7k5yUpSL9cU8g29/q8A4jXZsrZ/hNNlGVu6XDSf/wEYTqNDbKGXNkss2XDJQ0KIiACIiACIiACIiACbSLQyGPKx3BrKK4vpWC8yNsw4m/BGgVuM0VR+nUYnxXWqZaRZqBopShlu6y7B/a4TxeVoUr3gjXpGtOTZjRZa2pOrhmIj766R1hPEgEREAEREAEREIFULBZxoJjkNlGcxqcX82FvzXhKUTXZWorxu2Bcm8r9Ue+GXQZjuB3Ga7bLQC8sw5mwojJO+XNJwXYY26VdCLs6c30u0m0JznCDfWMm3HASm9LkirY0rEZEQAREQAREQAREQASqBBp5TLl/Kb2YDPRg8iv9rDG/KJxAIUXcERi9jFthl8OuhTFshD2apNKfF32aH1UVlbEtenB3wrbAKG65x+r1MOY/AKOgphe2HSHxlJ7kVD6DHZYwbQdVtSECIiACIiACIiACGQKNhCmFHQUiA6fkua4ya43u530Up2znOIzT9AxPw8K9WWGaFOIneDvrlZ0TKiEeg+2GXePzKHzp2b3KX7dFmFrrPaZm1AvTSMLUA1YkAiIgAiIgAiIgAu0i0OjjJ06RtyOwHYpE2p0wejfDs+nxfBCWDRSuDEVlLA/bWHHNK0PtdZo781v0PkEMz9T2qamXnxsbWrfBHK9gg330+u5/mGC/KLgVepfApt7tuno+TwIa83mC6+HbNOY9PHjq+mASCOKw9u05VR6EVzNex1HUn6htBNcUe5xSp3eUopB2C+wXffog4s2wEFb5BL2yeWXpB0hpRfYtKzZrr31z1SjvXdgGPa91A0TpXhZMRcuSd7x868iLV/zJidz6dRtR5lIkoDFciqPS2T5pzDvLdym2rjFfiqPSuT7lOpk690i13E4CYTq9ts2PIINijfuO5ok53sOp83FY3tZJz6GMYQeMz+JHS38JY+D6Va41vQ7GP0gUybtgDBSCRWVJpW79hI+fTprhSvrMWFP53YKv54iACIiACIiACAwMgTyP6RdBgGLym54E14b+LYyCkV7Nn4LxI6MQeF0v8KOn98G+BGN7DPSEnpKkjLkL8aUwelVDCKKvqCzU7UocYbsoqvQJM0IxjaA1pikH/YqACIiACIiACIhA+wjkeUw59YEFleZW/yiKUE6/U1zSwxlE6e8jvQxWu0YUWdVwD1Jsix8qUXTSa3oAFsIVSDCfYpX9CUsIkDR5ZRTI9OR6oZgI5qJrtjXv4PwG+5NmKH1eHHOpg4IIiIAIiIAIiIAIiEAbCeR5TPkIirAPw7gF00bYm2HrYdyL9O9ge2B0JDYT2Ba9p3mBYjQrSLP1isqy9TqW5pGk2GRfHtOOEVbDIiACIiACIiACIjDzZXwti9XIOLMm89s116/GNT2rfR8iG00557DB/pAX4lpj2veDrhcUAREQAREQARHoOoE8jymn6zlt3yhwij5Mpzeq27PlUKOT7PyEG/HC1HLpgYIIiIAIiIAIiIAIiEAbCeQJU3789DMwfqjEwGn4w7DLeTFoIY7dVIQVrFhjynWsWAnrtMZ00P4Q6H1FQAREQAREQAQ6TiBPmHKK/g01Ty9nrrm+9AJY33tL+c6lCF/lYyp/yoSpfHlMyUVBBERABERABERABNpJIE+YZp/xXlxwu6cQLkLi4XAxCLHDdlH8zqvqMTVOU/mDMPB6RxEQAREQAREQga4SKBKm56In2f1Fr8T1XV3t3RJ5WFyJpyJsZDXlSulUvmmPx9R9bOUZZsRYWz72/BJ5VXVDBERABERABERABBaNQJ4w3Y4e7cz06gGkuW3T1Zk8eg15vKj/IChT0mfJ0hCm8mNnpjGpn7yaW/gaU3fDaavN8OTzpIf0mL3pxaLttPqMqF5HBERABERABERABOYSyBOmp9dUfSeuabWBG/D3vTB10/GUwddP07aNHtMVk2dUyY1MvgIcJUxr/3TpWgREQAREQAREYKAI5AnT3/MU9hXQ4NGkfS9K+f7xsJ2MKsZMu4jbYyG0YSq/4s6g2E2bcxSm300v9CsCIiACIiACIiACg0kgT5jSe3fzYCKZ+9ZD06Wp2FawBcGQF6Zt+PipVILH1G9qUIrOmPtU5YiACIiACIiACIjAYBFI10zOfWeuJaU3tJHl3T+3xR7OieOpZIP92NhUmFqz8H1MY3hMqyHxmFavlBABERABERABERCBQSQwEMJyoQNbGeV2UcZUTJjKNwvfLsrOEqOcylcQAREQAREQAREQgYEmkDeV/7ugck8TZAZig/3RiZEpzObDfWwDr3YIU0zl+zWmxkqYNvGHTVVEQAREQAREQAT6m0AQWrVvyQ+b1tVmZq7pQRyB7c/k9W2yMn180pSGKUyH/UsuXJg6risN347N8p72LUe9mAiIgAiIgAiIgAgUEcgTpr+Mm24qutGXcc1l33tNTwytmFqGA0khI1Nerg1rTC3WmFZ1qcmsN22CuqqIgAiIgAiIgAiIQB8SyBOmx2vedQ+uvwE7CqM3dQy2GhakFZKFgXXfAXsZ9gjsACyE9SHh42nE2fKtuKZwq73PV+989MrDp06+uC45nIle4sRb7MpmxJaT9Pw64LJiVB7T+UHUXSIgAiIgAiIgAv1EIE+YPlvzkhtxTXHKdac8BYrlzXpKa482xa3mSthdMC6y3AerDcxnoEBdm6TSn/MQPZG57krykS/82PTGjzxAET6M6fyD1jgI1LWYzj+YfBQ1r05YCNMg6609A0mLlw4582pSN4mACIiACIiACIhALxPI+yr/PrwUxSE9oxfBfh+2EfZ52DMwbDefiKggIHGZG57yJVyfuQy2B7YTxnBaGhkKzs3ezvJ52xBTlPL5fA7ve9ynEXUzWArGKT5x0g2dSJ88CWE6v5AcR+rMSrzVMbTwMkgOm/LqU+fXmu4SAREQAREQAREQgf4gkCdMw9txo/2HYR+AvRVGgZoNFIxFIbRP4ckp+gkYBScDy1bCDsLoBR33thcxw+2wu2F8PsMFaWTORMwPsx6DbYdRNNIuhGX3X6Wntp0hFaZm6GTS6PTw/Pcy5XGkDM68AHH6QpI2ms5POehXBERABERABERgUAkE4Vj7/szn2s4ybBxG4fcD2C/CGL4O+yCs0XQ+y+nx/EdYCO/xCbZ5MYxe0XEYrx+CrYYxbIQ9mqTSnxd9GtPoycdHFLs7YVtgFLC7YNfDmM/lBvTUNhLOqNJ0SKbtp+xw6jG10/P2mBoeR8qQiFKXLF41MU6CUhABERABERABERCBASaQt8b0OjC5KcOF0+h3wP4M9n0YPZ8Uks2E4PGkh/HPYW+DUdTy/tNhDDfAvgf7GuwwLIi+rDBFdhLOwe8/+PQY4iOwa2CXwyimeX0VbB+MwrTZfqJqYUiE6UTwmJpS6GPhTXULw3GkztJjmgpdeUzrolKmCIiACIiACIjA4BDIE6a1X+XTe0mhSsuGZreLugU3Uexy2v4sWJiuvxnpW2HB87oeaU6ZU2Ay0Bv6YJKa+Xl6JpnsEsDLYZ/HXQMYwnV6NfNbJFKLp/7jiuPBTwcnh+yZ8Nne990K6/PjrJbDd56f3vr6MyLzwrHKyenYTpy12pqnXqrwnYv70PKTdEMBgU0FZSrqTwIa8/4c16K30pgX0VGZCCxBAnnC9Nk29vV+tPVOGEXm7pp2t+P6K7D9Pp8fVTHQi0gRu5kXPqzyceK59Olaj2jtdbg3xCyvFyhYa/s2u15USsS6HVp2iAXbXlN6qeE9s1uoXkGUJksozlgZ7TbWHaNPd8tpQxT5xX2otqBEmwiId5tA9lAzGvMeGqw2dVVj3iaQPdKMHDw9MlB53UwEUp3C+5BHEdfIgqezThNJFqfvKUrpFaV3kX9gaMEjehvS/EuD9diXO2EMzLsWRi8r61NA74Ix7E2jbv+65OOnCTPMZQwIdv5T+dxcn8HZ5+Er1sdPCQz9iIAIiIAIiIAIDDqBPI9pu7iED3ooMGnZQA8hRec+WHbpAD2rFH93wS6FPQULYf5iMLQwz9gaO0m36nEzOpk24ebfl3AcaYQ1prE9lrTHfU0VREAEREAEREAERGCACeR5TNuFZBwN5Xld6W3lFD7Lub8phR7T2WmXK3z+KYjZV07xM9BryrrBY9vomvcsKECUJh7T42bEC9M2eEwrlReMi9Kv8o19xYI6qJtFQAREQAREQAREoMcJdNpj2iwePz1etzrFaBCkdSt0J9NBkFpz0ixLBKqJHZcfzC+E40hL8JhOTh3hR1WY15cwnR9N3SUCIiACIiACItAnBDrtMe0TTNxzyiae0qNuNBWmC1pj6qftj4+8YL577EXo3QoecKr7QO5uAn3DUS8iAiIgAiIgAiIgAnkEJEzzyNTkY91AIkhPBGFq43mtMc0eR2pvevGIvTcRpVzSEJn1y0+veawuRUAEREAEREAERGBgCEiYNj3UnMo3ONx+OY9WRZjnGtPscaRpQ/xN15kOj+gDqBkmSomACIiACIiACAwYAQnTJgfcWpt6TM2yVJjaea4xnXUcafXh/gMorTOtElFCBERABERABERg4AhImDY55M6Z1GPqlvtDAObpMeVxpAw8jnQmSJjOsFBKBERABERABERgQAlImDY78C5dY3rULfPCdJ77mMZ+c30bZ4Sp88I00pf5zY6H6omACIiACIiACPQdAQnTJofURanH9IhZ7vdOnafH1Prp+qzH1AbvqabymxwOVRMBERABERABEehDAhKmTQ6qjdMjSY+7ZTwAClPx81xjmj2OdObZqcc07G86k6+UCIiACIiACIiACAwMAQnTZoc68vuYmtGFeUyT40jxUB5HWg06/amKQgkREAEREAEREIGBJSBh2uTQh4+fjtvR1GM63+2igseUx5FWQ0UfP1VZKCECIiACIiACIjCoBCRMmxx569Kp/BNmhfeYzvPjpzBdz+NIq6HkhanRx09VJkqIgAiIgAiIgAgMGgEJ02ZH3E/ln3BDOAQKwZrlzd46q57NHEfqC2z54EEkT8LGXHnDvE6UmvUMXYiACIiACIiACIhADxKQMG1y0FycbrB/3C33U/mmZQFZexzprEdbk3pQp4/p9KdZYHQhAiIgAiIgAiIwKAQkTJsc6cjEyQb7E24k9Zia1oWpqX8caehBOp1vY03nByKKRUAEREAEREAEBoqAhGmTw+38kaRTNirhlqPGmWFXPmWsydvTavWPIw1NeGGqdaYBiGIREAEREAEREIHBItAtYboaWN8LuwS2rg7irfMsq9NUZ7KcS7eLstaNYH3pc+lTpja09LT6x5H6JnT6U0ssVVkEREAEREAERKDvCAx14Y3OxTOeqnnOlbi+y+cdQLzWpxmdB3vCXxeV+SrdiSLrpri4NI6jEXz5tBc77L/axKUzkfVPTfeAx5FyIcCs40j93eEkqNRZJ68AAC/xSURBVHAyVNONqqIIiIAIiIAIiIAI9AeBbnhMgygdBrJlsD2wnTCGbTCK0otglGwse9yni8pQpbsh7GMKj+kwTn3yHtO4NY9pEJ1BhGZfwVk/lQ/xqiACIiACIiACIiACA0ig08I0tE8v6DRsAkbBycCy22F3wx6GMVyQRoaeyKIyLgd4DLYdRkcm7ULY1ZlremrbFpxJp/INp/KNgccUIXLsZ/MhbK4fROisO+NUmBqrj59mcdGFCIiACIiACIjAoBAIwrFT78vN6OkN/cfMA97j0xSTG2GPZspe9GmKv6Iy7iFKsbsTtgVGcbsLdj2M+Q/A6KmlF7YtgVP5SUPODkOceo+pbc1jWvc4Ut89Z7ww1cdPbRkwNSICIiACIiACItBzBDotTAmE3lB6SykmH4LtgH0QFkRjVpgiOwnB21mv7JxQCfEYbDfsGp/Hj6i4PvUqfx2e4S/nH4WpfLQA0cw1pggu8ewmyaZ+gsd01nGk/k4XSZg2BVGVREAEREAEREAE+pVANz5+IrtbYNfBeMLRWTAKuyCK6fF8EJYNT/uLojJWOerrDfu49tpnVyN6afNCEMN1y088++3Tl28837iJ42vv21MpbdtSMscmzQ+jcuF92cYmp80rR7DZ1Ge/McnN+Wfd95b/eHz5375/uXFx4jGdVZZtQ+m2EdjUtpbUUK8Q0Jj3yki1r58a8/axVEsi0BUC3RCm9+NN3gmjyKR3MwRO81Oobg4ZiFf5NI/nzCubzNSnRzQrNmuvM1WTZJ4HlW1k+1Z7n1mx8fzNyYNGV0xtOyf6O1ZYOZp8uFV4X7ahkSFzCnv7sR9b/d/+7V9NHMmW/d0PuFJg+WFreSzp2n3+mNJsFaXbT6DpsWv/o9XiIhHQmC8S+EV8rMZ8EeEvwqPl2FkE6O18ZPBatrPNbFucvqcovRV2AMY/MDROuTNcC6MnlXkUybtgDHthRWVJpW7+OBsla0yhbEfM0HC6xrSFqfzC40hnXsRP51f0AdQME6VEQAREQAREQAQGhECnhWnY+ojicx+MHyTRuCUUn30X7G4Y8yj8+OFSOIO+qAzVuhucq0z6Jw7b8suHkeaygVVNn/5UfBypb9pvGWVKEqaeiCIREAEREAEREIHBIdBpYToOlJw+r2ecyme4AkYxegqM/TkBCyGvjB5VthnaaHQd2pt3XHImEaaYiR9JGmn19Kfi40h9v8LpT7GE6bxHSjeKgAiIgAiIgAj0KoFOC9NmuVCMcqo/WcZZc1NRWU3Vzl3GUclP5WOD/ST4L/PT058aP7jwOFJ/uzUv+JSEaWOiqiECIiACIiACItBnBJaKMF3yWJ2b9lP5NvWYtnr6E48jZah3HGn17f1UvtPpT1UkSoiACIiACIiACAwMAQnTJod6yHtMUT1sTcXlA82f/lR0HGnSEH90+lMVhRIiIAIiIAIiIAIDR0DCtMkhj13kPaZhjWmLpz+FzfXrHkcaOhE+ftLpT4GIYhEQAREQAREQgcEhIGHa5FhPl6aTNaaoPnuNabNbRhUdR1rtg4RpFYUSIiACIiACIiACA0dAwrTJIV8Wl2Z7TOMo3cvU2A1NNRE8pvWOIw0NTE3q46fAQrEIiIAIiIAIiMDAEZAwbXLIp09MBI9p+vFTXEnXmFp3ZlNNOJN+/FSyQXzOvW3/Ce71GmMjrPXuMoPDSxVEQAREQAREQAREYHAISJg2OdaTa4eCxzSdym/19CfrhenxkVxhar+AQwaseQmbZpXMq1ee1mTXVE0EREAEREAEREAE+oKAhGmTw7jXrArCNPGYtnL6U5PHkfqe+HWmpWHtZdrk2KiaCIiACIiACIhAfxCQMG12HMtvn0ZVHgAwhIinTvHsKb/OdKp4nWlTx5EmLeLHn/5kdfpTIKJYBERABERABERgMAhImLY2zonX9JwPfTVdZ2qaPP2pqeNIfUdcOP3JyWPa2tiotgiIgAiIgAiIQI8TkDBtbQCTD6AOnXo0XWfa7OlPzRxHWu2Hn8qP/JrUar4SIiACIiACIiACItDfBCRMWxvfxGM6emjMe0xN+mV+1ODL/KaOI/UdcTr9qbUhUW0REAEREAEREIF+ISBh2tpIJh7T4eUTqcfUNnn6U1PHkYaOaJP9QEKxCIiACIiACIjAYBGQMG1hvJ2zicc0sm72GtNGpz+FzfULjyMNHQnCVGtMAxHFIiACIiACIiACg0FAwrSFcbbWJR7TycmRVJg2e/pTU8eR+o5EFb/PqdXHTy2MjaqKgAiIgAiIgAj0PoFuC9M31UG2HnlZW1dTZyuuL4HV5tdU68pl4jEtDVXSqfxmT38KHtOi40ir3fce03BSVDVfCREQAREQAREQARHobwLdFKYUpd+EZZ/J/UB5DGfWXsZ1CAeQeBz2NRjzKVIXL9h0Kt9OxanHtNnTn4LILDqONLxV+chL2B+VJ0Ctcx8yoyFbsQiIgAiIgAiIgAj0O4GsSOzUu9IbeguMorQ2hGM3z0PBZm9n+UrbEK+FXQSjgN0Do0hlenGCS6fyK0OlxGPawulP6bR8wXGk4YXwcs449wJ+rTl13RkhX7EIiIAIiIAIiIAI9DuBbgjTtwPidbCDdWCu9PlPIB73lm7BZMztuL4b9jCM4YI0Mmci5rT+Y7DtMJ7GRLsQdnXm+lyk2xucSabysdY09Ziy9QanP7ny2k2otQL19tubXjzCWxqH8AFURetMG8NSDREQAREQAREQgT4h0A1heg9Y0cv5ujrMLkYevaLjMIrLh2CrYQwbYY8mqfTnRZ+mKFwOo5d1J2wLjAJ2F+x6GPMfgD0F43PbFnAQafLxUxRH6RrTpOUGpz/FLhXUrq7HuH7frE5/qg9GuSIgAiIgAiIgAv1MoBvCtIjf6b7wBsQUcBSVh2EUngxZYZrmGHNOSCAeg+2GXePzuAaV3ter/HVbhSkaSzymlazHtNHpT9aED76+5fvUROQ9pnGkqfwmaKmKCIiACIiACIhAfxAYWuTXuBnPvxUW+36sR0yvJAUmA72hDyapmZ+nZ5LmqE8HD2btdaZqkqRXNi80nPqPJ46NRKMrzcT4o5vRSFL/B4fdibPGrHnq5QpF9TdrGz8yGV+8esSav3pmem+4p7ZO7fWeg5XJjWsi872DFXqZG/ar9n5dN0VgU1O1VKmfCGjM+2k0m3sXjXlznFRLBJYMgcUWpttB4iuw/Z5IxccnEHNNKgVgCKt8IvFa+jQ9olmxWXsd7g0xy+sFtkHPa2Gwoyu5S4BZvvFHuYtAUv+sNeYf2IMtpyUfRM1qw5WxA4Gzr+U9P/lK+6eI/B6lzMkPG9dGSZtnr0uWDMxqM/8ulcyDgNjOA1qP36Ix7/EBnEf3NebzgNbDt8iZ08ODx64v9lT+begD/9Lg1D37cieMgXnXwq6D8Q8ZBfQuGAM9j4sS0MFJPjjOTuUbv8a03ulP06vo8eVyg2ftZ481JUrZPpbGPp/GOv0p5aBfERABERABERCBQSCwGB7TrIeTopPex+MZ2BRzE7C7YJfC+BFTCCtCYjFi7OOEZQbWWGfD0gGo1Og5Y7kSwW6Y06coSteXWjdnin9O3WxGJX7BRHTu6vSnLBalRUAEREAEREAE+ptANz2m9HTWTrVzCp95y2AUnUzTWxrCFUgw/xQY+8opfobQVlib2ug6vWuBv9haNPGYWmtGqk0Vnf4UhS2uohY+fELLzntMrdHHT1XQSoiACIiACIiACPQ7gW4K0yKW9JAG0Vlbj/lc25n1tNbW6co1VHOyXVTs4hmPadHpT84L03juR1GFHZ4cSafyndnoyuvD2trCW1QoAiIgAiIgAiIgAr1OYKkI0x7h6FKPqYmqHtO805/cBwzF6/nwATtMyz/SygumG/Hb7+AeLLWY+INW7lVdERABERABERABEehVAhKmLYwcpvITj6l1bsZjyvvrnf50xtrXo4RLFHbb8kHuMNBaqMT/O244DFn7Plce44dgCiIgAiIgAiIgAiLQ1wQkTFsY3sivMZ39VT4bqHP6U8mf+GRMa+tLfX/sZ448ZZy7MvG4GnOT+8Tq/7mFrqqqCIiACIiACIiACPQcAQnTFobMWX6VTwfpzFR+cnu905/mu7400x/7qSP/BUtrb4LXdBjLAe5xH1txZqZYSREQAREQAREQARHoKwISpi0Mp3PpPqZYNjp7Kj/srRq5GeFoYx6xir0EWtwqqrY/Tx75txCn/z+yX2GGh+7xa1dra+laBERABERABERABHqegIRpC0NYnco3rvrxU3K7dc+lzaR7mbryhhXY8onHiU4bs/rbLTxiTlV7r6mYSfezcNN+H4UXmQ1jN8+ppAwREAEREAEREAER6AMCEqYtDGJ1Kj+7wX5yf+3pT0fOR/YQduJ/0pb3Zg8PaOFpM1XtbxzdD3l6GcTpBKb1r3E3jl0+U6qUCIiACIiACIiACPQHAQnTFsYxTOVDIM72mPL0pyT4059im07ju6i1E58K+mI/ffhvsbo1/To/Mr/vyqv41b+CCIiACIiACIiACPQNAQnTFoYy8h8/uVqPae3pT+HEJ9vixvoN+mLLh34PovgueE1XGhf9ifvIujUNblGxCIiACIiACIiACPQMAQnTFoZq5kjSmjWmc09/elPSbMW2zWNa7eahw/8aaa5bPdcsi3fiOCwe49pzoVf73XOg1WEREAEREAER6CECOFlIoVkCLrZT1iYno86ayufpT27H2FG0s8qV1/2QMZVzkT5pXjjI05vaGuznzAlXtu/Bl/rfwj6n7zblNR8x5UO/2daHdLgx94mxC0zJ3IWlEZvxqIOQ1jxy9iA8wWlczXMHTcy86KCJfNk04mnUXXbokC2jVEEEREAEREAERKBvCEiYtjCUURRPYhofetDWbhdFv+VzEFavhih9F2J6Mb9tv2CSfU9beERTVTGl/z2cBvVzqHwfBOpnXHnNt5D3YFM3L2IlV8baXDf2CXTho2AU/uy9AulXzOlWov+BMfEH4yK5Rq2SNzcWux04GSsrYhNhCxGbCN0gaiFiWSdyELjIm4wPmBPHD0LgK4iACIiACIiACCwxAkEcLLFuLc3uQBtNsmfwms7ymKa95Zf57tUQUP8yuW7z+tL0GTO/tnz4KxCnn8bzduDr/y/CU/smWz7w7EyNpZVyn1j7Boj2u9ArxPR02luwx8BnTTS1zIxE63C91sQWsVmbWGTSPMc4XguxmblO6oxV6zqzyStYZGVEbCJqkcWAf1AkdYaxemV4zJz8t+Z7oyXzK/ZThx9IyvUjAiIgAiIgAiKw6AQkTFsYAleJpmwE4WPnbLAP4RP2MjX/a9Jk3N4Pn+p38/CnjB17M7TYOyD67nUfMm+zt0PuLaEAL+mQibHcwMY3QhxS0D8NfldBWO/KdNPvapDJaZBEu5E5uWYNWl8LSwVt7EXsLFGL8kTUesGbCF1z2uiQORuPuB/i/i9MJb7Ofvrokw0eqWIREAEREAEREIEOE5AwbQFwqRRPxslUfs2RpGkb8JhmQhx/K3PVkSTXWEKg/RyEF9abmjebU1f/jjFHfrkjD5tHo+7jq16L3QN2Qoi+GX3kZPwd5ujyj9hbXjg2j+Zm3ZKuLz10AJm0Z2YVNrjg6Vn/NLyq/Jr10b9Gry41UfSTEKh3mgm3w/7GkZca3K5iERABERABERCBDhHo9lf56dfqc19mK7IugdHzVRuKymrrdvQ6xsdPfECdI0npRc16/Q6boaNPdbQzvnF4Hl82leg9WF9wAh7JX3I3rr6qG88tega9mejHdaYU/T3qUZTuwTT9T6GvH2yHKC16djNlXPv72juO3gUh+mr07Y7kHmeuNqP2uxCo1+rY12Yoqo4IiIAIiIAItJ9AN4UpRSm3T6p9Jj1ej8O+BnsZRiEaQlFZqNO12JWiZI0pvGyckq4J/vQn5lr3SOrRq6nSoUv76YOPYh3mv0maj6I7sJ7zjR16VMNm3cfX/DA+cHoIQvkWVF6G+D/gEKzz7KcO/VXDm7tcgd5RimUTx+dDoP4FxnUd7HM49vUJnK71zi53R48TAREQAREQgYEnUCsSOwFkPRqlSKm3p+c25PNjl4tg/DplD4wilemiMhR3P5QqlcRjitn84TlPr57+hJI2nvg05zk5GbZ8ZCcE8Z1Y67rclOI/gefvlJyqHclOVt6Wx/6NGXKP4QEcz70YxXdit4D3czutjjy0TY1yfSkE6k9DlL4LfX4K8RbEXH/65+4Tq17XpseoGREQAREQAREQgQYEuiFM344+XAfjtj214XZk3A172Bdc4OMzEReVccqfAmg7jGsXaRfCrs5cn4t0W4NzqccUqnmuxzSqzEzlO6z5XIzw0pFr8Ni/A43NiP+IU+rd6Eayd2t57Gt47h2wlRB1f2wm7Osh9r7Sjee36xnJF/p7D29F/38VdgDvwvWn34ZA/bz7tdWntus5akcEREAEREAERKA+gW4Il3vwaHpA63meNiIf09DV8KJPUfgVlS1H+XmwnbAtMIpbfuV9PYz5D8C4xpPPbVuIo9RjigbnClMzvLf6oKgDJz5VG89PJF/k29JleOv9EFX4Un/sxvza7Slx5dX/l3GVJ/C8n0SL+4yN3gNB+nP2N5MPk9rzkC62wvWn6P9tWn/aReh6lAiIgAiIgAh4At0Qpnmww7OzwjTUPdcn6pWdEyohHoPthtFTyMD1qU/AruIFQnuFaWzTfUyxE2ba/Myvn64+SlFoywfHZ0q6m0r3MrVXoB8ViMVPwNv3v3WiB668YgPavh8fXGENKcbBuj/B5vXwkh78/zrxvG63qfWn3Sau54mACIiACIgAvkpZAhDo8aw9tehp36+iMlbhMaAMQSjWXqelM7+c8s8LQQznlZuj//WPN6y5+Ofwrcz0SlSaU3+iYl46Oe3G65XlNtqBAqzrfPa7H1p92zmn2Osqsbn7k29f9q92/PXJ77frUf/t/Su3VSqlT5QiswbtH/rWc5VP//gfHLsP7XOJBa0XwqZmOon1p1xX/CsP/p8rfuKijUMfxab8yfrTQx9dveuuRyd/81f+YuK7zbSjOkuCwKYl0Qt1opsENObdpK1niUAbCCymMOU551x3ujnzHqt8+mRBWfplfFqRHtGs2Ky99s1VozwPKtug57UwrH7b5YnwtVGCbU790ZLdA+HyEBqZU1bYcAcKz7n9yPVmx5ofLln37hvfNnLLjeeffCuO4TyxkEe5X1u1Hlsq/Xt4Sd+TtGPNA6XS9Ad+/A+OzyxjWMgDun9v0+N0yR8e341tpP4jvtjnOuYbx0btRR/68dE//dD/NKr9T7s/bgt5YtNjvpCH6N4lRUBjvqSGo+OdmeM06vgT9YC2EgjT6W1ttIXGrkXd62D8g0S1twvGQKFTVJZU6vZPpeK3i6q7xpS9cfgS3X2z2/2q9zwocGdORleijH8pn2/WjP37evWazXPltf8KR4d+x4vSw3jPX8BazHfZcs+K0mZfvVpP60+rKJQQAREQAREQgY4QWAxhmvVw3oW34odL/FCJU6b8cGkFjKGoLK3R5d9lQ8vYR4awdCC9mvndi3WWS0KYskv2tw4cwlGg2HzfHINM3e7Ka1o+Fcp9dM06rCX9I+PiP0GTp6OtvzK2tBXbU/0/M689WCmtPx2s8dbbioAIiIAIdI9AN4UpvaBw5M2aeuebXgGjGOW+m+xPdro5ryy0xeUADI2u01oL/J2cPDDpmxip25S1f2s/e+yFumWLlGnLR79jYvOL6ePdbe4TY29ptivJh1OjDl5S838k4tZiO67yYZzgdODZZtvo53ra/7SfR1fvJgIiIAIisBgEuilMi96PYvQALOtNDfWLykKdrsT/vGZ9A49p6c+70pEWH4L9Oe+GsPwd0B01JXNvsla0oA1XPmUM3tU/QH1uu7UBtstM2zdg6v538S+LemNU0Fr/F2n/0/4fY72hCIiACIhAdwgsFWHanbdd6FPKb59GE/TSDplyeQ67JX3C0d7D3ON1F2Tlq8yI/U/uMkjUOsHduAb7kU4/jhOkfgHFJxF/2NjDF9vPHPrvdaoryxPQ+lP9URABERABERCBhROYI64W3mTft5B4TTeZi+tP5y/R16dwMlPT70X3nseKiv/FvG71r2e76j58xkpM3X/eRI4nOG1E2d+ZSvyj9lNHbrXlRIxnqyudQ0DrT3PAKFsEREAEREAEmiAgYdoEpJoqyTrT0eOH8z6Aqqm+dC7tZ48/Z2L3XkzrQ1zbG9yNq3+GvcPpTReZlSe+DUF6NWwKX95/HF7SC+1njv7j0ul9b/VE6097a7zUWxEQAREQgaVBQMK09XFIhOnRUqWnPKbhNe2nj/xXpG+AALXGYkq/PPZlCNGHkHcOtoB6zMTRm+2nDv06vKRctqCwQAJaf7pAgLpdBERABERgoAhImLY+3MlU/vD0aM95TMOrJmfBW/MlXI9AoG5DzHWznzHmyJvtpw8+Fuopbg8BrT9tD0e1IgIiIAIi0P8EJExbH+PEYxpFcU96TGded/T98Jk+CftHyNK32k8e/gS8pMm7zdRRqp0EtP60nTTVlgiIgAiIQD8SkDBtfVQTj+lUXOpZjylf2Zb343jVoZ825vCP2k8fXjKHArQ+HL13h9af9t6YqcciIAIiIALdISBh2jrn1GNqe91jSnH68j/DS3qydQS6ox0EtP60HRTVhgiIgAiIQD8RkDBtdTStTTymkXU9PpXf6ourficIaP1pJ6iqTREQAREQgV4lIGHa6sg5l3hMp13U01P5rb626neWgNafdpavWhcBERABEegNAhKmrY+Tn8qXx7R1dLqjEQGtP21ESOUiIAIiIAL9TEDCtNXR9VP5No7lMW2Vneo3TUDrT5tGpYoiIAIiIAJ9REDCtMXBtH4qP7YlrTFtkZ2qt0ZA609b46XaIiACIiACvU9AwrTFMXQ8shPBluQxbRGdqs+TgNafzhOcbhMBERABEeg5AhKmrQ6ZtckaU1vBqUkKItBFAlp/2kXYepQIiIAIiMCiEFgKwnQ93jxr62pIbMX1JbDa/Jpq3bm0xiUe0ziyWmPaHeR6Sg0BrT+tAaJLERABERCBviGw2MLUguS+Gns5Q/cA0o/DvgZjPkXqogZn02M7bayv8hd1IAb84Vp/OuB/APT6IiACItCnBBZbmJ7muZ6HeLO3s3zeNsRrYRfBKGD3wChSmV60UF1jajSVv2iDoAdXCWj9aRWFEiIgAiIgAn1AYLGF6UowPAh7AjbubS9ihtthd8Me5gXCBWlkzkTMaf3HYNthztuFiK/OXJ+LdNtD5FKPqaby245WDS6AgNafLgCebhUBERABEVgyBBZbmF4MEvSKjsMoMB+CrYYxbIQ9mqTSnxd9mh8dLYfRy7oTtgVGAbsLdj2M+Q/AnoK13bvqIk3lg6vCEiWg9adLdGDULREQAREQgaYILLYwPd338gbE9IhSVB6GUXgyZIVpmmPMOSGBeAy2G3aNz+MaVHpfr/LXbRemkM/Jx0820pGknrGiJUZA60+X2ICoOyIgAiIgAk0TGGq6Zmcq3oxmb4XFvvn1iCn8KDAZ6A19MEnN/Dw9kzRHfTp8IV97namaJOmVzQtNTf1XDu1fVVqz3kwe+MEGNNTUPXkPVP6iEti0qE/vwsOx/pRP+Xf/7tLRr25/48hHx0Yt12t/zpw59rnKjeY41ktPOONOVpydZDo2bsLF9mTs3EQFS1Zi5CFOrqdjO1lhfmwmpiv25HTsJiYrZjKJY3MSFSdOTJlJFEwcm7Anj07GE4cnzMSLx+3k88cqJ79/2Ew89eLUxHNHTaULr573iE15BcrvWwIa874dWr1YvxJYbGG6HWC/AtvvAYf/aZ3ANdeebvb5jFb5dLKPqE/TI5oVm7XXvlo1yvOgsg16XhuG0prTXmCl4bVnHULU1D0NG1WFxSIwEOP3K38xsRv2gLtx7J1Y3HILYL8msmYF0iu42mXmL4Hwn0eIw7CE6xAzP5sO9RrFyUTINGqdxO38b/xkavakcQ7XiCGUZ5fz2ufHNlMHeVG4J0ad6KSpxL68gvZKsKGTZngKectQvv+kLaNl/TebQBiwn4H473zAxrTodeUwKqLTA2Uz/09anM7ehsfSNsAmYL8HY+BfJNfCdsK+APsebBeMYS+M9RcnwLuU/D/ZaruoxRkAPXW+BJL1p1h/7crUouuXmYkTy000sszYaVhlmSmVoBzjZcaUYIjjyF9HuHYwuwxiEHk+zdhZqk2UBUMdyzqIk3rId8nSnFCHf+esQl74hyYuw78tQ5zJqgpgCOFEC4c6iJMkf1iAGGo7DSUfQwNjf7f0r5YxE9/oTtjIPo+Mfaj+Au7DVnXxPmMt0siLoxeMm95nSrw+8jKEbJjJ8e0pEgEREAER6DSBxRam/JcN9zE9nnlRTt9TpN4FuxTGj5hCgIdncYO1dgrTn8bG2i5qcUdCT58vAQguKLb9XPYSlr7Mt6mW71tMUYz/dpfjP93N6DQNwYvaRODiEv9RG4tl98n12DT6ipkch7+fIGATIYu/q6zzIhbi1SE/Lu0zh1/eZ29P/s5KWtWPCIiACIjA/AkstjDlFD5dGqMwfoiFabdZ4Qpc/QKM3hZO7Yf/hdBryvtCaHQd6i04diae5KOd1clPC4apBgaOwGKK4p/YNHz+X1+54jCm/M/Af8Knmyg6A3+l8APM0/E3S5q2FjGujTkFeWfiv3VYJiQeWFzzbx+Lv44iaPxTx5zbYQ4hLxWtjuIVojW5plcWVqmkQnZo6AVbfpkfeCqIgAiIgAjUIbDYwjR0iR7SvECxSlsagfuYpv9TGlkaHVIvREAEmiHwN+NTx235EJcF0QqD+4AZNhtWrDeVEYjYOBWvkYNojSBiGUO8WopZi7JE3K6FkF2LfMz4+H8zJ/+M5g8FbMibNm7HGNfYpiI2WVbApQPZJQX0xvolBU8efcneu6gfjBVyUqEIiIAItJvAUhGm7X6vjrVnbZRM5eMr5uGOPUQNi4AILCoBbrmFFUaYiaEVB8hOa8pjOPQDArUC0WpL8MZ68ZqIWIhZ669T7+xqaNWNaJWGgBYoZhkx0BNrsU6W168bq7gfMS+l62CxrKC6pMCL2ZgiNlkfu88MH4Q3NvmgLGlGPyIgAiLQiwQkTFsdNYepfGvxvxErj2mr7FRfBPqQQCIpy4dfxqvR/qnRK7ryBqyVPw4RyyUFELMRPa8QrxS2tUsKnDkV7dFLS6/tTMC/jBMxS0csLcL6WDcGbyz2gQ7e2LCkgF7ZsKygEqdCdqj0gikfPJSRwzNtKyUCIiACi0hAwrRF+DE+fuL/B/B/AXlMW2Sn6iIgAtCR5b3HwWHcWyGS9GOxlaeZynAqYileq95Yvx42WVKQrIulsIU6TQ4eeXWqWH3ziTcWP9UlBRCy5THsY5t8fIp1sIjDTgUW4pXrYhNvLOLSFK6PvQhvLLf6UhABERCBjhKQMG0ZL7aLwt/gEKfymLbMTjeIgAi0QiAVg8e4xRWtMEBkYknB2jXGVOCNhRfWwgtbFbFzPu6iiEVd8ypvvm20Uv3Aiy3ymttvjcXwxr6cel65pIBLCeiJ9UsKmK5wWQGE7NBKLClIxLdvU5EIiIAINE9AwrR5VklNa+Mp/sWNv67lMW2RnaqLgAh0jkAyLV8+yN1LaLsbPQneWOx2sg4itgIRC49r+LiruqSAH3jRK5ssIzgNf+mlZuzrqm2HJQXMSA64ppA9yg+8jqYilp7YIF4hWqtLCrhLAcqGsHds+fABLSmoElVCBAaegIRpq38EuMF+GuQxbZWd6ouACCwZAvDG4kSsA8+iQ7TC4C4zJfO6VaemH3cNpethk7WxXBfrt92qLinATgXOhUMUfhhKFW2nC6CSJH+iRMWmReUxfFAK0ZrsGUsh6w8/qC4pwLWLsLRgYp/Zf2Jf+mFaYXdVKAIi0MMEJExbHLzIuqkYHlNN5bcITtVFQAR6lkCyZdW9R9O1qE28hSufMmbMtN+lgN5YL15rP+5K94zFVlvuLDRL84GeVy9mGfHwgwiTVBuGuWfsgarnNfnAC2I2eGK57VbFC9mhESwpCKddh3YVi4AILHUCEqYtjhD+ukw8pog1ld8iO1UXAREYDAL+EAEeJPB0ozfGkoIRY0/xSwrqfNyV7BnL7bboiTXr0R4PPzgFyvW11bbxF3LqfoWKrS4pmMAxtKt1DG0VkhIi0BsEJExbHCenqfwWiam6CIiACOQTwJIC/GP/5X9GDVphgIiF7Fx9yuwlBWHPWL8eNiwpgHcWx9CugIjdjEZpCFSwEK+JkGUSnlgdQ0swCiKwZAhImLY4FDE+fopchL/adCRpi+hUXQREQAQWRAAiFkryyItohPZko8Zyj6FNlhTM+riL+8TqGNpGQFUuAl0gIGHaIuQheEzxNyP+we308VOL7FRdBERABLpJYEHH0NZ+3JUsKdAxtN0cPz1rMAlImLY47pWoNGVjSlOtMW0RnaqLgAiIwJIlMK9jaKd5ehc9rwXH0KbeWe5SsBEvT0OoXVKAax1Dm6LR78ATkDBt8Y9A7KYmS9g5BUEe0xbZqboIiIAI9AOBZJXqzDG0/9jonXKPoXV+faz1H3clx8/qGNpGPFXe3wQkTFsc3+Gpoal4iP+6lce0RXSqLgIiIAIDSWBBx9AmBx9k9outLimApzYVsjqGdiD/VPXvS0uYtji28Whp0lRwZLSTx7RFdKouAiIgAiLQgAA+8ML/YFo8hnban95V8vvFpoI1PbUrOb0LIpZLCpzB/rI6hrbBEKh4kQn0gjDdCkb4D8w8AjuwyLzM1MTU1PAQJnI0lb/YQ6Hni4AIiMBAE0iXFFSPoX2qEQxst1X/GNrqkoJZH3fpGNpGQFXeEQJLXZhSiK7NvPl5SD+Rue56csXq0uTUCX381HXweqAIiIAIiMCCCMAbO/9jaJMlBfS8cj1slB5Lm9kzFh3TMbQLGh3dHAgsZWG6DZ2kKL0I9jBsHPY4jOd68JPGRQnTZjk2gz7GZ+vjp0UZAT1UBERABESg0wTmdQzt9DR2KcB62FIpXUZguDaWXljE2SUFxqzRMbSdHsHebT+Zk16i3R9Hv74Bu8L3j0fR8axmnqd8AvY3sN+G7YQxUMCeD/s8LxC2wHYnqcY/fu+OxhXNZfeUNp69AmuAFERABERABERABJYSgVf90c+cv+sHU48tpT6pL60RSE4Vbu2WrtXeiCc9mnkaT/pgoKdyOYzT+jthFKB3w3bBrocx/wEY19u0X3jf+94K2qUpiIAIiIAIiIAILCECEKV0XCmIQNsJhOn6SzItU2TSs8m8DT69GjEDvaksq71uVnjzXoXBInDuYL2u3hYENOaD98dAY64xHzwCPf7GS3mNKdHSG/pgDeOnM9dHfXrYx7XXmapJskiAFpXVtqNrERABERABERCBpUmg/bOlS/M9+7JXS1WY8rP3g7DNGeqrfBofH1VD8KKGjNrrkB/ivD+sFKV5ZeFexf1FQGPeX+PZzNtozJuh1F91NOb9NZ7NvA3HXEEEOkJgO1rlHzBOxVBAczFz+AMXpvLDVH2ja9xaGEK7hZVU2FcENOZ9NZxNvYzGvClMfVVJY95Xw9nUy2jMm8KkSvMl8EXcyD9kwfjRE0MjIVpbnt6V/6s/yPls+rVEY96vI5v/XhrzfDb9WqIx79eRzX8vjXk+G5W0iQDF6DpYJ6fa9Qe5TYPVQ81ozHtosNrUVY15m0D2UDMa8x4arDZ1VWPeJpBqZnEJ6A/y4vJfjKdrzBeD+uI+U2O+uPwX4+ka88WgvrjP1JgvLn89vU0Eym1qR82IgAiIgAiIgAgsHoHy4j1aTxYBERABERABERABERABERABERABERABERABERABERCBdhIotbOxHm1rK/rNY0wPwE726Duo28UEOMZvgfGEsB/AsmuQRnH94zAe5vAMTKG/CLzJv86RzGtpzDMw+ijJk//eBjsb9iyM+2GHoDEPJPor5naSb4SthO2reTWNeQ0QXfYGAYpRipRgFDAK/UXgFrxOGN8Qc5cHhnCUbchnHPbGTSrop6cJhG3jypm30JhnYPRR8kK8S/a/Y6bD9oL/o72zB7GjCsPwBiWoGBOFNDZOIMRKK61stxHZwkLb7WxsFUvX1kqwi31qi4iF1QoWYmdhIlrMItiIoBYGg4jvs3s+OFzursmurjmzzwfvPX/zc+b57p37zTn3zujzBTm6O5S6t3n5ne/zMn1eJEyHIrCV3vKG5oSGzRHl//K2VNm8dooE6uR0ve2TCw98/HYr14mNK2uCVdputjaT8Qngz97fHJE+H9+v644AP8+toT73b7ayPm8gFpQwOo7Py8evtzIjqJg+P+Dg62AE5vSXG/iX1cmMURZtGQQqEO1HQeccGiIY5cTGMmU1ulpl03EJcDGCf1FdiOjzcf15VM9rZLx/xDaBCoMO+vwocuO2Tek6n+3eKG9F+rynYn4oAv0XFh1npJS6KdKWQYCpvD7wLB9zQTJF+JuR0rJ1gWy1mY5DgN+V4lsuNnejCkyn5PV5ICzMGDXDr6RM585R/bZ4Sl6fB8LCjHM7fuVczoVJDSpcSH6K9HkgaGMRYASNN+5m1+0KWvq6rtns4ASY4sHniIClRlnwe1nV9SOs1WY6BgFGzfAxI2bYHNV0X/lXn0NmOVaBKX5nlLSClO3k9fly/Lx6JPh7VYyW6vNVUoOVz/oX8LNr/PX9mjqrxiXAlfUcfRt9FHEnip+i8xHWBykHNb6OTODj1vlbSbkYuRi9HPFlpc8DYcHGZ/uL6K3ok+iDSJ8HwgKtLjafyrFxDn+sHeOHSfV5gzFqclYDU24j8kt0pXPc4y1/t6szOzYB3t+/R79Gj0RvRHULmZ+Tx64eJPuvr7Z8LdM1mR2EwG/p5170efRldCl6Jfo00ueBsGBj9Kzsdsvo8yKyvJTv8Pon/p3k+cy/FunzQNDGJMA0DyeyaxHTf/UvvmS1hRAoH/N7synC14jRM4yTGu8B/D+1PNOA2nIIzDmU+o0pR6XPobAsu5DD4XO8E3ExOkWUa1RNnwfGwqz+hb+Z4zoXTRE+r8+6Pg8MbUwCN9Jt3swlpn215RDgi6l826dzO0R+j9TX32z1JsshwAVnfVlxVPp8Ob7tj4SLz/6zfL1r1OcdjAVl8XHv8/78rc8X5OizeCgEo09GXHVpZ5MAIy5IOzsE9PnyfM05nPM5Qck60+frqIxdx2wXPj9sUEmfj+1fey8BCUhAAhKQgAQkIAEJSEACEpCABCQgAQlIQAISkIAEJCABCUhAAhKQgAQkIAEJSEACEpCABCQgAQlIQAISkIAEJCABCUhAAhKQgAQk8KAR8M4cD5pH7I8EJCABCUhAAhI4gwTmHDP3ROQG7Vg9jOGg5KsEJCABCUhAAhKQgAROiQA36+Zm/DVqWjfuPqXduxsJSEACEoBAjQ5IQwISkMD/QYCHW+xGPKWLx8HOEU9wuRzxZLY5osyNtMt4khOPHESs+1yE7UQsS4q9FFFmO/9kP2aBH9pC/fKsT7DKuZLglX3OEX2tIHYzefqxFRHcbkeaBCQgAQlIQAISkMBgBKb0t0Yn6/nWVe5TAj+MoJF6lq08ZZ72w1Neap3tLl+Ba6oOtTktrPtQVHnK7IcAtLZL4Dm3Mim2+ujbnf1aXyQgAQlIQAISkIAEhiLwdHpL0De3Xl9rZQJCbIr6dkYlGa2kntHWCmZZD2P0soJI0p3oXoyAk+VrFqm2wbrVR/a1GgDThwpM2cbDkSYBCUhAAhKQgAQkMCCBCvoI6rDDyvN+68bGC0krGK3gkXRq7SR9e1d9ZLYPTAlOa9vkK/Csuj4lIK52gmJNAhKQgAROQKBGB06wCVeVgAQkcGICF1e2sFqu5q+SuRRdic5Fe1FvBK60l/F71Ps1As+yv5L5rBW+Tso5k+n+56MXo++ism8qYyoBCUhAAscjYGB6PG6uJQEJnD4BAtGyq8kQdD7TKu4mZZqdwBV77yDZeD/p1PL3mvT72clKe21FgtF3I7ZJkFr7SnbfzlfGVAISkIAEJCABCUhgPAKHTd3X1P7lHBIjmFUmGKWMmLLnX/Pkt6PdlqcO24lqWcpHWT+Vz3I3olqXC/ipK1c9fcdqKn/aL/kiAQlIQALHJtCPDBx7I64oAQlI4BQJ8AcjptP/uI99HvXP/FvZzp9rtlXnRwLRskdb5k5VmEpAAhKQwL9HgBO8JgEJSGAkAgSR6wLJw46BAHP7kMYnUv9OxOjrqvUBabUZkBYJUwlIQAISkIAEJCABCUhAAhKQgAQkIAEJSEACEpCABCQgAQlIQAISkIAEJCABCUhAAhKQgAQkIAEJSEACEpCABCQgAQlIQAISkIAEJCABCUhAAhKQgAQkIAEJSOAUCfwNzhOGYjmFcZsAAAAASUVORK5CYII="
},
"metadata": {
"jupyter-vega": "#adad4346-4c33-4fd9-8d7c-22b6667ed6ac"
},
"output_type": "display_data"
}
],
"source": [
"from altair import Chart\n",
"Chart(df).mark_line().encode(\n",
" x='max_iter', y='min(rel_error)', color='approx_grad')"
]
},
{
"cell_type": "code",
"execution_count": 8,
"metadata": {
"ExecuteTime": {
"end_time": "2017-09-01T19:51:08.904476Z",
"start_time": "2017-09-01T19:51:08.876499Z"
}
},
"outputs": [
{
"data": {
"text/html": [
"<div class=\"vega-embed\" id=\"97805eea-eb5e-4b51-b2b9-bc9c98c81a05\"></div>\n",
"\n",
"<style>\n",
".vega-embed svg, .vega-embed canvas {\n",
" border: 1px dotted gray;\n",
"}\n",
"\n",
".vega-embed .vega-actions a {\n",
" margin-right: 6px;\n",
"}\n",
"</style>\n"
]
},
"metadata": {
"jupyter-vega": "#97805eea-eb5e-4b51-b2b9-bc9c98c81a05"
},
"output_type": "display_data"
},
{
"data": {
"application/javascript": [
"var spec = {\"config\": {\"cell\": {\"width\": 500, \"height\": 350}}, \"encoding\": {\"color\": {\"field\": \"approx_grad\", \"type\": \"nominal\"}, \"x\": {\"field\": \"max_iter\", \"type\": \"quantitative\"}, \"y\": {\"aggregate\": \"median\", \"bin\": false, \"field\": \"wall_time\", \"type\": \"quantitative\"}}, \"mark\": \"line\", \"data\": {\"values\": [{\"approx_grad\": true, \"beta\": [-0.3377254481236899, 0.2301321067680867, 0.333146114043381, -0.20971334143549528, -0.30575032752005804, 0.3740567200034455, -0.06305069711196563, 0.18395187863790194, 0.12185236088423221, -0.5561521713524271], \"chunks\": 2000, \"d\": 10, \"max_iter\": 1, \"n\": 8000000, \"pointwise_loss\": 404087362.5696646, \"rel_error\": 0.5786328175760411, \"seed\": 0, \"sigma\": 7.0, \"wall_time\": 9.872261762619019}, {\"approx_grad\": true, \"beta\": [-0.4826127929824557, 0.3292431113973369, 0.3473824920275945, -0.15459457027121878, -0.19415953367371733, 0.42306454754726186, -0.17252527019265015, 0.05592581965829817, 0.15623894385874282, -0.43401916546021496], \"chunks\": 2000, \"d\": 10, \"max_iter\": 1, \"n\": 8000000, \"pointwise_loss\": 403110631.53552437, \"rel_error\": 0.5542360022939973, \"seed\": 1, \"sigma\": 7.0, \"wall_time\": 7.958230018615723}, {\"approx_grad\": true, \"beta\": [-0.3867231542110521, 0.18469768851413176, 0.2816929331607336, -0.054056285360848305, -0.22192282415811285, 0.33552478632731875, -0.11270602798687447, 0.009945650699710791, 0.1491485983684059, -0.5102915167728846], \"chunks\": 2000, \"d\": 10, \"max_iter\": 1, \"n\": 8000000, \"pointwise_loss\": 405328298.33456, \"rel_error\": 0.6073403494210304, \"seed\": 2, \"sigma\": 7.0, \"wall_time\": 7.640789985656738}, {\"approx_grad\": true, \"beta\": [-0.38046753748948453, 0.23782463200335902, 0.18564595937362763, -0.1889852272595575, -0.20634251350348562, 0.38498398172180837, -0.2799477671469581, 0.05463026794689662, 0.1997645016076342, -0.4443484968294167], \"chunks\": 2000, \"d\": 10, \"max_iter\": 1, \"n\": 8000000, \"pointwise_loss\": 405016186.76993334, \"rel_error\": 0.6002291017741715, \"seed\": 3, \"sigma\": 7.0, \"wall_time\": 9.718455076217651}, {\"approx_grad\": true, \"beta\": [-0.5025513159335334, 0.3680979167244689, 0.3038675340941493, -0.09045097996902372, -0.10998030666990562, 0.24256309845221122, -0.1839670635653993, 0.07396880196168064, 0.1942692091907179, -0.6048743288747124], \"chunks\": 2000, \"d\": 10, \"max_iter\": 1, \"n\": 8000000, \"pointwise_loss\": 403628905.38647205, \"rel_error\": 0.5669817044056502, \"seed\": 4, \"sigma\": 7.0, \"wall_time\": 9.272586107254028}, {\"approx_grad\": false, \"beta\": [-1.4148599575143694, 0.9848581358758047, 1.0303035090932562, -0.5169708152326576, -0.8651982685784295, 1.5001343392240916, -0.47727207669491456, 0.0171080426595288, 0.8173125569729787, -1.8128364003611475], \"chunks\": 2000, \"d\": 10, \"max_iter\": 1, \"n\": 8000000, \"pointwise_loss\": 405024480.2475509, \"rel_error\": 0.5995344874321145, \"seed\": 0, \"sigma\": 7.0, \"wall_time\": 88.51962184906006}, {\"approx_grad\": false, \"beta\": [-1.4148599575143694, 0.9848581358758047, 1.0303035090932562, -0.5169708152326576, -0.8651982685784295, 1.5001343392240916, -0.47727207669491456, 0.0171080426595288, 0.8173125569729787, -1.8128364003611475], \"chunks\": 2000, \"d\": 10, \"max_iter\": 1, \"n\": 8000000, \"pointwise_loss\": 405024480.2475509, \"rel_error\": 0.5995344874321145, \"seed\": 1, \"sigma\": 7.0, \"wall_time\": 72.5514748096466}, {\"approx_grad\": true, \"beta\": [-0.594331803280871, 0.43971841399477063, 0.5008066717190522, -0.24237453836988399, -0.4465450331221631, 0.654942040865184, -0.02191295874530111, 0.17804981679861503, 0.3397522624591079, -0.8989588947159972], \"chunks\": 2000, \"d\": 10, \"max_iter\": 2, \"n\": 8000000, \"pointwise_loss\": 395448737.1142796, \"rel_error\": 0.30623972828770346, \"seed\": 0, \"sigma\": 7.0, \"wall_time\": 10.498352766036987}, {\"approx_grad\": true, \"beta\": [-0.8128070379894679, 0.38924159720017654, 0.4144969044449924, -0.19841532859246383, -0.4710042101673059, 0.734437493524168, -0.2930097249427355, 0.034101664921965966, 0.5329805523482001, -0.9469222291091087], \"chunks\": 2000, \"d\": 10, \"max_iter\": 2, \"n\": 8000000, \"pointwise_loss\": 393745566.9921506, \"rel_error\": 0.2138264606829495, \"seed\": 1, \"sigma\": 7.0, \"wall_time\": 10.346577882766724}, {\"approx_grad\": true, \"beta\": [-0.6790975376476712, 0.5450407181605879, 0.5322319335883364, -0.14396416217709798, -0.5102201147023271, 0.7565770006251672, -0.11488155620503593, -0.03861813027138468, 0.3069642283910206, -0.9001416737560651], \"chunks\": 2000, \"d\": 10, \"max_iter\": 2, \"n\": 8000000, \"pointwise_loss\": 394154971.6913496, \"rel_error\": 0.23966802233459264, \"seed\": 2, \"sigma\": 7.0, \"wall_time\": 10.392645120620728}, {\"approx_grad\": true, \"beta\": [-0.811140159228611, 0.6276138384907568, 0.473085872816493, -0.1838171446679991, -0.40054954722887764, 0.7558401907634041, -0.42296970180408683, 0.06692302612485297, 0.24404777524362825, -0.9590326212179863], \"chunks\": 2000, \"d\": 10, \"max_iter\": 2, \"n\": 8000000, \"pointwise_loss\": 393915237.67307943, \"rel_error\": 0.2250417472016826, \"seed\": 3, \"sigma\": 7.0, \"wall_time\": 10.574932098388672}, {\"approx_grad\": true, \"beta\": [-0.8095071525177628, 0.39580708437907686, 0.5526693122538857, -0.30895738289937397, -0.4200459055558351, 0.6794734103735167, -0.14089336983588296, -0.042578922137064476, 0.37934092991144963, -0.994111196953386], \"chunks\": 2000, \"d\": 10, \"max_iter\": 2, \"n\": 8000000, \"pointwise_loss\": 393761112.8971651, \"rel_error\": 0.21548270480883275, \"seed\": 4, \"sigma\": 7.0, \"wall_time\": 11.583020210266113}, {\"approx_grad\": false, \"beta\": [-0.8838851099696825, 0.6156234649452773, 0.642426897118965, -0.32293019293604086, -0.541623528391047, 0.9373094310885044, -0.2978533442281764, 0.011540620405336708, 0.5121800257134298, -1.1326823717915948], \"chunks\": 2000, \"d\": 10, \"max_iter\": 2, \"n\": 8000000, \"pointwise_loss\": 392098738.7839557, \"rel_error\": 0.0031580153358463157, \"seed\": 0, \"sigma\": 7.0, \"wall_time\": 111.78914093971252}, {\"approx_grad\": false, \"beta\": [-0.8838851099696825, 0.6156234649452773, 0.642426897118965, -0.32293019293604086, -0.541623528391047, 0.9373094310885044, -0.2978533442281764, 0.011540620405336708, 0.5121800257134298, -1.1326823717915948], \"chunks\": 2000, \"d\": 10, \"max_iter\": 2, \"n\": 8000000, \"pointwise_loss\": 392098738.7839557, \"rel_error\": 0.0031580153358463157, \"seed\": 1, \"sigma\": 7.0, \"wall_time\": 115.92996788024902}, {\"approx_grad\": true, \"beta\": [-0.9046560108081807, 0.6326613755118812, 0.5504089312086718, -0.29349505832691863, -0.6703244988401252, 0.997889006713227, -0.2968343677535923, -0.00451966296877129, 0.6390163185945504, -1.1380061562065575], \"chunks\": 2000, \"d\": 10, \"max_iter\": 3, \"n\": 8000000, \"pointwise_loss\": 392474274.2668188, \"rel_error\": 0.10178542461115976, \"seed\": 0, \"sigma\": 7.0, \"wall_time\": 16.024066925048828}, {\"approx_grad\": true, \"beta\": [-0.9083172935353033, 0.7288582671124717, 0.5878503446288098, -0.2563745573662875, -0.5116395655415267, 0.8216666253332333, -0.2935442589357671, 0.006857256737447226, 0.36823156748019614, -1.066219790184447], \"chunks\": 2000, \"d\": 10, \"max_iter\": 3, \"n\": 8000000, \"pointwise_loss\": 392580710.7184185, \"rel_error\": 0.11609047177517823, \"seed\": 1, \"sigma\": 7.0, \"wall_time\": 17.255850076675415}, {\"approx_grad\": true, \"beta\": [-0.8389011499255122, 0.5948283477142385, 0.7238175294286159, -0.2997997038890462, -0.6591657691232473, 0.8288540320772136, -0.3125267990645646, 0.1449823186819636, 0.5704160565098788, -1.2720900725201265], \"chunks\": 2000, \"d\": 10, \"max_iter\": 3, \"n\": 8000000, \"pointwise_loss\": 392708106.3297756, \"rel_error\": 0.1317116829867297, \"seed\": 2, \"sigma\": 7.0, \"wall_time\": 15.214453935623169}, {\"approx_grad\": true, \"beta\": [-0.9414078608218889, 0.7179276231792979, 0.5555960772034156, -0.49420754272047746, -0.40115295600901557, 0.8294078054465174, -0.2042147327239715, -0.15308573285461974, 0.38000894740318536, -1.05377066222109], \"chunks\": 2000, \"d\": 10, \"max_iter\": 3, \"n\": 8000000, \"pointwise_loss\": 393229638.7225437, \"rel_error\": 0.17764704001228399, \"seed\": 3, \"sigma\": 7.0, \"wall_time\": 18.07921314239502}, {\"approx_grad\": true, \"beta\": [-0.926476804277389, 0.6474259797624344, 0.46509357607425766, -0.31031412639931977, -0.2912526589717621, 0.7152394416847999, -0.22762164094017698, 0.17662110468719983, 0.4446096993058229, -1.1348464288890303], \"chunks\": 2000, \"d\": 10, \"max_iter\": 3, \"n\": 8000000, \"pointwise_loss\": 393565654.71435046, \"rel_error\": 0.20169795281469685, \"seed\": 4, \"sigma\": 7.0, \"wall_time\": 16.644001722335815}, {\"approx_grad\": false, \"beta\": [-0.8840752372471408, 0.6155811769804287, 0.6431364460452087, -0.32301339020313813, -0.5412142113891235, 0.9374405298154759, -0.2980612954990996, 0.011143191960726204, 0.5115415216339614, -1.132843412924105], \"chunks\": 2000, \"d\": 10, \"max_iter\": 3, \"n\": 8000000, \"pointwise_loss\": 392098732.22278774, \"rel_error\": 0.0031039884258993638, \"seed\": 0, \"sigma\": 7.0, \"wall_time\": 145.35059690475464}, {\"approx_grad\": false, \"beta\": [-0.8840752372471408, 0.6155811769804287, 0.6431364460452087, -0.32301339020313813, -0.5412142113891235, 0.9374405298154759, -0.2980612954990996, 0.011143191960726204, 0.5115415216339614, -1.132843412924105], \"chunks\": 2000, \"d\": 10, \"max_iter\": 3, \"n\": 8000000, \"pointwise_loss\": 392098732.22278774, \"rel_error\": 0.0031039884258993633, \"seed\": 1, \"sigma\": 7.0, \"wall_time\": 147.1575231552124}, {\"approx_grad\": true, \"beta\": [-0.8271726726227692, 0.49951926049589246, 0.704512362421464, -0.2850862602450924, -0.6863328958031757, 0.8601722207246763, -0.508975427468857, -0.06062527242346647, 0.5160043265879157, -1.2795063115296572], \"chunks\": 2000, \"d\": 10, \"max_iter\": 4, \"n\": 8000000, \"pointwise_loss\": 393059855.81411487, \"rel_error\": 0.16445766469948245, \"seed\": 0, \"sigma\": 7.0, \"wall_time\": 22.655385971069336}, {\"approx_grad\": true, \"beta\": [-0.9616765072338722, 0.5343757451320267, 0.5564692898481249, -0.20815412080423262, -0.6599511653081934, 1.00051922660928, -0.47053635410121947, 0.01242918279858982, 0.5791592404773261, -1.2908888698588328], \"chunks\": 2000, \"d\": 10, \"max_iter\": 4, \"n\": 8000000, \"pointwise_loss\": 392983624.5057327, \"rel_error\": 0.156090826714074, \"seed\": 1, \"sigma\": 7.0, \"wall_time\": 19.843119859695435}, {\"approx_grad\": true, \"beta\": [-0.8311171789400179, 0.521410421564028, 0.9462671660740188, -0.3456457641986597, -0.29334977323977884, 0.9560068287397671, -0.22130619909174298, -0.0049895114959127105, 0.8629391239905513, -1.1745132638244187], \"chunks\": 2000, \"d\": 10, \"max_iter\": 4, \"n\": 8000000, \"pointwise_loss\": 394477000.17289954, \"rel_error\": 0.25672332004490195, \"seed\": 2, \"sigma\": 7.0, \"wall_time\": 20.78614115715027}, {\"approx_grad\": true, \"beta\": [-0.801139199063629, 0.16532940342659558, 0.6954539332675801, -0.28936401002679035, -0.7217519603155139, 1.091651439217816, -0.43199373306063593, -0.11066348276909367, 0.2855398064677888, -1.355992899941729], \"chunks\": 2000, \"d\": 10, \"max_iter\": 4, \"n\": 8000000, \"pointwise_loss\": 395327809.9076261, \"rel_error\": 0.3000645202606274, \"seed\": 3, \"sigma\": 7.0, \"wall_time\": 23.399425983428955}, {\"approx_grad\": true, \"beta\": [-0.7053576581148432, 0.6241930091984895, 0.5853791964035107, -0.2521215481878079, -0.8336804942073975, 0.7948488717641425, -0.278253920773729, 0.018089168528746924, 0.22836203225370016, -0.9739507392001161], \"chunks\": 2000, \"d\": 10, \"max_iter\": 4, \"n\": 8000000, \"pointwise_loss\": 394115898.34814155, \"rel_error\": 0.23886453574779987, \"seed\": 4, \"sigma\": 7.0, \"wall_time\": 22.683278799057007}, {\"approx_grad\": false, \"beta\": [-0.8840149290810264, 0.6155952392872667, 0.6429144576863511, -0.32298713855219047, -0.5413417425959787, 0.9373981467172503, -0.2979957802729954, 0.01126732427321426, 0.5117417330680154, -1.1327922152299434], \"chunks\": 2000, \"d\": 10, \"max_iter\": 4, \"n\": 8000000, \"pointwise_loss\": 392098731.9225391, \"rel_error\": 0.003110684443249633, \"seed\": 0, \"sigma\": 7.0, \"wall_time\": 189.257798910141}, {\"approx_grad\": false, \"beta\": [-0.8840149290810264, 0.6155952392872667, 0.6429144576863511, -0.32298713855219047, -0.5413417425959787, 0.9373981467172503, -0.2979957802729954, 0.01126732427321426, 0.5117417330680154, -1.1327922152299434], \"chunks\": 2000, \"d\": 10, \"max_iter\": 4, \"n\": 8000000, \"pointwise_loss\": 392098731.9225391, \"rel_error\": 0.003110684443249633, \"seed\": 1, \"sigma\": 7.0, \"wall_time\": 186.81083416938782}, {\"approx_grad\": true, \"beta\": [-0.9407374451959918, 0.5995469772485784, 0.3829421849180021, -0.5000865670359426, -0.5737165925062941, 1.1179590375704331, -0.05387964227593586, -0.062217309307136766, -0.002549105064196923, -0.8238232877423541], \"chunks\": 2000, \"d\": 10, \"max_iter\": 5, \"n\": 8000000, \"pointwise_loss\": 396586220.0473553, \"rel_error\": 0.3538455628010787, \"seed\": 0, \"sigma\": 7.0, \"wall_time\": 25.987020015716553}, {\"approx_grad\": true, \"beta\": [-0.4964414415005384, 0.6718447193669304, 0.6846026851661291, -0.010461025842725123, -0.34100874172096585, 0.9586468990831749, -0.21698443097406914, -0.13017291237255, 0.9125536530331779, -0.6087769509034419], \"chunks\": 2000, \"d\": 10, \"max_iter\": 5, \"n\": 8000000, \"pointwise_loss\": 398139813.413357, \"rel_error\": 0.4096640005375313, \"seed\": 1, \"sigma\": 7.0, \"wall_time\": 22.02472710609436}, {\"approx_grad\": true, \"beta\": [-0.8829626205083301, 1.1212609577680581, 0.5684551025155554, -0.6505691911598741, -0.9372315683339582, 0.842119326116806, -0.42196789947305235, -0.08570587191028647, 0.08368869163034408, -1.2792551410778739], \"chunks\": 2000, \"d\": 10, \"max_iter\": 5, \"n\": 8000000, \"pointwise_loss\": 398210025.845274, \"rel_error\": 0.41423200443526725, \"seed\": 2, \"sigma\": 7.0, \"wall_time\": 22.424404859542847}, {\"approx_grad\": true, \"beta\": [-0.8276390460755914, 0.6254867138609531, 0.3250671199876772, -0.17258921033454458, -0.19899367174226415, 0.802372686454787, -0.29783989269792505, 0.15952477773340476, 0.7608196944681033, -1.0561404959330822], \"chunks\": 2000, \"d\": 10, \"max_iter\": 5, \"n\": 8000000, \"pointwise_loss\": 394918666.8340963, \"rel_error\": 0.2788153092040237, \"seed\": 3, \"sigma\": 7.0, \"wall_time\": 23.654550075531006}, {\"approx_grad\": true, \"beta\": [-1.0663379690958243, 0.5137748168705915, 0.7515449969486085, -0.8673948569046709, -0.5042643587430429, 1.400751405530614, -0.45931404840107765, 0.13862134013387145, 0.3718303171564792, -1.255472840503655], \"chunks\": 2000, \"d\": 10, \"max_iter\": 5, \"n\": 8000000, \"pointwise_loss\": 397256096.00919074, \"rel_error\": 0.37940869669671184, \"seed\": 4, \"sigma\": 7.0, \"wall_time\": 30.06166672706604}, {\"approx_grad\": false, \"beta\": [-0.8840363068073565, 0.6155901299411422, 0.6429925519141578, -0.32299641680865165, -0.5412969800789815, 0.937413323810281, -0.298018917144369, 0.011223695205511832, 0.5116712128875367, -1.132810384813737], \"chunks\": 2000, \"d\": 10, \"max_iter\": 5, \"n\": 8000000, \"pointwise_loss\": 392098731.7830447, \"rel_error\": 0.003107192716169649, \"seed\": 0, \"sigma\": 7.0, \"wall_time\": 201.68372106552124}, {\"approx_grad\": false, \"beta\": [-0.8840363068073565, 0.6155901299411422, 0.6429925519141578, -0.32299641680865165, -0.5412969800789815, 0.937413323810281, -0.298018917144369, 0.011223695205511832, 0.5116712128875367, -1.132810384813737], \"chunks\": 2000, \"d\": 10, \"max_iter\": 5, \"n\": 8000000, \"pointwise_loss\": 392098731.7830447, \"rel_error\": 0.003107192716169649, \"seed\": 1, \"sigma\": 7.0, \"wall_time\": 204.56695413589478}, {\"approx_grad\": true, \"beta\": [-0.7871260042083935, 0.3961733101780085, 0.6958527709901753, -0.3142060076706436, -0.6376767465113462, 0.960881014904799, -0.31250840955083714, 0.08209674189561741, 0.515565851142813, -1.2308461475839017], \"chunks\": 2000, \"d\": 10, \"max_iter\": 6, \"n\": 8000000, \"pointwise_loss\": 392779349.549917, \"rel_error\": 0.13854581491536738, \"seed\": 0, \"sigma\": 7.0, \"wall_time\": 36.94024991989136}, {\"approx_grad\": true, \"beta\": [-0.6031158370785182, 0.6709487169693519, 0.7702163302750972, -0.4654554489878465, -0.7768143279733025, 0.9003608368757029, -0.37337880686709085, -0.07810376917934554, 0.5054273482702586, -1.1163010862552663], \"chunks\": 2000, \"d\": 10, \"max_iter\": 6, \"n\": 8000000, \"pointwise_loss\": 393612977.4438118, \"rel_error\": 0.20778431721208035, \"seed\": 1, \"sigma\": 7.0, \"wall_time\": 28.863558053970337}, {\"approx_grad\": true, \"beta\": [-0.8664624880186267, 0.4366107494098207, 0.7988147992751405, -0.2926785272602952, -0.5428020755237238, 0.8156204506426754, -0.19816820095111792, -0.034020062784066066, 0.37078322939922836, -1.137769387721782], \"chunks\": 2000, \"d\": 10, \"max_iter\": 6, \"n\": 8000000, \"pointwise_loss\": 392932772.7201239, \"rel_error\": 0.15320684314729843, \"seed\": 2, \"sigma\": 7.0, \"wall_time\": 28.86776614189148}, {\"approx_grad\": true, \"beta\": [-0.9508191916738734, 0.5803069168850657, 0.6081637705377964, -0.4316316463999406, -0.5012948159484267, 0.968552261019256, -0.24577223705663653, 0.06299812960488932, 0.2590370207057818, -1.252127732722464], \"chunks\": 2000, \"d\": 10, \"max_iter\": 6, \"n\": 8000000, \"pointwise_loss\": 392936926.844349, \"rel_error\": 0.15315234339841183, \"seed\": 3, \"sigma\": 7.0, \"wall_time\": 27.614052772521973}, {\"approx_grad\": true, \"beta\": [-0.8091028799637264, 0.45004047894098886, 0.7061596317906195, -0.3256998918927045, -0.49755334068387674, 0.9253875802460394, -0.11361957804608215, 0.24122491463255788, 0.5365126582799239, -1.1292519285384686], \"chunks\": 2000, \"d\": 10, \"max_iter\": 6, \"n\": 8000000, \"pointwise_loss\": 393111480.63514334, \"rel_error\": 0.1684587208741917, \"seed\": 4, \"sigma\": 7.0, \"wall_time\": 26.95098090171814}, {\"approx_grad\": false, \"beta\": [-0.8840368556424033, 0.6155900530326254, 0.6429948167902779, -0.3229966670664642, -0.541295637133796, 0.937413646700167, -0.2980195490623621, 0.011222412409941856, 0.5116692056516307, -1.132810842239684], \"chunks\": 2000, \"d\": 10, \"max_iter\": 6, \"n\": 8000000, \"pointwise_loss\": 392098731.78297484, \"rel_error\": 0.00310712691370173, \"seed\": 0, \"sigma\": 7.0, \"wall_time\": 229.23673105239868}, {\"approx_grad\": false, \"beta\": [-0.8840368556424033, 0.6155900530326254, 0.6429948167902779, -0.3229966670664642, -0.541295637133796, 0.937413646700167, -0.2980195490623621, 0.011222412409941856, 0.5116692056516307, -1.132810842239684], \"chunks\": 2000, \"d\": 10, \"max_iter\": 6, \"n\": 8000000, \"pointwise_loss\": 392098731.78297484, \"rel_error\": 0.00310712691370173, \"seed\": 1, \"sigma\": 7.0, \"wall_time\": 227.5106761455536}, {\"approx_grad\": true, \"beta\": [-0.9289146412636535, 0.589207394081429, 0.7012400145186803, -0.4208002014380348, -0.43386871861124676, 0.933089665396426, -0.4005966465091005, 0.06483459323652249, 0.5673006082227218, -1.225626019604088], \"chunks\": 2000, \"d\": 10, \"max_iter\": 10, \"n\": 8000000, \"pointwise_loss\": 392517421.8296241, \"rel_error\": 0.10793165451244967, \"seed\": 0, \"sigma\": 7.0, \"wall_time\": 52.241666078567505}, {\"approx_grad\": true, \"beta\": [-0.9743804887017291, 0.5410402380568015, 0.7704433890334813, -0.35185475181030945, -0.46356642753222593, 0.9481604622933278, -0.40504543584553954, 0.06890603470985716, 0.4346690052333042, -1.0354939209957497], \"chunks\": 2000, \"d\": 10, \"max_iter\": 10, \"n\": 8000000, \"pointwise_loss\": 392636046.76928383, \"rel_error\": 0.12258435738129043, \"seed\": 1, \"sigma\": 7.0, \"wall_time\": 49.099812269210815}, {\"approx_grad\": true, \"beta\": [-0.9691976834229626, 0.6475594312600076, 0.7508500065046348, -0.32858753685093967, -0.5462316324992845, 0.8922110086481335, -0.29444047752987906, -0.007789285206273814, 0.5428055414686452, -1.0599010679159921], \"chunks\": 2000, \"d\": 10, \"max_iter\": 10, \"n\": 8000000, \"pointwise_loss\": 392327909.8895484, \"rel_error\": 0.08030740531555673, \"seed\": 2, \"sigma\": 7.0, \"wall_time\": 53.91449022293091}, {\"approx_grad\": true, \"beta\": [-0.910264983109954, 0.5498026119249138, 0.7398351291370956, -0.29663314818807934, -0.6484180771824207, 0.9684016736573872, -0.10330011565812211, 0.05913743859741222, 0.7403852268345604, -1.1810385316788785], \"chunks\": 2000, \"d\": 10, \"max_iter\": 10, \"n\": 8000000, \"pointwise_loss\": 393077610.3466072, \"rel_error\": 0.16494774361606168, \"seed\": 3, \"sigma\": 7.0, \"wall_time\": 56.24317216873169}, {\"approx_grad\": true, \"beta\": [-0.8675186275412994, 0.4551384495672032, 0.604807568567115, -0.2831505725214636, -0.5860885866345346, 0.8674112928558778, -0.416804818232409, -0.1134532111679631, 0.38210984739072146, -1.075781297833849], \"chunks\": 2000, \"d\": 10, \"max_iter\": 10, \"n\": 8000000, \"pointwise_loss\": 392784311.2327362, \"rel_error\": 0.1385175667205294, \"seed\": 4, \"sigma\": 7.0, \"wall_time\": 62.55457830429077}, {\"approx_grad\": false, \"beta\": [-0.8840367015486226, 0.6155900769031977, 0.6429941913527001, -0.3229965972154067, -0.5412960062534135, 0.9374135532285174, -0.2980193731251264, 0.011222765961551154, 0.511669761456385, -1.1328107133068281], \"chunks\": 2000, \"d\": 10, \"max_iter\": 10, \"n\": 8000000, \"pointwise_loss\": 392098731.7829722, \"rel_error\": 0.003107145577421732, \"seed\": 0, \"sigma\": 7.0, \"wall_time\": 301.74830627441406}, {\"approx_grad\": false, \"beta\": [-0.8840367015486226, 0.6155900769031977, 0.6429941913527001, -0.3229965972154067, -0.5412960062534135, 0.9374135532285174, -0.2980193731251264, 0.011222765961551154, 0.511669761456385, -1.1328107133068281], \"chunks\": 2000, \"d\": 10, \"max_iter\": 10, \"n\": 8000000, \"pointwise_loss\": 392098731.7829722, \"rel_error\": 0.003107145577421732, \"seed\": 1, \"sigma\": 7.0, \"wall_time\": 283.68756103515625}, {\"approx_grad\": true, \"beta\": [-0.879226020902641, 0.7785837679404635, 0.5854598664790497, -0.2652160310122725, -0.5033963847997842, 0.9101222961481134, -0.3230869875520656, 0.0726667810267524, 0.6090404727113902, -1.015491282425678], \"chunks\": 2000, \"d\": 10, \"max_iter\": 15, \"n\": 8000000, \"pointwise_loss\": 392603554.26582336, \"rel_error\": 0.1182117141794517, \"seed\": 0, \"sigma\": 7.0, \"wall_time\": 88.87447690963745}, {\"approx_grad\": true, \"beta\": [-0.9264615710300912, 0.5793376114442538, 0.5459787720950554, -0.3485187459810788, -0.737088753955497, 1.0098549802053038, -0.2697041886515856, -0.15363978801505496, 0.38003248055376626, -1.2006135031985268], \"chunks\": 2000, \"d\": 10, \"max_iter\": 15, \"n\": 8000000, \"pointwise_loss\": 392952025.853789, \"rel_error\": 0.15435494003660002, \"seed\": 1, \"sigma\": 7.0, \"wall_time\": 82.69940185546875}, {\"approx_grad\": true, \"beta\": [-0.9206027132484186, 0.739878760663106, 0.7270095070520408, -0.38870409165038816, -0.611151472242097, 0.715224504550175, -0.441189414563162, 0.11140238970762492, 0.5215692165038381, -1.0797497203817892], \"chunks\": 2000, \"d\": 10, \"max_iter\": 15, \"n\": 8000000, \"pointwise_loss\": 393025962.4803908, \"rel_error\": 0.1625869001127405, \"seed\": 2, \"sigma\": 7.0, \"wall_time\": 92.84537315368652}, {\"approx_grad\": true, \"beta\": [-0.787874281047773, 0.6799359185856565, 0.5416287983104555, -0.37731162844352173, -0.5493588312272013, 0.8054093051255071, -0.41572109145020375, -0.10753865241218799, 0.5014026377711878, -1.027600255568214], \"chunks\": 2000, \"d\": 10, \"max_iter\": 15, \"n\": 8000000, \"pointwise_loss\": 392764542.0845498, \"rel_error\": 0.13723331555685436, \"seed\": 3, \"sigma\": 7.0, \"wall_time\": 90.04433107376099}, {\"approx_grad\": true, \"beta\": [-0.8744565169107531, 0.604436206417696, 0.46990150193534835, -0.2077964962204974, -0.5355731471077425, 0.7940482312493866, -0.2473564612549025, -0.13633621899376802, 0.5697639944472284, -1.1292764001824989], \"chunks\": 2000, \"d\": 10, \"max_iter\": 15, \"n\": 8000000, \"pointwise_loss\": 392833172.4833296, \"rel_error\": 0.14196395307157753, \"seed\": 4, \"sigma\": 7.0, \"wall_time\": 80.92051577568054}, {\"approx_grad\": false, \"beta\": [-0.8840367015486226, 0.6155900769031977, 0.6429941913527001, -0.3229965972154067, -0.5412960062534135, 0.9374135532285174, -0.2980193731251264, 0.011222765961551154, 0.511669761456385, -1.1328107133068281], \"chunks\": 2000, \"d\": 10, \"max_iter\": 15, \"n\": 8000000, \"pointwise_loss\": 392098731.7829722, \"rel_error\": 0.003107145577421732, \"seed\": 0, \"sigma\": 7.0, \"wall_time\": 265.0206460952759}, {\"approx_grad\": false, \"beta\": [-0.8840367015486226, 0.6155900769031977, 0.6429941913527001, -0.3229965972154067, -0.5412960062534135, 0.9374135532285174, -0.2980193731251264, 0.011222765961551154, 0.511669761456385, -1.1328107133068281], \"chunks\": 2000, \"d\": 10, \"max_iter\": 15, \"n\": 8000000, \"pointwise_loss\": 392098731.7829722, \"rel_error\": 0.003107145577421732, \"seed\": 1, \"sigma\": 7.0, \"wall_time\": 259.3626790046692}, {\"approx_grad\": true, \"beta\": [-0.9042011637589843, 0.7029405553155955, 0.7110057308292758, -0.3704724339708527, -0.43615660607115325, 0.9149258394399113, -0.2950345455604007, 0.12055345649464859, 0.5209865340313132, -0.9256352343235037], \"chunks\": 2000, \"d\": 10, \"max_iter\": 20, \"n\": 8000000, \"pointwise_loss\": 392750046.25591063, \"rel_error\": 0.13527978657110445, \"seed\": 0, \"sigma\": 7.0, \"wall_time\": 108.67117500305176}, {\"approx_grad\": true, \"beta\": [-0.9719847430972252, 0.561696376565175, 0.7420408719990494, -0.34829215067329033, -0.46395348346167564, 1.0175211166551112, -0.5115341156758816, 0.1059482707651871, 0.46804927489393644, -1.2154401929156522], \"chunks\": 2000, \"d\": 10, \"max_iter\": 20, \"n\": 8000000, \"pointwise_loss\": 392873611.206111, \"rel_error\": 0.14661583431505454, \"seed\": 1, \"sigma\": 7.0, \"wall_time\": 117.8687379360199}, {\"approx_grad\": true, \"beta\": [-0.7238922364851894, 0.48440091636092486, 0.5483989511032223, -0.28848806460103726, -0.5291904133328238, 0.966457907630273, -0.4282109825815653, 0.09406156210969836, 0.5318905979292694, -1.1103801072529718], \"chunks\": 2000, \"d\": 10, \"max_iter\": 20, \"n\": 8000000, \"pointwise_loss\": 392728243.39322346, \"rel_error\": 0.1327308431752905, \"seed\": 2, \"sigma\": 7.0, \"wall_time\": 114.08178377151489}, {\"approx_grad\": true, \"beta\": [-0.9438060751341817, 0.5639914647299185, 0.5472097964866391, -0.42914470178028863, -0.4964936943584284, 0.8407517247734084, -0.276133337484286, -0.0008974746619392729, 0.6534416631567372, -1.1682447312995101], \"chunks\": 2000, \"d\": 10, \"max_iter\": 20, \"n\": 8000000, \"pointwise_loss\": 392578910.63095534, \"rel_error\": 0.11538053946876133, \"seed\": 3, \"sigma\": 7.0, \"wall_time\": 117.60523080825806}, {\"approx_grad\": true, \"beta\": [-0.9221378142827324, 0.7839502847556458, 0.48953742742167167, -0.3521669540471849, -0.6727637426480056, 0.9333889365145457, -0.2866852634483545, 0.05098493792076283, 0.5652542241558753, -1.1840084441981549], \"chunks\": 2000, \"d\": 10, \"max_iter\": 20, \"n\": 8000000, \"pointwise_loss\": 392728530.95368534, \"rel_error\": 0.1327486638410105, \"seed\": 4, \"sigma\": 7.0, \"wall_time\": 118.73969268798828}, {\"approx_grad\": false, \"beta\": [-0.8840367015486226, 0.6155900769031977, 0.6429941913527001, -0.3229965972154067, -0.5412960062534135, 0.9374135532285174, -0.2980193731251264, 0.011222765961551154, 0.511669761456385, -1.1328107133068281], \"chunks\": 2000, \"d\": 10, \"max_iter\": 20, \"n\": 8000000, \"pointwise_loss\": 392098731.7829722, \"rel_error\": 0.003107145577421732, \"seed\": 0, \"sigma\": 7.0, \"wall_time\": 267.0480582714081}, {\"approx_grad\": false, \"beta\": [-0.8840367015486226, 0.6155900769031977, 0.6429941913527001, -0.3229965972154067, -0.5412960062534135, 0.9374135532285174, -0.2980193731251264, 0.011222765961551154, 0.511669761456385, -1.1328107133068281], \"chunks\": 2000, \"d\": 10, \"max_iter\": 20, \"n\": 8000000, \"pointwise_loss\": 392098731.7829722, \"rel_error\": 0.003107145577421732, \"seed\": 1, \"sigma\": 7.0, \"wall_time\": 262.2125267982483}, {\"approx_grad\": true, \"beta\": [-0.8829139465236511, 0.625917698868711, 0.5610035845124403, -0.29290341592114466, -0.5451773218871103, 0.9795265198140615, -0.2515854796694451, 3.103563452222811e-05, 0.5432008109787038, -1.100514723843423], \"chunks\": 2000, \"d\": 10, \"max_iter\": 30, \"n\": 8000000, \"pointwise_loss\": 392209513.1652502, \"rel_error\": 0.05404568835450826, \"seed\": 0, \"sigma\": 7.0, \"wall_time\": 195.57359433174133}, {\"approx_grad\": true, \"beta\": [-0.8635894856734007, 0.6248311725057915, 0.630928540527987, -0.3554994869075913, -0.5729866601651461, 0.950780176643394, -0.33218680175151266, -0.031223046549164638, 0.5121029562335224, -1.070479358507842], \"chunks\": 2000, \"d\": 10, \"max_iter\": 30, \"n\": 8000000, \"pointwise_loss\": 392176658.7978555, \"rel_error\": 0.04775019218448381, \"seed\": 1, \"sigma\": 7.0, \"wall_time\": 185.2518310546875}, {\"approx_grad\": true, \"beta\": [-0.854022071521578, 0.6031780361923385, 0.6422737722072802, -0.36969844012775666, -0.6124105215693041, 0.8853708036980525, -0.2445671723441642, 0.012866940525705072, 0.6060871681055311, -1.1131080448188513], \"chunks\": 2000, \"d\": 10, \"max_iter\": 30, \"n\": 8000000, \"pointwise_loss\": 392284167.6319425, \"rel_error\": 0.07326080170913572, \"seed\": 2, \"sigma\": 7.0, \"wall_time\": 195.97774267196655}, {\"approx_grad\": true, \"beta\": [-0.8827942112640698, 0.6166268364806528, 0.5828977133369241, -0.2929787420823192, -0.6212947554007182, 0.9072455211672876, -0.2748304348371266, 0.060147878868150505, 0.5433172421187823, -1.1121212867854486], \"chunks\": 2000, \"d\": 10, \"max_iter\": 30, \"n\": 8000000, \"pointwise_loss\": 392228262.4553793, \"rel_error\": 0.06074344122218579, \"seed\": 3, \"sigma\": 7.0, \"wall_time\": 206.70327305793762}, {\"approx_grad\": true, \"beta\": [-0.8629285267307552, 0.5671430205245978, 0.5939363055152163, -0.3136346831101807, -0.48504145478632693, 0.9055869404266276, -0.3386675928486884, 0.024980442735857315, 0.49453810816927624, -1.0657688750768044], \"chunks\": 2000, \"d\": 10, \"max_iter\": 30, \"n\": 8000000, \"pointwise_loss\": 392227507.05134, \"rel_error\": 0.059750956746531476, \"seed\": 4, \"sigma\": 7.0, \"wall_time\": 202.11913299560547}, {\"approx_grad\": false, \"beta\": [-0.8840367015486226, 0.6155900769031977, 0.6429941913527001, -0.3229965972154067, -0.5412960062534135, 0.9374135532285174, -0.2980193731251264, 0.011222765961551154, 0.511669761456385, -1.1328107133068281], \"chunks\": 2000, \"d\": 10, \"max_iter\": 30, \"n\": 8000000, \"pointwise_loss\": 392098731.7829722, \"rel_error\": 0.003107145577421732, \"seed\": 0, \"sigma\": 7.0, \"wall_time\": 256.4530289173126}, {\"approx_grad\": false, \"beta\": [-0.8840367015486226, 0.6155900769031977, 0.6429941913527001, -0.3229965972154067, -0.5412960062534135, 0.9374135532285174, -0.2980193731251264, 0.011222765961551154, 0.511669761456385, -1.1328107133068281], \"chunks\": 2000, \"d\": 10, \"max_iter\": 30, \"n\": 8000000, \"pointwise_loss\": 392098731.7829722, \"rel_error\": 0.003107145577421732, \"seed\": 1, \"sigma\": 7.0, \"wall_time\": 271.3524467945099}, {\"approx_grad\": true, \"beta\": [-0.9144115552180049, 0.603015498404459, 0.5713064287707974, -0.34545575275615004, -0.6424885972652146, 0.8931970146541598, -0.2331797791375983, -0.02907522461562066, 0.46925773932688525, -1.1219401556826756], \"chunks\": 2000, \"d\": 10, \"max_iter\": 40, \"n\": 8000000, \"pointwise_loss\": 392312074.65794194, \"rel_error\": 0.0779374005493023, \"seed\": 0, \"sigma\": 7.0, \"wall_time\": 264.6048061847687}, {\"approx_grad\": true, \"beta\": [-0.8973887214484578, 0.625628029190906, 0.6978309024390227, -0.32447667279242576, -0.5497618115288487, 0.9206360070908075, -0.2452788563571229, -0.02033314724967066, 0.4922056424940327, -1.0984816585576087], \"chunks\": 2000, \"d\": 10, \"max_iter\": 40, \"n\": 8000000, \"pointwise_loss\": 392170503.4222042, \"rel_error\": 0.04541517197333465, \"seed\": 1, \"sigma\": 7.0, \"wall_time\": 269.0821568965912}, {\"approx_grad\": true, \"beta\": [-0.8823393235027647, 0.5492224550551256, 0.7019713370961441, -0.3435038812509715, -0.5853327207643598, 0.9835950273865081, -0.27844770423239806, -0.055921215247640724, 0.4887343295786643, -1.1194613758252137], \"chunks\": 2000, \"d\": 10, \"max_iter\": 40, \"n\": 8000000, \"pointwise_loss\": 392242635.91384065, \"rel_error\": 0.06373818373096356, \"seed\": 2, \"sigma\": 7.0, \"wall_time\": 256.9263689517975}, {\"approx_grad\": true, \"beta\": [-0.8279929314291558, 0.5785406092292175, 0.6888273335787916, -0.354480652670551, -0.5278944818518453, 0.9341898266525525, -0.28833210613646065, 0.12581324637761054, 0.47795866600785925, -1.159984780697032], \"chunks\": 2000, \"d\": 10, \"max_iter\": 40, \"n\": 8000000, \"pointwise_loss\": 392281900.7215971, \"rel_error\": 0.0729637904656608, \"seed\": 3, \"sigma\": 7.0, \"wall_time\": 243.45107984542847}, {\"approx_grad\": true, \"beta\": [-0.9078119382540301, 0.5449053559924391, 0.6857141065778634, -0.33039721408797773, -0.546340031592623, 0.9589346749302237, -0.2378234765841817, 0.01719609554537357, 0.4794984490149854, -1.1048310139934197], \"chunks\": 2000, \"d\": 10, \"max_iter\": 40, \"n\": 8000000, \"pointwise_loss\": 392205987.4079259, \"rel_error\": 0.05482705526360315, \"seed\": 4, \"sigma\": 7.0, \"wall_time\": 247.5673270225525}, {\"approx_grad\": false, \"beta\": [-0.8840367015486226, 0.6155900769031977, 0.6429941913527001, -0.3229965972154067, -0.5412960062534135, 0.9374135532285174, -0.2980193731251264, 0.011222765961551154, 0.511669761456385, -1.1328107133068281], \"chunks\": 2000, \"d\": 10, \"max_iter\": 40, \"n\": 8000000, \"pointwise_loss\": 392098731.7829722, \"rel_error\": 0.003107145577421732, \"seed\": 0, \"sigma\": 7.0, \"wall_time\": 261.3412778377533}, {\"approx_grad\": false, \"beta\": [-0.8840367015486226, 0.6155900769031977, 0.6429941913527001, -0.3229965972154067, -0.5412960062534135, 0.9374135532285174, -0.2980193731251264, 0.011222765961551154, 0.511669761456385, -1.1328107133068281], \"chunks\": 2000, \"d\": 10, \"max_iter\": 40, \"n\": 8000000, \"pointwise_loss\": 392098731.7829722, \"rel_error\": 0.003107145577421732, \"seed\": 1, \"sigma\": 7.0, \"wall_time\": 243.74875688552856}, {\"approx_grad\": true, \"beta\": [-0.8854073087268456, 0.6172178631773021, 0.6450145855969107, -0.32189737380866645, -0.5447108173125104, 0.9348775941459327, -0.29811605878941355, 0.013981237885834294, 0.5143839539544327, -1.1370906424701517], \"chunks\": 2000, \"d\": 10, \"max_iter\": 80, \"n\": 8000000, \"pointwise_loss\": 392099221.5401707, \"rel_error\": 0.005300432688336602, \"seed\": 0, \"sigma\": 7.0, \"wall_time\": 925.5044710636139}, {\"approx_grad\": true, \"beta\": [-0.8818837818078422, 0.613442176739526, 0.6419890655824715, -0.32382572208301896, -0.5415035086126159, 0.9344285595975396, -0.29847295104227245, 0.00894236421481264, 0.5093510115674734, -1.1237886908468033], \"chunks\": 2000, \"d\": 10, \"max_iter\": 80, \"n\": 8000000, \"pointwise_loss\": 392099628.5426975, \"rel_error\": 0.006657967894667196, \"seed\": 1, \"sigma\": 7.0, \"wall_time\": 1053.5675048828125}, {\"approx_grad\": true, \"beta\": [-0.8841524006885965, 0.6192790187917091, 0.6435924079837052, -0.3224913282377281, -0.5405742823931305, 0.9331562752015514, -0.303804688437872, 0.012429471065645424, 0.5122031575325743, -1.1330440943007851], \"chunks\": 2000, \"d\": 10, \"max_iter\": 80, \"n\": 8000000, \"pointwise_loss\": 392099277.1428425, \"rel_error\": 0.005527006304765571, \"seed\": 2, \"sigma\": 7.0, \"wall_time\": 1059.2264840602875}, {\"approx_grad\": true, \"beta\": [-0.8825335911850323, 0.6098598438138241, 0.6441103657737104, -0.3210903115661359, -0.5390939638365305, 0.933140324813863, -0.2976977255708058, 0.012605030831765042, 0.5152813665905337, -1.1273924354863225], \"chunks\": 2000, \"d\": 10, \"max_iter\": 80, \"n\": 8000000, \"pointwise_loss\": 392099592.01777464, \"rel_error\": 0.005867858009065959, \"seed\": 3, \"sigma\": 7.0, \"wall_time\": 1028.4987251758575}, {\"approx_grad\": true, \"beta\": [-0.8871790409057009, 0.6196366303695788, 0.6431369705966894, -0.3268682376868445, -0.5390371654462093, 0.9393368181590982, -0.294816062643927, 0.010144475140230992, 0.5120191693819195, -1.1346528130151012], \"chunks\": 2000, \"d\": 10, \"max_iter\": 80, \"n\": 8000000, \"pointwise_loss\": 392099251.88777584, \"rel_error\": 0.004409698151563556, \"seed\": 4, \"sigma\": 7.0, \"wall_time\": 1020.5811431407928}, {\"approx_grad\": false, \"beta\": [-0.8840367015486226, 0.6155900769031977, 0.6429941913527001, -0.3229965972154067, -0.5412960062534135, 0.9374135532285174, -0.2980193731251264, 0.011222765961551154, 0.511669761456385, -1.1328107133068281], \"chunks\": 2000, \"d\": 10, \"max_iter\": 80, \"n\": 8000000, \"pointwise_loss\": 392098731.7829722, \"rel_error\": 0.003107145577421732, \"seed\": 0, \"sigma\": 7.0, \"wall_time\": 273.7842438220978}, {\"approx_grad\": false, \"beta\": [-0.8840367015486226, 0.6155900769031977, 0.6429941913527001, -0.3229965972154067, -0.5412960062534135, 0.9374135532285174, -0.2980193731251264, 0.011222765961551154, 0.511669761456385, -1.1328107133068281], \"chunks\": 2000, \"d\": 10, \"max_iter\": 80, \"n\": 8000000, \"pointwise_loss\": 392098731.7829722, \"rel_error\": 0.003107145577421732, \"seed\": 1, \"sigma\": 7.0, \"wall_time\": 267.38317227363586}]}};\n",
"var selector = \"#97805eea-eb5e-4b51-b2b9-bc9c98c81a05\";\n",
"var type = \"vega-lite\";\n",
"\n",
"var output_area = this;\n",
"require(['nbextensions/jupyter-vega/index'], function(vega) {\n",
" vega.render(selector, spec, type, output_area);\n",
"}, function (err) {\n",
" if (err.requireType !== 'scripterror') {\n",
" throw(err);\n",
" }\n",
"});\n"
]
},
"metadata": {
"jupyter-vega": "#97805eea-eb5e-4b51-b2b9-bc9c98c81a05"
},
"output_type": "display_data"
},
{
"data": {
"image/png": "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"
},
"metadata": {
"jupyter-vega": "#97805eea-eb5e-4b51-b2b9-bc9c98c81a05"
},
"output_type": "display_data"
}
],
"source": [
"from altair import Chart, Y, Scale\n",
"Chart(df).mark_line().encode(\n",
" x='max_iter', y=Y('median(wall_time)', bin=False), color='approx_grad')"
]
},
{
"cell_type": "code",
"execution_count": 9,
"metadata": {
"ExecuteTime": {
"end_time": "2017-09-01T19:51:15.918830Z",
"start_time": "2017-09-01T19:51:15.886095Z"
}
},
"outputs": [
{
"data": {
"text/html": [
"<div class=\"vega-embed\" id=\"7d6838e3-87a2-41de-b65c-bbd065e82350\"></div>\n",
"\n",
"<style>\n",
".vega-embed svg, .vega-embed canvas {\n",
" border: 1px dotted gray;\n",
"}\n",
"\n",
".vega-embed .vega-actions a {\n",
" margin-right: 6px;\n",
"}\n",
"</style>\n"
]
},
"metadata": {
"jupyter-vega": "#7d6838e3-87a2-41de-b65c-bbd065e82350"
},
"output_type": "display_data"
},
{
"data": {
"application/javascript": [
"var spec = {\"config\": {\"cell\": {\"width\": 500, \"height\": 350}}, \"encoding\": {\"color\": {\"field\": \"approx_grad\", \"type\": \"nominal\"}, \"x\": {\"field\": \"wall_time\", \"type\": \"quantitative\"}, \"y\": {\"field\": \"rel_error\", \"scale\": {\"type\": \"log\"}, \"type\": \"quantitative\"}}, \"mark\": \"line\", \"data\": {\"values\": [{\"approx_grad\": true, \"beta\": [-0.3377254481236899, 0.2301321067680867, 0.333146114043381, -0.20971334143549528, -0.30575032752005804, 0.3740567200034455, -0.06305069711196563, 0.18395187863790194, 0.12185236088423221, -0.5561521713524271], \"chunks\": 2000, \"d\": 10, \"max_iter\": 1, \"n\": 8000000, \"pointwise_loss\": 404087362.5696646, \"rel_error\": 0.5786328175760411, \"seed\": 0, \"sigma\": 7.0, \"wall_time\": 9.872261762619019}, {\"approx_grad\": true, \"beta\": [-0.4826127929824557, 0.3292431113973369, 0.3473824920275945, -0.15459457027121878, -0.19415953367371733, 0.42306454754726186, -0.17252527019265015, 0.05592581965829817, 0.15623894385874282, -0.43401916546021496], \"chunks\": 2000, \"d\": 10, \"max_iter\": 1, \"n\": 8000000, \"pointwise_loss\": 403110631.53552437, \"rel_error\": 0.5542360022939973, \"seed\": 1, \"sigma\": 7.0, \"wall_time\": 7.958230018615723}, {\"approx_grad\": true, \"beta\": [-0.3867231542110521, 0.18469768851413176, 0.2816929331607336, -0.054056285360848305, -0.22192282415811285, 0.33552478632731875, -0.11270602798687447, 0.009945650699710791, 0.1491485983684059, -0.5102915167728846], \"chunks\": 2000, \"d\": 10, \"max_iter\": 1, \"n\": 8000000, \"pointwise_loss\": 405328298.33456, \"rel_error\": 0.6073403494210304, \"seed\": 2, \"sigma\": 7.0, \"wall_time\": 7.640789985656738}, {\"approx_grad\": true, \"beta\": [-0.38046753748948453, 0.23782463200335902, 0.18564595937362763, -0.1889852272595575, -0.20634251350348562, 0.38498398172180837, -0.2799477671469581, 0.05463026794689662, 0.1997645016076342, -0.4443484968294167], \"chunks\": 2000, \"d\": 10, \"max_iter\": 1, \"n\": 8000000, \"pointwise_loss\": 405016186.76993334, \"rel_error\": 0.6002291017741715, \"seed\": 3, \"sigma\": 7.0, \"wall_time\": 9.718455076217651}, {\"approx_grad\": true, \"beta\": [-0.5025513159335334, 0.3680979167244689, 0.3038675340941493, -0.09045097996902372, -0.10998030666990562, 0.24256309845221122, -0.1839670635653993, 0.07396880196168064, 0.1942692091907179, -0.6048743288747124], \"chunks\": 2000, \"d\": 10, \"max_iter\": 1, \"n\": 8000000, \"pointwise_loss\": 403628905.38647205, \"rel_error\": 0.5669817044056502, \"seed\": 4, \"sigma\": 7.0, \"wall_time\": 9.272586107254028}, {\"approx_grad\": false, \"beta\": [-1.4148599575143694, 0.9848581358758047, 1.0303035090932562, -0.5169708152326576, -0.8651982685784295, 1.5001343392240916, -0.47727207669491456, 0.0171080426595288, 0.8173125569729787, -1.8128364003611475], \"chunks\": 2000, \"d\": 10, \"max_iter\": 1, \"n\": 8000000, \"pointwise_loss\": 405024480.2475509, \"rel_error\": 0.5995344874321145, \"seed\": 0, \"sigma\": 7.0, \"wall_time\": 88.51962184906006}, {\"approx_grad\": false, \"beta\": [-1.4148599575143694, 0.9848581358758047, 1.0303035090932562, -0.5169708152326576, -0.8651982685784295, 1.5001343392240916, -0.47727207669491456, 0.0171080426595288, 0.8173125569729787, -1.8128364003611475], \"chunks\": 2000, \"d\": 10, \"max_iter\": 1, \"n\": 8000000, \"pointwise_loss\": 405024480.2475509, \"rel_error\": 0.5995344874321145, \"seed\": 1, \"sigma\": 7.0, \"wall_time\": 72.5514748096466}, {\"approx_grad\": true, \"beta\": [-0.594331803280871, 0.43971841399477063, 0.5008066717190522, -0.24237453836988399, -0.4465450331221631, 0.654942040865184, -0.02191295874530111, 0.17804981679861503, 0.3397522624591079, -0.8989588947159972], \"chunks\": 2000, \"d\": 10, \"max_iter\": 2, \"n\": 8000000, \"pointwise_loss\": 395448737.1142796, \"rel_error\": 0.30623972828770346, \"seed\": 0, \"sigma\": 7.0, \"wall_time\": 10.498352766036987}, {\"approx_grad\": true, \"beta\": [-0.8128070379894679, 0.38924159720017654, 0.4144969044449924, -0.19841532859246383, -0.4710042101673059, 0.734437493524168, -0.2930097249427355, 0.034101664921965966, 0.5329805523482001, -0.9469222291091087], \"chunks\": 2000, \"d\": 10, \"max_iter\": 2, \"n\": 8000000, \"pointwise_loss\": 393745566.9921506, \"rel_error\": 0.2138264606829495, \"seed\": 1, \"sigma\": 7.0, \"wall_time\": 10.346577882766724}, {\"approx_grad\": true, \"beta\": [-0.6790975376476712, 0.5450407181605879, 0.5322319335883364, -0.14396416217709798, -0.5102201147023271, 0.7565770006251672, -0.11488155620503593, -0.03861813027138468, 0.3069642283910206, -0.9001416737560651], \"chunks\": 2000, \"d\": 10, \"max_iter\": 2, \"n\": 8000000, \"pointwise_loss\": 394154971.6913496, \"rel_error\": 0.23966802233459264, \"seed\": 2, \"sigma\": 7.0, \"wall_time\": 10.392645120620728}, {\"approx_grad\": true, \"beta\": [-0.811140159228611, 0.6276138384907568, 0.473085872816493, -0.1838171446679991, -0.40054954722887764, 0.7558401907634041, -0.42296970180408683, 0.06692302612485297, 0.24404777524362825, -0.9590326212179863], \"chunks\": 2000, \"d\": 10, \"max_iter\": 2, \"n\": 8000000, \"pointwise_loss\": 393915237.67307943, \"rel_error\": 0.2250417472016826, \"seed\": 3, \"sigma\": 7.0, \"wall_time\": 10.574932098388672}, {\"approx_grad\": true, \"beta\": [-0.8095071525177628, 0.39580708437907686, 0.5526693122538857, -0.30895738289937397, -0.4200459055558351, 0.6794734103735167, -0.14089336983588296, -0.042578922137064476, 0.37934092991144963, -0.994111196953386], \"chunks\": 2000, \"d\": 10, \"max_iter\": 2, \"n\": 8000000, \"pointwise_loss\": 393761112.8971651, \"rel_error\": 0.21548270480883275, \"seed\": 4, \"sigma\": 7.0, \"wall_time\": 11.583020210266113}, {\"approx_grad\": false, \"beta\": [-0.8838851099696825, 0.6156234649452773, 0.642426897118965, -0.32293019293604086, -0.541623528391047, 0.9373094310885044, -0.2978533442281764, 0.011540620405336708, 0.5121800257134298, -1.1326823717915948], \"chunks\": 2000, \"d\": 10, \"max_iter\": 2, \"n\": 8000000, \"pointwise_loss\": 392098738.7839557, \"rel_error\": 0.0031580153358463157, \"seed\": 0, \"sigma\": 7.0, \"wall_time\": 111.78914093971252}, {\"approx_grad\": false, \"beta\": [-0.8838851099696825, 0.6156234649452773, 0.642426897118965, -0.32293019293604086, -0.541623528391047, 0.9373094310885044, -0.2978533442281764, 0.011540620405336708, 0.5121800257134298, -1.1326823717915948], \"chunks\": 2000, \"d\": 10, \"max_iter\": 2, \"n\": 8000000, \"pointwise_loss\": 392098738.7839557, \"rel_error\": 0.0031580153358463157, \"seed\": 1, \"sigma\": 7.0, \"wall_time\": 115.92996788024902}, {\"approx_grad\": true, \"beta\": [-0.9046560108081807, 0.6326613755118812, 0.5504089312086718, -0.29349505832691863, -0.6703244988401252, 0.997889006713227, -0.2968343677535923, -0.00451966296877129, 0.6390163185945504, -1.1380061562065575], \"chunks\": 2000, \"d\": 10, \"max_iter\": 3, \"n\": 8000000, \"pointwise_loss\": 392474274.2668188, \"rel_error\": 0.10178542461115976, \"seed\": 0, \"sigma\": 7.0, \"wall_time\": 16.024066925048828}, {\"approx_grad\": true, \"beta\": [-0.9083172935353033, 0.7288582671124717, 0.5878503446288098, -0.2563745573662875, -0.5116395655415267, 0.8216666253332333, -0.2935442589357671, 0.006857256737447226, 0.36823156748019614, -1.066219790184447], \"chunks\": 2000, \"d\": 10, \"max_iter\": 3, \"n\": 8000000, \"pointwise_loss\": 392580710.7184185, \"rel_error\": 0.11609047177517823, \"seed\": 1, \"sigma\": 7.0, \"wall_time\": 17.255850076675415}, {\"approx_grad\": true, \"beta\": [-0.8389011499255122, 0.5948283477142385, 0.7238175294286159, -0.2997997038890462, -0.6591657691232473, 0.8288540320772136, -0.3125267990645646, 0.1449823186819636, 0.5704160565098788, -1.2720900725201265], \"chunks\": 2000, \"d\": 10, \"max_iter\": 3, \"n\": 8000000, \"pointwise_loss\": 392708106.3297756, \"rel_error\": 0.1317116829867297, \"seed\": 2, \"sigma\": 7.0, \"wall_time\": 15.214453935623169}, {\"approx_grad\": true, \"beta\": [-0.9414078608218889, 0.7179276231792979, 0.5555960772034156, -0.49420754272047746, -0.40115295600901557, 0.8294078054465174, -0.2042147327239715, -0.15308573285461974, 0.38000894740318536, -1.05377066222109], \"chunks\": 2000, \"d\": 10, \"max_iter\": 3, \"n\": 8000000, \"pointwise_loss\": 393229638.7225437, \"rel_error\": 0.17764704001228399, \"seed\": 3, \"sigma\": 7.0, \"wall_time\": 18.07921314239502}, {\"approx_grad\": true, \"beta\": [-0.926476804277389, 0.6474259797624344, 0.46509357607425766, -0.31031412639931977, -0.2912526589717621, 0.7152394416847999, -0.22762164094017698, 0.17662110468719983, 0.4446096993058229, -1.1348464288890303], \"chunks\": 2000, \"d\": 10, \"max_iter\": 3, \"n\": 8000000, \"pointwise_loss\": 393565654.71435046, \"rel_error\": 0.20169795281469685, \"seed\": 4, \"sigma\": 7.0, \"wall_time\": 16.644001722335815}, {\"approx_grad\": false, \"beta\": [-0.8840752372471408, 0.6155811769804287, 0.6431364460452087, -0.32301339020313813, -0.5412142113891235, 0.9374405298154759, -0.2980612954990996, 0.011143191960726204, 0.5115415216339614, -1.132843412924105], \"chunks\": 2000, \"d\": 10, \"max_iter\": 3, \"n\": 8000000, \"pointwise_loss\": 392098732.22278774, \"rel_error\": 0.0031039884258993638, \"seed\": 0, \"sigma\": 7.0, \"wall_time\": 145.35059690475464}, {\"approx_grad\": false, \"beta\": [-0.8840752372471408, 0.6155811769804287, 0.6431364460452087, -0.32301339020313813, -0.5412142113891235, 0.9374405298154759, -0.2980612954990996, 0.011143191960726204, 0.5115415216339614, -1.132843412924105], \"chunks\": 2000, \"d\": 10, \"max_iter\": 3, \"n\": 8000000, \"pointwise_loss\": 392098732.22278774, \"rel_error\": 0.0031039884258993633, \"seed\": 1, \"sigma\": 7.0, \"wall_time\": 147.1575231552124}, {\"approx_grad\": true, \"beta\": [-0.8271726726227692, 0.49951926049589246, 0.704512362421464, -0.2850862602450924, -0.6863328958031757, 0.8601722207246763, -0.508975427468857, -0.06062527242346647, 0.5160043265879157, -1.2795063115296572], \"chunks\": 2000, \"d\": 10, \"max_iter\": 4, \"n\": 8000000, \"pointwise_loss\": 393059855.81411487, \"rel_error\": 0.16445766469948245, \"seed\": 0, \"sigma\": 7.0, \"wall_time\": 22.655385971069336}, {\"approx_grad\": true, \"beta\": [-0.9616765072338722, 0.5343757451320267, 0.5564692898481249, -0.20815412080423262, -0.6599511653081934, 1.00051922660928, -0.47053635410121947, 0.01242918279858982, 0.5791592404773261, -1.2908888698588328], \"chunks\": 2000, \"d\": 10, \"max_iter\": 4, \"n\": 8000000, \"pointwise_loss\": 392983624.5057327, \"rel_error\": 0.156090826714074, \"seed\": 1, \"sigma\": 7.0, \"wall_time\": 19.843119859695435}, {\"approx_grad\": true, \"beta\": [-0.8311171789400179, 0.521410421564028, 0.9462671660740188, -0.3456457641986597, -0.29334977323977884, 0.9560068287397671, -0.22130619909174298, -0.0049895114959127105, 0.8629391239905513, -1.1745132638244187], \"chunks\": 2000, \"d\": 10, \"max_iter\": 4, \"n\": 8000000, \"pointwise_loss\": 394477000.17289954, \"rel_error\": 0.25672332004490195, \"seed\": 2, \"sigma\": 7.0, \"wall_time\": 20.78614115715027}, {\"approx_grad\": true, \"beta\": [-0.801139199063629, 0.16532940342659558, 0.6954539332675801, -0.28936401002679035, -0.7217519603155139, 1.091651439217816, -0.43199373306063593, -0.11066348276909367, 0.2855398064677888, -1.355992899941729], \"chunks\": 2000, \"d\": 10, \"max_iter\": 4, \"n\": 8000000, \"pointwise_loss\": 395327809.9076261, \"rel_error\": 0.3000645202606274, \"seed\": 3, \"sigma\": 7.0, \"wall_time\": 23.399425983428955}, {\"approx_grad\": true, \"beta\": [-0.7053576581148432, 0.6241930091984895, 0.5853791964035107, -0.2521215481878079, -0.8336804942073975, 0.7948488717641425, -0.278253920773729, 0.018089168528746924, 0.22836203225370016, -0.9739507392001161], \"chunks\": 2000, \"d\": 10, \"max_iter\": 4, \"n\": 8000000, \"pointwise_loss\": 394115898.34814155, \"rel_error\": 0.23886453574779987, \"seed\": 4, \"sigma\": 7.0, \"wall_time\": 22.683278799057007}, {\"approx_grad\": false, \"beta\": [-0.8840149290810264, 0.6155952392872667, 0.6429144576863511, -0.32298713855219047, -0.5413417425959787, 0.9373981467172503, -0.2979957802729954, 0.01126732427321426, 0.5117417330680154, -1.1327922152299434], \"chunks\": 2000, \"d\": 10, \"max_iter\": 4, \"n\": 8000000, \"pointwise_loss\": 392098731.9225391, \"rel_error\": 0.003110684443249633, \"seed\": 0, \"sigma\": 7.0, \"wall_time\": 189.257798910141}, {\"approx_grad\": false, \"beta\": [-0.8840149290810264, 0.6155952392872667, 0.6429144576863511, -0.32298713855219047, -0.5413417425959787, 0.9373981467172503, -0.2979957802729954, 0.01126732427321426, 0.5117417330680154, -1.1327922152299434], \"chunks\": 2000, \"d\": 10, \"max_iter\": 4, \"n\": 8000000, \"pointwise_loss\": 392098731.9225391, \"rel_error\": 0.003110684443249633, \"seed\": 1, \"sigma\": 7.0, \"wall_time\": 186.81083416938782}, {\"approx_grad\": true, \"beta\": [-0.9407374451959918, 0.5995469772485784, 0.3829421849180021, -0.5000865670359426, -0.5737165925062941, 1.1179590375704331, -0.05387964227593586, -0.062217309307136766, -0.002549105064196923, -0.8238232877423541], \"chunks\": 2000, \"d\": 10, \"max_iter\": 5, \"n\": 8000000, \"pointwise_loss\": 396586220.0473553, \"rel_error\": 0.3538455628010787, \"seed\": 0, \"sigma\": 7.0, \"wall_time\": 25.987020015716553}, {\"approx_grad\": true, \"beta\": [-0.4964414415005384, 0.6718447193669304, 0.6846026851661291, -0.010461025842725123, -0.34100874172096585, 0.9586468990831749, -0.21698443097406914, -0.13017291237255, 0.9125536530331779, -0.6087769509034419], \"chunks\": 2000, \"d\": 10, \"max_iter\": 5, \"n\": 8000000, \"pointwise_loss\": 398139813.413357, \"rel_error\": 0.4096640005375313, \"seed\": 1, \"sigma\": 7.0, \"wall_time\": 22.02472710609436}, {\"approx_grad\": true, \"beta\": [-0.8829626205083301, 1.1212609577680581, 0.5684551025155554, -0.6505691911598741, -0.9372315683339582, 0.842119326116806, -0.42196789947305235, -0.08570587191028647, 0.08368869163034408, -1.2792551410778739], \"chunks\": 2000, \"d\": 10, \"max_iter\": 5, \"n\": 8000000, \"pointwise_loss\": 398210025.845274, \"rel_error\": 0.41423200443526725, \"seed\": 2, \"sigma\": 7.0, \"wall_time\": 22.424404859542847}, {\"approx_grad\": true, \"beta\": [-0.8276390460755914, 0.6254867138609531, 0.3250671199876772, -0.17258921033454458, -0.19899367174226415, 0.802372686454787, -0.29783989269792505, 0.15952477773340476, 0.7608196944681033, -1.0561404959330822], \"chunks\": 2000, \"d\": 10, \"max_iter\": 5, \"n\": 8000000, \"pointwise_loss\": 394918666.8340963, \"rel_error\": 0.2788153092040237, \"seed\": 3, \"sigma\": 7.0, \"wall_time\": 23.654550075531006}, {\"approx_grad\": true, \"beta\": [-1.0663379690958243, 0.5137748168705915, 0.7515449969486085, -0.8673948569046709, -0.5042643587430429, 1.400751405530614, -0.45931404840107765, 0.13862134013387145, 0.3718303171564792, -1.255472840503655], \"chunks\": 2000, \"d\": 10, \"max_iter\": 5, \"n\": 8000000, \"pointwise_loss\": 397256096.00919074, \"rel_error\": 0.37940869669671184, \"seed\": 4, \"sigma\": 7.0, \"wall_time\": 30.06166672706604}, {\"approx_grad\": false, \"beta\": [-0.8840363068073565, 0.6155901299411422, 0.6429925519141578, -0.32299641680865165, -0.5412969800789815, 0.937413323810281, -0.298018917144369, 0.011223695205511832, 0.5116712128875367, -1.132810384813737], \"chunks\": 2000, \"d\": 10, \"max_iter\": 5, \"n\": 8000000, \"pointwise_loss\": 392098731.7830447, \"rel_error\": 0.003107192716169649, \"seed\": 0, \"sigma\": 7.0, \"wall_time\": 201.68372106552124}, {\"approx_grad\": false, \"beta\": [-0.8840363068073565, 0.6155901299411422, 0.6429925519141578, -0.32299641680865165, -0.5412969800789815, 0.937413323810281, -0.298018917144369, 0.011223695205511832, 0.5116712128875367, -1.132810384813737], \"chunks\": 2000, \"d\": 10, \"max_iter\": 5, \"n\": 8000000, \"pointwise_loss\": 392098731.7830447, \"rel_error\": 0.003107192716169649, \"seed\": 1, \"sigma\": 7.0, \"wall_time\": 204.56695413589478}, {\"approx_grad\": true, \"beta\": [-0.7871260042083935, 0.3961733101780085, 0.6958527709901753, -0.3142060076706436, -0.6376767465113462, 0.960881014904799, -0.31250840955083714, 0.08209674189561741, 0.515565851142813, -1.2308461475839017], \"chunks\": 2000, \"d\": 10, \"max_iter\": 6, \"n\": 8000000, \"pointwise_loss\": 392779349.549917, \"rel_error\": 0.13854581491536738, \"seed\": 0, \"sigma\": 7.0, \"wall_time\": 36.94024991989136}, {\"approx_grad\": true, \"beta\": [-0.6031158370785182, 0.6709487169693519, 0.7702163302750972, -0.4654554489878465, -0.7768143279733025, 0.9003608368757029, -0.37337880686709085, -0.07810376917934554, 0.5054273482702586, -1.1163010862552663], \"chunks\": 2000, \"d\": 10, \"max_iter\": 6, \"n\": 8000000, \"pointwise_loss\": 393612977.4438118, \"rel_error\": 0.20778431721208035, \"seed\": 1, \"sigma\": 7.0, \"wall_time\": 28.863558053970337}, {\"approx_grad\": true, \"beta\": [-0.8664624880186267, 0.4366107494098207, 0.7988147992751405, -0.2926785272602952, -0.5428020755237238, 0.8156204506426754, -0.19816820095111792, -0.034020062784066066, 0.37078322939922836, -1.137769387721782], \"chunks\": 2000, \"d\": 10, \"max_iter\": 6, \"n\": 8000000, \"pointwise_loss\": 392932772.7201239, \"rel_error\": 0.15320684314729843, \"seed\": 2, \"sigma\": 7.0, \"wall_time\": 28.86776614189148}, {\"approx_grad\": true, \"beta\": [-0.9508191916738734, 0.5803069168850657, 0.6081637705377964, -0.4316316463999406, -0.5012948159484267, 0.968552261019256, -0.24577223705663653, 0.06299812960488932, 0.2590370207057818, -1.252127732722464], \"chunks\": 2000, \"d\": 10, \"max_iter\": 6, \"n\": 8000000, \"pointwise_loss\": 392936926.844349, \"rel_error\": 0.15315234339841183, \"seed\": 3, \"sigma\": 7.0, \"wall_time\": 27.614052772521973}, {\"approx_grad\": true, \"beta\": [-0.8091028799637264, 0.45004047894098886, 0.7061596317906195, -0.3256998918927045, -0.49755334068387674, 0.9253875802460394, -0.11361957804608215, 0.24122491463255788, 0.5365126582799239, -1.1292519285384686], \"chunks\": 2000, \"d\": 10, \"max_iter\": 6, \"n\": 8000000, \"pointwise_loss\": 393111480.63514334, \"rel_error\": 0.1684587208741917, \"seed\": 4, \"sigma\": 7.0, \"wall_time\": 26.95098090171814}, {\"approx_grad\": false, \"beta\": [-0.8840368556424033, 0.6155900530326254, 0.6429948167902779, -0.3229966670664642, -0.541295637133796, 0.937413646700167, -0.2980195490623621, 0.011222412409941856, 0.5116692056516307, -1.132810842239684], \"chunks\": 2000, \"d\": 10, \"max_iter\": 6, \"n\": 8000000, \"pointwise_loss\": 392098731.78297484, \"rel_error\": 0.00310712691370173, \"seed\": 0, \"sigma\": 7.0, \"wall_time\": 229.23673105239868}, {\"approx_grad\": false, \"beta\": [-0.8840368556424033, 0.6155900530326254, 0.6429948167902779, -0.3229966670664642, -0.541295637133796, 0.937413646700167, -0.2980195490623621, 0.011222412409941856, 0.5116692056516307, -1.132810842239684], \"chunks\": 2000, \"d\": 10, \"max_iter\": 6, \"n\": 8000000, \"pointwise_loss\": 392098731.78297484, \"rel_error\": 0.00310712691370173, \"seed\": 1, \"sigma\": 7.0, \"wall_time\": 227.5106761455536}, {\"approx_grad\": true, \"beta\": [-0.9289146412636535, 0.589207394081429, 0.7012400145186803, -0.4208002014380348, -0.43386871861124676, 0.933089665396426, -0.4005966465091005, 0.06483459323652249, 0.5673006082227218, -1.225626019604088], \"chunks\": 2000, \"d\": 10, \"max_iter\": 10, \"n\": 8000000, \"pointwise_loss\": 392517421.8296241, \"rel_error\": 0.10793165451244967, \"seed\": 0, \"sigma\": 7.0, \"wall_time\": 52.241666078567505}, {\"approx_grad\": true, \"beta\": [-0.9743804887017291, 0.5410402380568015, 0.7704433890334813, -0.35185475181030945, -0.46356642753222593, 0.9481604622933278, -0.40504543584553954, 0.06890603470985716, 0.4346690052333042, -1.0354939209957497], \"chunks\": 2000, \"d\": 10, \"max_iter\": 10, \"n\": 8000000, \"pointwise_loss\": 392636046.76928383, \"rel_error\": 0.12258435738129043, \"seed\": 1, \"sigma\": 7.0, \"wall_time\": 49.099812269210815}, {\"approx_grad\": true, \"beta\": [-0.9691976834229626, 0.6475594312600076, 0.7508500065046348, -0.32858753685093967, -0.5462316324992845, 0.8922110086481335, -0.29444047752987906, -0.007789285206273814, 0.5428055414686452, -1.0599010679159921], \"chunks\": 2000, \"d\": 10, \"max_iter\": 10, \"n\": 8000000, \"pointwise_loss\": 392327909.8895484, \"rel_error\": 0.08030740531555673, \"seed\": 2, \"sigma\": 7.0, \"wall_time\": 53.91449022293091}, {\"approx_grad\": true, \"beta\": [-0.910264983109954, 0.5498026119249138, 0.7398351291370956, -0.29663314818807934, -0.6484180771824207, 0.9684016736573872, -0.10330011565812211, 0.05913743859741222, 0.7403852268345604, -1.1810385316788785], \"chunks\": 2000, \"d\": 10, \"max_iter\": 10, \"n\": 8000000, \"pointwise_loss\": 393077610.3466072, \"rel_error\": 0.16494774361606168, \"seed\": 3, \"sigma\": 7.0, \"wall_time\": 56.24317216873169}, {\"approx_grad\": true, \"beta\": [-0.8675186275412994, 0.4551384495672032, 0.604807568567115, -0.2831505725214636, -0.5860885866345346, 0.8674112928558778, -0.416804818232409, -0.1134532111679631, 0.38210984739072146, -1.075781297833849], \"chunks\": 2000, \"d\": 10, \"max_iter\": 10, \"n\": 8000000, \"pointwise_loss\": 392784311.2327362, \"rel_error\": 0.1385175667205294, \"seed\": 4, \"sigma\": 7.0, \"wall_time\": 62.55457830429077}, {\"approx_grad\": false, \"beta\": [-0.8840367015486226, 0.6155900769031977, 0.6429941913527001, -0.3229965972154067, -0.5412960062534135, 0.9374135532285174, -0.2980193731251264, 0.011222765961551154, 0.511669761456385, -1.1328107133068281], \"chunks\": 2000, \"d\": 10, \"max_iter\": 10, \"n\": 8000000, \"pointwise_loss\": 392098731.7829722, \"rel_error\": 0.003107145577421732, \"seed\": 0, \"sigma\": 7.0, \"wall_time\": 301.74830627441406}, {\"approx_grad\": false, \"beta\": [-0.8840367015486226, 0.6155900769031977, 0.6429941913527001, -0.3229965972154067, -0.5412960062534135, 0.9374135532285174, -0.2980193731251264, 0.011222765961551154, 0.511669761456385, -1.1328107133068281], \"chunks\": 2000, \"d\": 10, \"max_iter\": 10, \"n\": 8000000, \"pointwise_loss\": 392098731.7829722, \"rel_error\": 0.003107145577421732, \"seed\": 1, \"sigma\": 7.0, \"wall_time\": 283.68756103515625}, {\"approx_grad\": true, \"beta\": [-0.879226020902641, 0.7785837679404635, 0.5854598664790497, -0.2652160310122725, -0.5033963847997842, 0.9101222961481134, -0.3230869875520656, 0.0726667810267524, 0.6090404727113902, -1.015491282425678], \"chunks\": 2000, \"d\": 10, \"max_iter\": 15, \"n\": 8000000, \"pointwise_loss\": 392603554.26582336, \"rel_error\": 0.1182117141794517, \"seed\": 0, \"sigma\": 7.0, \"wall_time\": 88.87447690963745}, {\"approx_grad\": true, \"beta\": [-0.9264615710300912, 0.5793376114442538, 0.5459787720950554, -0.3485187459810788, -0.737088753955497, 1.0098549802053038, -0.2697041886515856, -0.15363978801505496, 0.38003248055376626, -1.2006135031985268], \"chunks\": 2000, \"d\": 10, \"max_iter\": 15, \"n\": 8000000, \"pointwise_loss\": 392952025.853789, \"rel_error\": 0.15435494003660002, \"seed\": 1, \"sigma\": 7.0, \"wall_time\": 82.69940185546875}, {\"approx_grad\": true, \"beta\": [-0.9206027132484186, 0.739878760663106, 0.7270095070520408, -0.38870409165038816, -0.611151472242097, 0.715224504550175, -0.441189414563162, 0.11140238970762492, 0.5215692165038381, -1.0797497203817892], \"chunks\": 2000, \"d\": 10, \"max_iter\": 15, \"n\": 8000000, \"pointwise_loss\": 393025962.4803908, \"rel_error\": 0.1625869001127405, \"seed\": 2, \"sigma\": 7.0, \"wall_time\": 92.84537315368652}, {\"approx_grad\": true, \"beta\": [-0.787874281047773, 0.6799359185856565, 0.5416287983104555, -0.37731162844352173, -0.5493588312272013, 0.8054093051255071, -0.41572109145020375, -0.10753865241218799, 0.5014026377711878, -1.027600255568214], \"chunks\": 2000, \"d\": 10, \"max_iter\": 15, \"n\": 8000000, \"pointwise_loss\": 392764542.0845498, \"rel_error\": 0.13723331555685436, \"seed\": 3, \"sigma\": 7.0, \"wall_time\": 90.04433107376099}, {\"approx_grad\": true, \"beta\": [-0.8744565169107531, 0.604436206417696, 0.46990150193534835, -0.2077964962204974, -0.5355731471077425, 0.7940482312493866, -0.2473564612549025, -0.13633621899376802, 0.5697639944472284, -1.1292764001824989], \"chunks\": 2000, \"d\": 10, \"max_iter\": 15, \"n\": 8000000, \"pointwise_loss\": 392833172.4833296, \"rel_error\": 0.14196395307157753, \"seed\": 4, \"sigma\": 7.0, \"wall_time\": 80.92051577568054}, {\"approx_grad\": false, \"beta\": [-0.8840367015486226, 0.6155900769031977, 0.6429941913527001, -0.3229965972154067, -0.5412960062534135, 0.9374135532285174, -0.2980193731251264, 0.011222765961551154, 0.511669761456385, -1.1328107133068281], \"chunks\": 2000, \"d\": 10, \"max_iter\": 15, \"n\": 8000000, \"pointwise_loss\": 392098731.7829722, \"rel_error\": 0.003107145577421732, \"seed\": 0, \"sigma\": 7.0, \"wall_time\": 265.0206460952759}, {\"approx_grad\": false, \"beta\": [-0.8840367015486226, 0.6155900769031977, 0.6429941913527001, -0.3229965972154067, -0.5412960062534135, 0.9374135532285174, -0.2980193731251264, 0.011222765961551154, 0.511669761456385, -1.1328107133068281], \"chunks\": 2000, \"d\": 10, \"max_iter\": 15, \"n\": 8000000, \"pointwise_loss\": 392098731.7829722, \"rel_error\": 0.003107145577421732, \"seed\": 1, \"sigma\": 7.0, \"wall_time\": 259.3626790046692}, {\"approx_grad\": true, \"beta\": [-0.9042011637589843, 0.7029405553155955, 0.7110057308292758, -0.3704724339708527, -0.43615660607115325, 0.9149258394399113, -0.2950345455604007, 0.12055345649464859, 0.5209865340313132, -0.9256352343235037], \"chunks\": 2000, \"d\": 10, \"max_iter\": 20, \"n\": 8000000, \"pointwise_loss\": 392750046.25591063, \"rel_error\": 0.13527978657110445, \"seed\": 0, \"sigma\": 7.0, \"wall_time\": 108.67117500305176}, {\"approx_grad\": true, \"beta\": [-0.9719847430972252, 0.561696376565175, 0.7420408719990494, -0.34829215067329033, -0.46395348346167564, 1.0175211166551112, -0.5115341156758816, 0.1059482707651871, 0.46804927489393644, -1.2154401929156522], \"chunks\": 2000, \"d\": 10, \"max_iter\": 20, \"n\": 8000000, \"pointwise_loss\": 392873611.206111, \"rel_error\": 0.14661583431505454, \"seed\": 1, \"sigma\": 7.0, \"wall_time\": 117.8687379360199}, {\"approx_grad\": true, \"beta\": [-0.7238922364851894, 0.48440091636092486, 0.5483989511032223, -0.28848806460103726, -0.5291904133328238, 0.966457907630273, -0.4282109825815653, 0.09406156210969836, 0.5318905979292694, -1.1103801072529718], \"chunks\": 2000, \"d\": 10, \"max_iter\": 20, \"n\": 8000000, \"pointwise_loss\": 392728243.39322346, \"rel_error\": 0.1327308431752905, \"seed\": 2, \"sigma\": 7.0, \"wall_time\": 114.08178377151489}, {\"approx_grad\": true, \"beta\": [-0.9438060751341817, 0.5639914647299185, 0.5472097964866391, -0.42914470178028863, -0.4964936943584284, 0.8407517247734084, -0.276133337484286, -0.0008974746619392729, 0.6534416631567372, -1.1682447312995101], \"chunks\": 2000, \"d\": 10, \"max_iter\": 20, \"n\": 8000000, \"pointwise_loss\": 392578910.63095534, \"rel_error\": 0.11538053946876133, \"seed\": 3, \"sigma\": 7.0, \"wall_time\": 117.60523080825806}, {\"approx_grad\": true, \"beta\": [-0.9221378142827324, 0.7839502847556458, 0.48953742742167167, -0.3521669540471849, -0.6727637426480056, 0.9333889365145457, -0.2866852634483545, 0.05098493792076283, 0.5652542241558753, -1.1840084441981549], \"chunks\": 2000, \"d\": 10, \"max_iter\": 20, \"n\": 8000000, \"pointwise_loss\": 392728530.95368534, \"rel_error\": 0.1327486638410105, \"seed\": 4, \"sigma\": 7.0, \"wall_time\": 118.73969268798828}, {\"approx_grad\": false, \"beta\": [-0.8840367015486226, 0.6155900769031977, 0.6429941913527001, -0.3229965972154067, -0.5412960062534135, 0.9374135532285174, -0.2980193731251264, 0.011222765961551154, 0.511669761456385, -1.1328107133068281], \"chunks\": 2000, \"d\": 10, \"max_iter\": 20, \"n\": 8000000, \"pointwise_loss\": 392098731.7829722, \"rel_error\": 0.003107145577421732, \"seed\": 0, \"sigma\": 7.0, \"wall_time\": 267.0480582714081}, {\"approx_grad\": false, \"beta\": [-0.8840367015486226, 0.6155900769031977, 0.6429941913527001, -0.3229965972154067, -0.5412960062534135, 0.9374135532285174, -0.2980193731251264, 0.011222765961551154, 0.511669761456385, -1.1328107133068281], \"chunks\": 2000, \"d\": 10, \"max_iter\": 20, \"n\": 8000000, \"pointwise_loss\": 392098731.7829722, \"rel_error\": 0.003107145577421732, \"seed\": 1, \"sigma\": 7.0, \"wall_time\": 262.2125267982483}, {\"approx_grad\": true, \"beta\": [-0.8829139465236511, 0.625917698868711, 0.5610035845124403, -0.29290341592114466, -0.5451773218871103, 0.9795265198140615, -0.2515854796694451, 3.103563452222811e-05, 0.5432008109787038, -1.100514723843423], \"chunks\": 2000, \"d\": 10, \"max_iter\": 30, \"n\": 8000000, \"pointwise_loss\": 392209513.1652502, \"rel_error\": 0.05404568835450826, \"seed\": 0, \"sigma\": 7.0, \"wall_time\": 195.57359433174133}, {\"approx_grad\": true, \"beta\": [-0.8635894856734007, 0.6248311725057915, 0.630928540527987, -0.3554994869075913, -0.5729866601651461, 0.950780176643394, -0.33218680175151266, -0.031223046549164638, 0.5121029562335224, -1.070479358507842], \"chunks\": 2000, \"d\": 10, \"max_iter\": 30, \"n\": 8000000, \"pointwise_loss\": 392176658.7978555, \"rel_error\": 0.04775019218448381, \"seed\": 1, \"sigma\": 7.0, \"wall_time\": 185.2518310546875}, {\"approx_grad\": true, \"beta\": [-0.854022071521578, 0.6031780361923385, 0.6422737722072802, -0.36969844012775666, -0.6124105215693041, 0.8853708036980525, -0.2445671723441642, 0.012866940525705072, 0.6060871681055311, -1.1131080448188513], \"chunks\": 2000, \"d\": 10, \"max_iter\": 30, \"n\": 8000000, \"pointwise_loss\": 392284167.6319425, \"rel_error\": 0.07326080170913572, \"seed\": 2, \"sigma\": 7.0, \"wall_time\": 195.97774267196655}, {\"approx_grad\": true, \"beta\": [-0.8827942112640698, 0.6166268364806528, 0.5828977133369241, -0.2929787420823192, -0.6212947554007182, 0.9072455211672876, -0.2748304348371266, 0.060147878868150505, 0.5433172421187823, -1.1121212867854486], \"chunks\": 2000, \"d\": 10, \"max_iter\": 30, \"n\": 8000000, \"pointwise_loss\": 392228262.4553793, \"rel_error\": 0.06074344122218579, \"seed\": 3, \"sigma\": 7.0, \"wall_time\": 206.70327305793762}, {\"approx_grad\": true, \"beta\": [-0.8629285267307552, 0.5671430205245978, 0.5939363055152163, -0.3136346831101807, -0.48504145478632693, 0.9055869404266276, -0.3386675928486884, 0.024980442735857315, 0.49453810816927624, -1.0657688750768044], \"chunks\": 2000, \"d\": 10, \"max_iter\": 30, \"n\": 8000000, \"pointwise_loss\": 392227507.05134, \"rel_error\": 0.059750956746531476, \"seed\": 4, \"sigma\": 7.0, \"wall_time\": 202.11913299560547}, {\"approx_grad\": false, \"beta\": [-0.8840367015486226, 0.6155900769031977, 0.6429941913527001, -0.3229965972154067, -0.5412960062534135, 0.9374135532285174, -0.2980193731251264, 0.011222765961551154, 0.511669761456385, -1.1328107133068281], \"chunks\": 2000, \"d\": 10, \"max_iter\": 30, \"n\": 8000000, \"pointwise_loss\": 392098731.7829722, \"rel_error\": 0.003107145577421732, \"seed\": 0, \"sigma\": 7.0, \"wall_time\": 256.4530289173126}, {\"approx_grad\": false, \"beta\": [-0.8840367015486226, 0.6155900769031977, 0.6429941913527001, -0.3229965972154067, -0.5412960062534135, 0.9374135532285174, -0.2980193731251264, 0.011222765961551154, 0.511669761456385, -1.1328107133068281], \"chunks\": 2000, \"d\": 10, \"max_iter\": 30, \"n\": 8000000, \"pointwise_loss\": 392098731.7829722, \"rel_error\": 0.003107145577421732, \"seed\": 1, \"sigma\": 7.0, \"wall_time\": 271.3524467945099}, {\"approx_grad\": true, \"beta\": [-0.9144115552180049, 0.603015498404459, 0.5713064287707974, -0.34545575275615004, -0.6424885972652146, 0.8931970146541598, -0.2331797791375983, -0.02907522461562066, 0.46925773932688525, -1.1219401556826756], \"chunks\": 2000, \"d\": 10, \"max_iter\": 40, \"n\": 8000000, \"pointwise_loss\": 392312074.65794194, \"rel_error\": 0.0779374005493023, \"seed\": 0, \"sigma\": 7.0, \"wall_time\": 264.6048061847687}, {\"approx_grad\": true, \"beta\": [-0.8973887214484578, 0.625628029190906, 0.6978309024390227, -0.32447667279242576, -0.5497618115288487, 0.9206360070908075, -0.2452788563571229, -0.02033314724967066, 0.4922056424940327, -1.0984816585576087], \"chunks\": 2000, \"d\": 10, \"max_iter\": 40, \"n\": 8000000, \"pointwise_loss\": 392170503.4222042, \"rel_error\": 0.04541517197333465, \"seed\": 1, \"sigma\": 7.0, \"wall_time\": 269.0821568965912}, {\"approx_grad\": true, \"beta\": [-0.8823393235027647, 0.5492224550551256, 0.7019713370961441, -0.3435038812509715, -0.5853327207643598, 0.9835950273865081, -0.27844770423239806, -0.055921215247640724, 0.4887343295786643, -1.1194613758252137], \"chunks\": 2000, \"d\": 10, \"max_iter\": 40, \"n\": 8000000, \"pointwise_loss\": 392242635.91384065, \"rel_error\": 0.06373818373096356, \"seed\": 2, \"sigma\": 7.0, \"wall_time\": 256.9263689517975}, {\"approx_grad\": true, \"beta\": [-0.8279929314291558, 0.5785406092292175, 0.6888273335787916, -0.354480652670551, -0.5278944818518453, 0.9341898266525525, -0.28833210613646065, 0.12581324637761054, 0.47795866600785925, -1.159984780697032], \"chunks\": 2000, \"d\": 10, \"max_iter\": 40, \"n\": 8000000, \"pointwise_loss\": 392281900.7215971, \"rel_error\": 0.0729637904656608, \"seed\": 3, \"sigma\": 7.0, \"wall_time\": 243.45107984542847}, {\"approx_grad\": true, \"beta\": [-0.9078119382540301, 0.5449053559924391, 0.6857141065778634, -0.33039721408797773, -0.546340031592623, 0.9589346749302237, -0.2378234765841817, 0.01719609554537357, 0.4794984490149854, -1.1048310139934197], \"chunks\": 2000, \"d\": 10, \"max_iter\": 40, \"n\": 8000000, \"pointwise_loss\": 392205987.4079259, \"rel_error\": 0.05482705526360315, \"seed\": 4, \"sigma\": 7.0, \"wall_time\": 247.5673270225525}, {\"approx_grad\": false, \"beta\": [-0.8840367015486226, 0.6155900769031977, 0.6429941913527001, -0.3229965972154067, -0.5412960062534135, 0.9374135532285174, -0.2980193731251264, 0.011222765961551154, 0.511669761456385, -1.1328107133068281], \"chunks\": 2000, \"d\": 10, \"max_iter\": 40, \"n\": 8000000, \"pointwise_loss\": 392098731.7829722, \"rel_error\": 0.003107145577421732, \"seed\": 0, \"sigma\": 7.0, \"wall_time\": 261.3412778377533}, {\"approx_grad\": false, \"beta\": [-0.8840367015486226, 0.6155900769031977, 0.6429941913527001, -0.3229965972154067, -0.5412960062534135, 0.9374135532285174, -0.2980193731251264, 0.011222765961551154, 0.511669761456385, -1.1328107133068281], \"chunks\": 2000, \"d\": 10, \"max_iter\": 40, \"n\": 8000000, \"pointwise_loss\": 392098731.7829722, \"rel_error\": 0.003107145577421732, \"seed\": 1, \"sigma\": 7.0, \"wall_time\": 243.74875688552856}, {\"approx_grad\": true, \"beta\": [-0.8854073087268456, 0.6172178631773021, 0.6450145855969107, -0.32189737380866645, -0.5447108173125104, 0.9348775941459327, -0.29811605878941355, 0.013981237885834294, 0.5143839539544327, -1.1370906424701517], \"chunks\": 2000, \"d\": 10, \"max_iter\": 80, \"n\": 8000000, \"pointwise_loss\": 392099221.5401707, \"rel_error\": 0.005300432688336602, \"seed\": 0, \"sigma\": 7.0, \"wall_time\": 925.5044710636139}, {\"approx_grad\": true, \"beta\": [-0.8818837818078422, 0.613442176739526, 0.6419890655824715, -0.32382572208301896, -0.5415035086126159, 0.9344285595975396, -0.29847295104227245, 0.00894236421481264, 0.5093510115674734, -1.1237886908468033], \"chunks\": 2000, \"d\": 10, \"max_iter\": 80, \"n\": 8000000, \"pointwise_loss\": 392099628.5426975, \"rel_error\": 0.006657967894667196, \"seed\": 1, \"sigma\": 7.0, \"wall_time\": 1053.5675048828125}, {\"approx_grad\": true, \"beta\": [-0.8841524006885965, 0.6192790187917091, 0.6435924079837052, -0.3224913282377281, -0.5405742823931305, 0.9331562752015514, -0.303804688437872, 0.012429471065645424, 0.5122031575325743, -1.1330440943007851], \"chunks\": 2000, \"d\": 10, \"max_iter\": 80, \"n\": 8000000, \"pointwise_loss\": 392099277.1428425, \"rel_error\": 0.005527006304765571, \"seed\": 2, \"sigma\": 7.0, \"wall_time\": 1059.2264840602875}, {\"approx_grad\": true, \"beta\": [-0.8825335911850323, 0.6098598438138241, 0.6441103657737104, -0.3210903115661359, -0.5390939638365305, 0.933140324813863, -0.2976977255708058, 0.012605030831765042, 0.5152813665905337, -1.1273924354863225], \"chunks\": 2000, \"d\": 10, \"max_iter\": 80, \"n\": 8000000, \"pointwise_loss\": 392099592.01777464, \"rel_error\": 0.005867858009065959, \"seed\": 3, \"sigma\": 7.0, \"wall_time\": 1028.4987251758575}, {\"approx_grad\": true, \"beta\": [-0.8871790409057009, 0.6196366303695788, 0.6431369705966894, -0.3268682376868445, -0.5390371654462093, 0.9393368181590982, -0.294816062643927, 0.010144475140230992, 0.5120191693819195, -1.1346528130151012], \"chunks\": 2000, \"d\": 10, \"max_iter\": 80, \"n\": 8000000, \"pointwise_loss\": 392099251.88777584, \"rel_error\": 0.004409698151563556, \"seed\": 4, \"sigma\": 7.0, \"wall_time\": 1020.5811431407928}, {\"approx_grad\": false, \"beta\": [-0.8840367015486226, 0.6155900769031977, 0.6429941913527001, -0.3229965972154067, -0.5412960062534135, 0.9374135532285174, -0.2980193731251264, 0.011222765961551154, 0.511669761456385, -1.1328107133068281], \"chunks\": 2000, \"d\": 10, \"max_iter\": 80, \"n\": 8000000, \"pointwise_loss\": 392098731.7829722, \"rel_error\": 0.003107145577421732, \"seed\": 0, \"sigma\": 7.0, \"wall_time\": 273.7842438220978}, {\"approx_grad\": false, \"beta\": [-0.8840367015486226, 0.6155900769031977, 0.6429941913527001, -0.3229965972154067, -0.5412960062534135, 0.9374135532285174, -0.2980193731251264, 0.011222765961551154, 0.511669761456385, -1.1328107133068281], \"chunks\": 2000, \"d\": 10, \"max_iter\": 80, \"n\": 8000000, \"pointwise_loss\": 392098731.7829722, \"rel_error\": 0.003107145577421732, \"seed\": 1, \"sigma\": 7.0, \"wall_time\": 267.38317227363586}]}};\n",
"var selector = \"#7d6838e3-87a2-41de-b65c-bbd065e82350\";\n",
"var type = \"vega-lite\";\n",
"\n",
"var output_area = this;\n",
"require(['nbextensions/jupyter-vega/index'], function(vega) {\n",
" vega.render(selector, spec, type, output_area);\n",
"}, function (err) {\n",
" if (err.requireType !== 'scripterror') {\n",
" throw(err);\n",
" }\n",
"});\n"
]
},
"metadata": {
"jupyter-vega": "#7d6838e3-87a2-41de-b65c-bbd065e82350"
},
"output_type": "display_data"
},
{
"data": {
"image/png": "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"
},
"metadata": {
"jupyter-vega": "#7d6838e3-87a2-41de-b65c-bbd065e82350"
},
"output_type": "display_data"
}
],
"source": [
"from altair import X, Y, Scale\n",
"Chart(df).mark_line().encode(\n",
" x='wall_time',\n",
" y=Y('rel_error', scale=Scale(type='log')),\n",
" color='approx_grad')"
]
},
{
"cell_type": "code",
"execution_count": 11,
"metadata": {
"ExecuteTime": {
"end_time": "2017-09-01T19:51:50.694517Z",
"start_time": "2017-09-01T19:51:50.652027Z"
}
},
"outputs": [
{
"data": {
"text/html": [
"<div class=\"vega-embed\" id=\"7100bd77-4025-4247-acdd-2c798f08cbc7\"></div>\n",
"\n",
"<style>\n",
".vega-embed svg, .vega-embed canvas {\n",
" border: 1px dotted gray;\n",
"}\n",
"\n",
".vega-embed .vega-actions a {\n",
" margin-right: 6px;\n",
"}\n",
"</style>\n"
]
},
"metadata": {
"jupyter-vega": "#7100bd77-4025-4247-acdd-2c798f08cbc7"
},
"output_type": "display_data"
},
{
"data": {
"application/javascript": [
"var spec = {\"config\": {\"cell\": {\"width\": 500, \"height\": 350}}, \"encoding\": {\"color\": {\"field\": \"approx_grad\", \"type\": \"nominal\"}, \"x\": {\"field\": \"wall_time\", \"scale\": {\"type\": \"log\"}, \"type\": \"quantitative\"}, \"y\": {\"field\": \"pointwise_loss\", \"scale\": {\"clamp\": true, \"domain\": [392098731.7829722, 405328298.33456]}, \"type\": \"quantitative\"}}, \"mark\": \"line\", \"data\": {\"values\": [{\"approx_grad\": true, \"beta\": [-0.3377254481236899, 0.2301321067680867, 0.333146114043381, -0.20971334143549528, -0.30575032752005804, 0.3740567200034455, -0.06305069711196563, 0.18395187863790194, 0.12185236088423221, -0.5561521713524271], \"chunks\": 2000, \"d\": 10, \"max_iter\": 1, \"n\": 8000000, \"pointwise_loss\": 404087362.5696646, \"rel_error\": 0.5786328175760411, \"seed\": 0, \"sigma\": 7.0, \"wall_time\": 9.872261762619019}, {\"approx_grad\": true, \"beta\": [-0.4826127929824557, 0.3292431113973369, 0.3473824920275945, -0.15459457027121878, -0.19415953367371733, 0.42306454754726186, -0.17252527019265015, 0.05592581965829817, 0.15623894385874282, -0.43401916546021496], \"chunks\": 2000, \"d\": 10, \"max_iter\": 1, \"n\": 8000000, \"pointwise_loss\": 403110631.53552437, \"rel_error\": 0.5542360022939973, \"seed\": 1, \"sigma\": 7.0, \"wall_time\": 7.958230018615723}, {\"approx_grad\": true, \"beta\": [-0.3867231542110521, 0.18469768851413176, 0.2816929331607336, -0.054056285360848305, -0.22192282415811285, 0.33552478632731875, -0.11270602798687447, 0.009945650699710791, 0.1491485983684059, -0.5102915167728846], \"chunks\": 2000, \"d\": 10, \"max_iter\": 1, \"n\": 8000000, \"pointwise_loss\": 405328298.33456, \"rel_error\": 0.6073403494210304, \"seed\": 2, \"sigma\": 7.0, \"wall_time\": 7.640789985656738}, {\"approx_grad\": true, \"beta\": [-0.38046753748948453, 0.23782463200335902, 0.18564595937362763, -0.1889852272595575, -0.20634251350348562, 0.38498398172180837, -0.2799477671469581, 0.05463026794689662, 0.1997645016076342, -0.4443484968294167], \"chunks\": 2000, \"d\": 10, \"max_iter\": 1, \"n\": 8000000, \"pointwise_loss\": 405016186.76993334, \"rel_error\": 0.6002291017741715, \"seed\": 3, \"sigma\": 7.0, \"wall_time\": 9.718455076217651}, {\"approx_grad\": true, \"beta\": [-0.5025513159335334, 0.3680979167244689, 0.3038675340941493, -0.09045097996902372, -0.10998030666990562, 0.24256309845221122, -0.1839670635653993, 0.07396880196168064, 0.1942692091907179, -0.6048743288747124], \"chunks\": 2000, \"d\": 10, \"max_iter\": 1, \"n\": 8000000, \"pointwise_loss\": 403628905.38647205, \"rel_error\": 0.5669817044056502, \"seed\": 4, \"sigma\": 7.0, \"wall_time\": 9.272586107254028}, {\"approx_grad\": false, \"beta\": [-1.4148599575143694, 0.9848581358758047, 1.0303035090932562, -0.5169708152326576, -0.8651982685784295, 1.5001343392240916, -0.47727207669491456, 0.0171080426595288, 0.8173125569729787, -1.8128364003611475], \"chunks\": 2000, \"d\": 10, \"max_iter\": 1, \"n\": 8000000, \"pointwise_loss\": 405024480.2475509, \"rel_error\": 0.5995344874321145, \"seed\": 0, \"sigma\": 7.0, \"wall_time\": 88.51962184906006}, {\"approx_grad\": false, \"beta\": [-1.4148599575143694, 0.9848581358758047, 1.0303035090932562, -0.5169708152326576, -0.8651982685784295, 1.5001343392240916, -0.47727207669491456, 0.0171080426595288, 0.8173125569729787, -1.8128364003611475], \"chunks\": 2000, \"d\": 10, \"max_iter\": 1, \"n\": 8000000, \"pointwise_loss\": 405024480.2475509, \"rel_error\": 0.5995344874321145, \"seed\": 1, \"sigma\": 7.0, \"wall_time\": 72.5514748096466}, {\"approx_grad\": true, \"beta\": [-0.594331803280871, 0.43971841399477063, 0.5008066717190522, -0.24237453836988399, -0.4465450331221631, 0.654942040865184, -0.02191295874530111, 0.17804981679861503, 0.3397522624591079, -0.8989588947159972], \"chunks\": 2000, \"d\": 10, \"max_iter\": 2, \"n\": 8000000, \"pointwise_loss\": 395448737.1142796, \"rel_error\": 0.30623972828770346, \"seed\": 0, \"sigma\": 7.0, \"wall_time\": 10.498352766036987}, {\"approx_grad\": true, \"beta\": [-0.8128070379894679, 0.38924159720017654, 0.4144969044449924, -0.19841532859246383, -0.4710042101673059, 0.734437493524168, -0.2930097249427355, 0.034101664921965966, 0.5329805523482001, -0.9469222291091087], \"chunks\": 2000, \"d\": 10, \"max_iter\": 2, \"n\": 8000000, \"pointwise_loss\": 393745566.9921506, \"rel_error\": 0.2138264606829495, \"seed\": 1, \"sigma\": 7.0, \"wall_time\": 10.346577882766724}, {\"approx_grad\": true, \"beta\": [-0.6790975376476712, 0.5450407181605879, 0.5322319335883364, -0.14396416217709798, -0.5102201147023271, 0.7565770006251672, -0.11488155620503593, -0.03861813027138468, 0.3069642283910206, -0.9001416737560651], \"chunks\": 2000, \"d\": 10, \"max_iter\": 2, \"n\": 8000000, \"pointwise_loss\": 394154971.6913496, \"rel_error\": 0.23966802233459264, \"seed\": 2, \"sigma\": 7.0, \"wall_time\": 10.392645120620728}, {\"approx_grad\": true, \"beta\": [-0.811140159228611, 0.6276138384907568, 0.473085872816493, -0.1838171446679991, -0.40054954722887764, 0.7558401907634041, -0.42296970180408683, 0.06692302612485297, 0.24404777524362825, -0.9590326212179863], \"chunks\": 2000, \"d\": 10, \"max_iter\": 2, \"n\": 8000000, \"pointwise_loss\": 393915237.67307943, \"rel_error\": 0.2250417472016826, \"seed\": 3, \"sigma\": 7.0, \"wall_time\": 10.574932098388672}, {\"approx_grad\": true, \"beta\": [-0.8095071525177628, 0.39580708437907686, 0.5526693122538857, -0.30895738289937397, -0.4200459055558351, 0.6794734103735167, -0.14089336983588296, -0.042578922137064476, 0.37934092991144963, -0.994111196953386], \"chunks\": 2000, \"d\": 10, \"max_iter\": 2, \"n\": 8000000, \"pointwise_loss\": 393761112.8971651, \"rel_error\": 0.21548270480883275, \"seed\": 4, \"sigma\": 7.0, \"wall_time\": 11.583020210266113}, {\"approx_grad\": false, \"beta\": [-0.8838851099696825, 0.6156234649452773, 0.642426897118965, -0.32293019293604086, -0.541623528391047, 0.9373094310885044, -0.2978533442281764, 0.011540620405336708, 0.5121800257134298, -1.1326823717915948], \"chunks\": 2000, \"d\": 10, \"max_iter\": 2, \"n\": 8000000, \"pointwise_loss\": 392098738.7839557, \"rel_error\": 0.0031580153358463157, \"seed\": 0, \"sigma\": 7.0, \"wall_time\": 111.78914093971252}, {\"approx_grad\": false, \"beta\": [-0.8838851099696825, 0.6156234649452773, 0.642426897118965, -0.32293019293604086, -0.541623528391047, 0.9373094310885044, -0.2978533442281764, 0.011540620405336708, 0.5121800257134298, -1.1326823717915948], \"chunks\": 2000, \"d\": 10, \"max_iter\": 2, \"n\": 8000000, \"pointwise_loss\": 392098738.7839557, \"rel_error\": 0.0031580153358463157, \"seed\": 1, \"sigma\": 7.0, \"wall_time\": 115.92996788024902}, {\"approx_grad\": true, \"beta\": [-0.9046560108081807, 0.6326613755118812, 0.5504089312086718, -0.29349505832691863, -0.6703244988401252, 0.997889006713227, -0.2968343677535923, -0.00451966296877129, 0.6390163185945504, -1.1380061562065575], \"chunks\": 2000, \"d\": 10, \"max_iter\": 3, \"n\": 8000000, \"pointwise_loss\": 392474274.2668188, \"rel_error\": 0.10178542461115976, \"seed\": 0, \"sigma\": 7.0, \"wall_time\": 16.024066925048828}, {\"approx_grad\": true, \"beta\": [-0.9083172935353033, 0.7288582671124717, 0.5878503446288098, -0.2563745573662875, -0.5116395655415267, 0.8216666253332333, -0.2935442589357671, 0.006857256737447226, 0.36823156748019614, -1.066219790184447], \"chunks\": 2000, \"d\": 10, \"max_iter\": 3, \"n\": 8000000, \"pointwise_loss\": 392580710.7184185, \"rel_error\": 0.11609047177517823, \"seed\": 1, \"sigma\": 7.0, \"wall_time\": 17.255850076675415}, {\"approx_grad\": true, \"beta\": [-0.8389011499255122, 0.5948283477142385, 0.7238175294286159, -0.2997997038890462, -0.6591657691232473, 0.8288540320772136, -0.3125267990645646, 0.1449823186819636, 0.5704160565098788, -1.2720900725201265], \"chunks\": 2000, \"d\": 10, \"max_iter\": 3, \"n\": 8000000, \"pointwise_loss\": 392708106.3297756, \"rel_error\": 0.1317116829867297, \"seed\": 2, \"sigma\": 7.0, \"wall_time\": 15.214453935623169}, {\"approx_grad\": true, \"beta\": [-0.9414078608218889, 0.7179276231792979, 0.5555960772034156, -0.49420754272047746, -0.40115295600901557, 0.8294078054465174, -0.2042147327239715, -0.15308573285461974, 0.38000894740318536, -1.05377066222109], \"chunks\": 2000, \"d\": 10, \"max_iter\": 3, \"n\": 8000000, \"pointwise_loss\": 393229638.7225437, \"rel_error\": 0.17764704001228399, \"seed\": 3, \"sigma\": 7.0, \"wall_time\": 18.07921314239502}, {\"approx_grad\": true, \"beta\": [-0.926476804277389, 0.6474259797624344, 0.46509357607425766, -0.31031412639931977, -0.2912526589717621, 0.7152394416847999, -0.22762164094017698, 0.17662110468719983, 0.4446096993058229, -1.1348464288890303], \"chunks\": 2000, \"d\": 10, \"max_iter\": 3, \"n\": 8000000, \"pointwise_loss\": 393565654.71435046, \"rel_error\": 0.20169795281469685, \"seed\": 4, \"sigma\": 7.0, \"wall_time\": 16.644001722335815}, {\"approx_grad\": false, \"beta\": [-0.8840752372471408, 0.6155811769804287, 0.6431364460452087, -0.32301339020313813, -0.5412142113891235, 0.9374405298154759, -0.2980612954990996, 0.011143191960726204, 0.5115415216339614, -1.132843412924105], \"chunks\": 2000, \"d\": 10, \"max_iter\": 3, \"n\": 8000000, \"pointwise_loss\": 392098732.22278774, \"rel_error\": 0.0031039884258993638, \"seed\": 0, \"sigma\": 7.0, \"wall_time\": 145.35059690475464}, {\"approx_grad\": false, \"beta\": [-0.8840752372471408, 0.6155811769804287, 0.6431364460452087, -0.32301339020313813, -0.5412142113891235, 0.9374405298154759, -0.2980612954990996, 0.011143191960726204, 0.5115415216339614, -1.132843412924105], \"chunks\": 2000, \"d\": 10, \"max_iter\": 3, \"n\": 8000000, \"pointwise_loss\": 392098732.22278774, \"rel_error\": 0.0031039884258993633, \"seed\": 1, \"sigma\": 7.0, \"wall_time\": 147.1575231552124}, {\"approx_grad\": true, \"beta\": [-0.8271726726227692, 0.49951926049589246, 0.704512362421464, -0.2850862602450924, -0.6863328958031757, 0.8601722207246763, -0.508975427468857, -0.06062527242346647, 0.5160043265879157, -1.2795063115296572], \"chunks\": 2000, \"d\": 10, \"max_iter\": 4, \"n\": 8000000, \"pointwise_loss\": 393059855.81411487, \"rel_error\": 0.16445766469948245, \"seed\": 0, \"sigma\": 7.0, \"wall_time\": 22.655385971069336}, {\"approx_grad\": true, \"beta\": [-0.9616765072338722, 0.5343757451320267, 0.5564692898481249, -0.20815412080423262, -0.6599511653081934, 1.00051922660928, -0.47053635410121947, 0.01242918279858982, 0.5791592404773261, -1.2908888698588328], \"chunks\": 2000, \"d\": 10, \"max_iter\": 4, \"n\": 8000000, \"pointwise_loss\": 392983624.5057327, \"rel_error\": 0.156090826714074, \"seed\": 1, \"sigma\": 7.0, \"wall_time\": 19.843119859695435}, {\"approx_grad\": true, \"beta\": [-0.8311171789400179, 0.521410421564028, 0.9462671660740188, -0.3456457641986597, -0.29334977323977884, 0.9560068287397671, -0.22130619909174298, -0.0049895114959127105, 0.8629391239905513, -1.1745132638244187], \"chunks\": 2000, \"d\": 10, \"max_iter\": 4, \"n\": 8000000, \"pointwise_loss\": 394477000.17289954, \"rel_error\": 0.25672332004490195, \"seed\": 2, \"sigma\": 7.0, \"wall_time\": 20.78614115715027}, {\"approx_grad\": true, \"beta\": [-0.801139199063629, 0.16532940342659558, 0.6954539332675801, -0.28936401002679035, -0.7217519603155139, 1.091651439217816, -0.43199373306063593, -0.11066348276909367, 0.2855398064677888, -1.355992899941729], \"chunks\": 2000, \"d\": 10, \"max_iter\": 4, \"n\": 8000000, \"pointwise_loss\": 395327809.9076261, \"rel_error\": 0.3000645202606274, \"seed\": 3, \"sigma\": 7.0, \"wall_time\": 23.399425983428955}, {\"approx_grad\": true, \"beta\": [-0.7053576581148432, 0.6241930091984895, 0.5853791964035107, -0.2521215481878079, -0.8336804942073975, 0.7948488717641425, -0.278253920773729, 0.018089168528746924, 0.22836203225370016, -0.9739507392001161], \"chunks\": 2000, \"d\": 10, \"max_iter\": 4, \"n\": 8000000, \"pointwise_loss\": 394115898.34814155, \"rel_error\": 0.23886453574779987, \"seed\": 4, \"sigma\": 7.0, \"wall_time\": 22.683278799057007}, {\"approx_grad\": false, \"beta\": [-0.8840149290810264, 0.6155952392872667, 0.6429144576863511, -0.32298713855219047, -0.5413417425959787, 0.9373981467172503, -0.2979957802729954, 0.01126732427321426, 0.5117417330680154, -1.1327922152299434], \"chunks\": 2000, \"d\": 10, \"max_iter\": 4, \"n\": 8000000, \"pointwise_loss\": 392098731.9225391, \"rel_error\": 0.003110684443249633, \"seed\": 0, \"sigma\": 7.0, \"wall_time\": 189.257798910141}, {\"approx_grad\": false, \"beta\": [-0.8840149290810264, 0.6155952392872667, 0.6429144576863511, -0.32298713855219047, -0.5413417425959787, 0.9373981467172503, -0.2979957802729954, 0.01126732427321426, 0.5117417330680154, -1.1327922152299434], \"chunks\": 2000, \"d\": 10, \"max_iter\": 4, \"n\": 8000000, \"pointwise_loss\": 392098731.9225391, \"rel_error\": 0.003110684443249633, \"seed\": 1, \"sigma\": 7.0, \"wall_time\": 186.81083416938782}, {\"approx_grad\": true, \"beta\": [-0.9407374451959918, 0.5995469772485784, 0.3829421849180021, -0.5000865670359426, -0.5737165925062941, 1.1179590375704331, -0.05387964227593586, -0.062217309307136766, -0.002549105064196923, -0.8238232877423541], \"chunks\": 2000, \"d\": 10, \"max_iter\": 5, \"n\": 8000000, \"pointwise_loss\": 396586220.0473553, \"rel_error\": 0.3538455628010787, \"seed\": 0, \"sigma\": 7.0, \"wall_time\": 25.987020015716553}, {\"approx_grad\": true, \"beta\": [-0.4964414415005384, 0.6718447193669304, 0.6846026851661291, -0.010461025842725123, -0.34100874172096585, 0.9586468990831749, -0.21698443097406914, -0.13017291237255, 0.9125536530331779, -0.6087769509034419], \"chunks\": 2000, \"d\": 10, \"max_iter\": 5, \"n\": 8000000, \"pointwise_loss\": 398139813.413357, \"rel_error\": 0.4096640005375313, \"seed\": 1, \"sigma\": 7.0, \"wall_time\": 22.02472710609436}, {\"approx_grad\": true, \"beta\": [-0.8829626205083301, 1.1212609577680581, 0.5684551025155554, -0.6505691911598741, -0.9372315683339582, 0.842119326116806, -0.42196789947305235, -0.08570587191028647, 0.08368869163034408, -1.2792551410778739], \"chunks\": 2000, \"d\": 10, \"max_iter\": 5, \"n\": 8000000, \"pointwise_loss\": 398210025.845274, \"rel_error\": 0.41423200443526725, \"seed\": 2, \"sigma\": 7.0, \"wall_time\": 22.424404859542847}, {\"approx_grad\": true, \"beta\": [-0.8276390460755914, 0.6254867138609531, 0.3250671199876772, -0.17258921033454458, -0.19899367174226415, 0.802372686454787, -0.29783989269792505, 0.15952477773340476, 0.7608196944681033, -1.0561404959330822], \"chunks\": 2000, \"d\": 10, \"max_iter\": 5, \"n\": 8000000, \"pointwise_loss\": 394918666.8340963, \"rel_error\": 0.2788153092040237, \"seed\": 3, \"sigma\": 7.0, \"wall_time\": 23.654550075531006}, {\"approx_grad\": true, \"beta\": [-1.0663379690958243, 0.5137748168705915, 0.7515449969486085, -0.8673948569046709, -0.5042643587430429, 1.400751405530614, -0.45931404840107765, 0.13862134013387145, 0.3718303171564792, -1.255472840503655], \"chunks\": 2000, \"d\": 10, \"max_iter\": 5, \"n\": 8000000, \"pointwise_loss\": 397256096.00919074, \"rel_error\": 0.37940869669671184, \"seed\": 4, \"sigma\": 7.0, \"wall_time\": 30.06166672706604}, {\"approx_grad\": false, \"beta\": [-0.8840363068073565, 0.6155901299411422, 0.6429925519141578, -0.32299641680865165, -0.5412969800789815, 0.937413323810281, -0.298018917144369, 0.011223695205511832, 0.5116712128875367, -1.132810384813737], \"chunks\": 2000, \"d\": 10, \"max_iter\": 5, \"n\": 8000000, \"pointwise_loss\": 392098731.7830447, \"rel_error\": 0.003107192716169649, \"seed\": 0, \"sigma\": 7.0, \"wall_time\": 201.68372106552124}, {\"approx_grad\": false, \"beta\": [-0.8840363068073565, 0.6155901299411422, 0.6429925519141578, -0.32299641680865165, -0.5412969800789815, 0.937413323810281, -0.298018917144369, 0.011223695205511832, 0.5116712128875367, -1.132810384813737], \"chunks\": 2000, \"d\": 10, \"max_iter\": 5, \"n\": 8000000, \"pointwise_loss\": 392098731.7830447, \"rel_error\": 0.003107192716169649, \"seed\": 1, \"sigma\": 7.0, \"wall_time\": 204.56695413589478}, {\"approx_grad\": true, \"beta\": [-0.7871260042083935, 0.3961733101780085, 0.6958527709901753, -0.3142060076706436, -0.6376767465113462, 0.960881014904799, -0.31250840955083714, 0.08209674189561741, 0.515565851142813, -1.2308461475839017], \"chunks\": 2000, \"d\": 10, \"max_iter\": 6, \"n\": 8000000, \"pointwise_loss\": 392779349.549917, \"rel_error\": 0.13854581491536738, \"seed\": 0, \"sigma\": 7.0, \"wall_time\": 36.94024991989136}, {\"approx_grad\": true, \"beta\": [-0.6031158370785182, 0.6709487169693519, 0.7702163302750972, -0.4654554489878465, -0.7768143279733025, 0.9003608368757029, -0.37337880686709085, -0.07810376917934554, 0.5054273482702586, -1.1163010862552663], \"chunks\": 2000, \"d\": 10, \"max_iter\": 6, \"n\": 8000000, \"pointwise_loss\": 393612977.4438118, \"rel_error\": 0.20778431721208035, \"seed\": 1, \"sigma\": 7.0, \"wall_time\": 28.863558053970337}, {\"approx_grad\": true, \"beta\": [-0.8664624880186267, 0.4366107494098207, 0.7988147992751405, -0.2926785272602952, -0.5428020755237238, 0.8156204506426754, -0.19816820095111792, -0.034020062784066066, 0.37078322939922836, -1.137769387721782], \"chunks\": 2000, \"d\": 10, \"max_iter\": 6, \"n\": 8000000, \"pointwise_loss\": 392932772.7201239, \"rel_error\": 0.15320684314729843, \"seed\": 2, \"sigma\": 7.0, \"wall_time\": 28.86776614189148}, {\"approx_grad\": true, \"beta\": [-0.9508191916738734, 0.5803069168850657, 0.6081637705377964, -0.4316316463999406, -0.5012948159484267, 0.968552261019256, -0.24577223705663653, 0.06299812960488932, 0.2590370207057818, -1.252127732722464], \"chunks\": 2000, \"d\": 10, \"max_iter\": 6, \"n\": 8000000, \"pointwise_loss\": 392936926.844349, \"rel_error\": 0.15315234339841183, \"seed\": 3, \"sigma\": 7.0, \"wall_time\": 27.614052772521973}, {\"approx_grad\": true, \"beta\": [-0.8091028799637264, 0.45004047894098886, 0.7061596317906195, -0.3256998918927045, -0.49755334068387674, 0.9253875802460394, -0.11361957804608215, 0.24122491463255788, 0.5365126582799239, -1.1292519285384686], \"chunks\": 2000, \"d\": 10, \"max_iter\": 6, \"n\": 8000000, \"pointwise_loss\": 393111480.63514334, \"rel_error\": 0.1684587208741917, \"seed\": 4, \"sigma\": 7.0, \"wall_time\": 26.95098090171814}, {\"approx_grad\": false, \"beta\": [-0.8840368556424033, 0.6155900530326254, 0.6429948167902779, -0.3229966670664642, -0.541295637133796, 0.937413646700167, -0.2980195490623621, 0.011222412409941856, 0.5116692056516307, -1.132810842239684], \"chunks\": 2000, \"d\": 10, \"max_iter\": 6, \"n\": 8000000, \"pointwise_loss\": 392098731.78297484, \"rel_error\": 0.00310712691370173, \"seed\": 0, \"sigma\": 7.0, \"wall_time\": 229.23673105239868}, {\"approx_grad\": false, \"beta\": [-0.8840368556424033, 0.6155900530326254, 0.6429948167902779, -0.3229966670664642, -0.541295637133796, 0.937413646700167, -0.2980195490623621, 0.011222412409941856, 0.5116692056516307, -1.132810842239684], \"chunks\": 2000, \"d\": 10, \"max_iter\": 6, \"n\": 8000000, \"pointwise_loss\": 392098731.78297484, \"rel_error\": 0.00310712691370173, \"seed\": 1, \"sigma\": 7.0, \"wall_time\": 227.5106761455536}, {\"approx_grad\": true, \"beta\": [-0.9289146412636535, 0.589207394081429, 0.7012400145186803, -0.4208002014380348, -0.43386871861124676, 0.933089665396426, -0.4005966465091005, 0.06483459323652249, 0.5673006082227218, -1.225626019604088], \"chunks\": 2000, \"d\": 10, \"max_iter\": 10, \"n\": 8000000, \"pointwise_loss\": 392517421.8296241, \"rel_error\": 0.10793165451244967, \"seed\": 0, \"sigma\": 7.0, \"wall_time\": 52.241666078567505}, {\"approx_grad\": true, \"beta\": [-0.9743804887017291, 0.5410402380568015, 0.7704433890334813, -0.35185475181030945, -0.46356642753222593, 0.9481604622933278, -0.40504543584553954, 0.06890603470985716, 0.4346690052333042, -1.0354939209957497], \"chunks\": 2000, \"d\": 10, \"max_iter\": 10, \"n\": 8000000, \"pointwise_loss\": 392636046.76928383, \"rel_error\": 0.12258435738129043, \"seed\": 1, \"sigma\": 7.0, \"wall_time\": 49.099812269210815}, {\"approx_grad\": true, \"beta\": [-0.9691976834229626, 0.6475594312600076, 0.7508500065046348, -0.32858753685093967, -0.5462316324992845, 0.8922110086481335, -0.29444047752987906, -0.007789285206273814, 0.5428055414686452, -1.0599010679159921], \"chunks\": 2000, \"d\": 10, \"max_iter\": 10, \"n\": 8000000, \"pointwise_loss\": 392327909.8895484, \"rel_error\": 0.08030740531555673, \"seed\": 2, \"sigma\": 7.0, \"wall_time\": 53.91449022293091}, {\"approx_grad\": true, \"beta\": [-0.910264983109954, 0.5498026119249138, 0.7398351291370956, -0.29663314818807934, -0.6484180771824207, 0.9684016736573872, -0.10330011565812211, 0.05913743859741222, 0.7403852268345604, -1.1810385316788785], \"chunks\": 2000, \"d\": 10, \"max_iter\": 10, \"n\": 8000000, \"pointwise_loss\": 393077610.3466072, \"rel_error\": 0.16494774361606168, \"seed\": 3, \"sigma\": 7.0, \"wall_time\": 56.24317216873169}, {\"approx_grad\": true, \"beta\": [-0.8675186275412994, 0.4551384495672032, 0.604807568567115, -0.2831505725214636, -0.5860885866345346, 0.8674112928558778, -0.416804818232409, -0.1134532111679631, 0.38210984739072146, -1.075781297833849], \"chunks\": 2000, \"d\": 10, \"max_iter\": 10, \"n\": 8000000, \"pointwise_loss\": 392784311.2327362, \"rel_error\": 0.1385175667205294, \"seed\": 4, \"sigma\": 7.0, \"wall_time\": 62.55457830429077}, {\"approx_grad\": false, \"beta\": [-0.8840367015486226, 0.6155900769031977, 0.6429941913527001, -0.3229965972154067, -0.5412960062534135, 0.9374135532285174, -0.2980193731251264, 0.011222765961551154, 0.511669761456385, -1.1328107133068281], \"chunks\": 2000, \"d\": 10, \"max_iter\": 10, \"n\": 8000000, \"pointwise_loss\": 392098731.7829722, \"rel_error\": 0.003107145577421732, \"seed\": 0, \"sigma\": 7.0, \"wall_time\": 301.74830627441406}, {\"approx_grad\": false, \"beta\": [-0.8840367015486226, 0.6155900769031977, 0.6429941913527001, -0.3229965972154067, -0.5412960062534135, 0.9374135532285174, -0.2980193731251264, 0.011222765961551154, 0.511669761456385, -1.1328107133068281], \"chunks\": 2000, \"d\": 10, \"max_iter\": 10, \"n\": 8000000, \"pointwise_loss\": 392098731.7829722, \"rel_error\": 0.003107145577421732, \"seed\": 1, \"sigma\": 7.0, \"wall_time\": 283.68756103515625}, {\"approx_grad\": true, \"beta\": [-0.879226020902641, 0.7785837679404635, 0.5854598664790497, -0.2652160310122725, -0.5033963847997842, 0.9101222961481134, -0.3230869875520656, 0.0726667810267524, 0.6090404727113902, -1.015491282425678], \"chunks\": 2000, \"d\": 10, \"max_iter\": 15, \"n\": 8000000, \"pointwise_loss\": 392603554.26582336, \"rel_error\": 0.1182117141794517, \"seed\": 0, \"sigma\": 7.0, \"wall_time\": 88.87447690963745}, {\"approx_grad\": true, \"beta\": [-0.9264615710300912, 0.5793376114442538, 0.5459787720950554, -0.3485187459810788, -0.737088753955497, 1.0098549802053038, -0.2697041886515856, -0.15363978801505496, 0.38003248055376626, -1.2006135031985268], \"chunks\": 2000, \"d\": 10, \"max_iter\": 15, \"n\": 8000000, \"pointwise_loss\": 392952025.853789, \"rel_error\": 0.15435494003660002, \"seed\": 1, \"sigma\": 7.0, \"wall_time\": 82.69940185546875}, {\"approx_grad\": true, \"beta\": [-0.9206027132484186, 0.739878760663106, 0.7270095070520408, -0.38870409165038816, -0.611151472242097, 0.715224504550175, -0.441189414563162, 0.11140238970762492, 0.5215692165038381, -1.0797497203817892], \"chunks\": 2000, \"d\": 10, \"max_iter\": 15, \"n\": 8000000, \"pointwise_loss\": 393025962.4803908, \"rel_error\": 0.1625869001127405, \"seed\": 2, \"sigma\": 7.0, \"wall_time\": 92.84537315368652}, {\"approx_grad\": true, \"beta\": [-0.787874281047773, 0.6799359185856565, 0.5416287983104555, -0.37731162844352173, -0.5493588312272013, 0.8054093051255071, -0.41572109145020375, -0.10753865241218799, 0.5014026377711878, -1.027600255568214], \"chunks\": 2000, \"d\": 10, \"max_iter\": 15, \"n\": 8000000, \"pointwise_loss\": 392764542.0845498, \"rel_error\": 0.13723331555685436, \"seed\": 3, \"sigma\": 7.0, \"wall_time\": 90.04433107376099}, {\"approx_grad\": true, \"beta\": [-0.8744565169107531, 0.604436206417696, 0.46990150193534835, -0.2077964962204974, -0.5355731471077425, 0.7940482312493866, -0.2473564612549025, -0.13633621899376802, 0.5697639944472284, -1.1292764001824989], \"chunks\": 2000, \"d\": 10, \"max_iter\": 15, \"n\": 8000000, \"pointwise_loss\": 392833172.4833296, \"rel_error\": 0.14196395307157753, \"seed\": 4, \"sigma\": 7.0, \"wall_time\": 80.92051577568054}, {\"approx_grad\": false, \"beta\": [-0.8840367015486226, 0.6155900769031977, 0.6429941913527001, -0.3229965972154067, -0.5412960062534135, 0.9374135532285174, -0.2980193731251264, 0.011222765961551154, 0.511669761456385, -1.1328107133068281], \"chunks\": 2000, \"d\": 10, \"max_iter\": 15, \"n\": 8000000, \"pointwise_loss\": 392098731.7829722, \"rel_error\": 0.003107145577421732, \"seed\": 0, \"sigma\": 7.0, \"wall_time\": 265.0206460952759}, {\"approx_grad\": false, \"beta\": [-0.8840367015486226, 0.6155900769031977, 0.6429941913527001, -0.3229965972154067, -0.5412960062534135, 0.9374135532285174, -0.2980193731251264, 0.011222765961551154, 0.511669761456385, -1.1328107133068281], \"chunks\": 2000, \"d\": 10, \"max_iter\": 15, \"n\": 8000000, \"pointwise_loss\": 392098731.7829722, \"rel_error\": 0.003107145577421732, \"seed\": 1, \"sigma\": 7.0, \"wall_time\": 259.3626790046692}, {\"approx_grad\": true, \"beta\": [-0.9042011637589843, 0.7029405553155955, 0.7110057308292758, -0.3704724339708527, -0.43615660607115325, 0.9149258394399113, -0.2950345455604007, 0.12055345649464859, 0.5209865340313132, -0.9256352343235037], \"chunks\": 2000, \"d\": 10, \"max_iter\": 20, \"n\": 8000000, \"pointwise_loss\": 392750046.25591063, \"rel_error\": 0.13527978657110445, \"seed\": 0, \"sigma\": 7.0, \"wall_time\": 108.67117500305176}, {\"approx_grad\": true, \"beta\": [-0.9719847430972252, 0.561696376565175, 0.7420408719990494, -0.34829215067329033, -0.46395348346167564, 1.0175211166551112, -0.5115341156758816, 0.1059482707651871, 0.46804927489393644, -1.2154401929156522], \"chunks\": 2000, \"d\": 10, \"max_iter\": 20, \"n\": 8000000, \"pointwise_loss\": 392873611.206111, \"rel_error\": 0.14661583431505454, \"seed\": 1, \"sigma\": 7.0, \"wall_time\": 117.8687379360199}, {\"approx_grad\": true, \"beta\": [-0.7238922364851894, 0.48440091636092486, 0.5483989511032223, -0.28848806460103726, -0.5291904133328238, 0.966457907630273, -0.4282109825815653, 0.09406156210969836, 0.5318905979292694, -1.1103801072529718], \"chunks\": 2000, \"d\": 10, \"max_iter\": 20, \"n\": 8000000, \"pointwise_loss\": 392728243.39322346, \"rel_error\": 0.1327308431752905, \"seed\": 2, \"sigma\": 7.0, \"wall_time\": 114.08178377151489}, {\"approx_grad\": true, \"beta\": [-0.9438060751341817, 0.5639914647299185, 0.5472097964866391, -0.42914470178028863, -0.4964936943584284, 0.8407517247734084, -0.276133337484286, -0.0008974746619392729, 0.6534416631567372, -1.1682447312995101], \"chunks\": 2000, \"d\": 10, \"max_iter\": 20, \"n\": 8000000, \"pointwise_loss\": 392578910.63095534, \"rel_error\": 0.11538053946876133, \"seed\": 3, \"sigma\": 7.0, \"wall_time\": 117.60523080825806}, {\"approx_grad\": true, \"beta\": [-0.9221378142827324, 0.7839502847556458, 0.48953742742167167, -0.3521669540471849, -0.6727637426480056, 0.9333889365145457, -0.2866852634483545, 0.05098493792076283, 0.5652542241558753, -1.1840084441981549], \"chunks\": 2000, \"d\": 10, \"max_iter\": 20, \"n\": 8000000, \"pointwise_loss\": 392728530.95368534, \"rel_error\": 0.1327486638410105, \"seed\": 4, \"sigma\": 7.0, \"wall_time\": 118.73969268798828}, {\"approx_grad\": false, \"beta\": [-0.8840367015486226, 0.6155900769031977, 0.6429941913527001, -0.3229965972154067, -0.5412960062534135, 0.9374135532285174, -0.2980193731251264, 0.011222765961551154, 0.511669761456385, -1.1328107133068281], \"chunks\": 2000, \"d\": 10, \"max_iter\": 20, \"n\": 8000000, \"pointwise_loss\": 392098731.7829722, \"rel_error\": 0.003107145577421732, \"seed\": 0, \"sigma\": 7.0, \"wall_time\": 267.0480582714081}, {\"approx_grad\": false, \"beta\": [-0.8840367015486226, 0.6155900769031977, 0.6429941913527001, -0.3229965972154067, -0.5412960062534135, 0.9374135532285174, -0.2980193731251264, 0.011222765961551154, 0.511669761456385, -1.1328107133068281], \"chunks\": 2000, \"d\": 10, \"max_iter\": 20, \"n\": 8000000, \"pointwise_loss\": 392098731.7829722, \"rel_error\": 0.003107145577421732, \"seed\": 1, \"sigma\": 7.0, \"wall_time\": 262.2125267982483}, {\"approx_grad\": true, \"beta\": [-0.8829139465236511, 0.625917698868711, 0.5610035845124403, -0.29290341592114466, -0.5451773218871103, 0.9795265198140615, -0.2515854796694451, 3.103563452222811e-05, 0.5432008109787038, -1.100514723843423], \"chunks\": 2000, \"d\": 10, \"max_iter\": 30, \"n\": 8000000, \"pointwise_loss\": 392209513.1652502, \"rel_error\": 0.05404568835450826, \"seed\": 0, \"sigma\": 7.0, \"wall_time\": 195.57359433174133}, {\"approx_grad\": true, \"beta\": [-0.8635894856734007, 0.6248311725057915, 0.630928540527987, -0.3554994869075913, -0.5729866601651461, 0.950780176643394, -0.33218680175151266, -0.031223046549164638, 0.5121029562335224, -1.070479358507842], \"chunks\": 2000, \"d\": 10, \"max_iter\": 30, \"n\": 8000000, \"pointwise_loss\": 392176658.7978555, \"rel_error\": 0.04775019218448381, \"seed\": 1, \"sigma\": 7.0, \"wall_time\": 185.2518310546875}, {\"approx_grad\": true, \"beta\": [-0.854022071521578, 0.6031780361923385, 0.6422737722072802, -0.36969844012775666, -0.6124105215693041, 0.8853708036980525, -0.2445671723441642, 0.012866940525705072, 0.6060871681055311, -1.1131080448188513], \"chunks\": 2000, \"d\": 10, \"max_iter\": 30, \"n\": 8000000, \"pointwise_loss\": 392284167.6319425, \"rel_error\": 0.07326080170913572, \"seed\": 2, \"sigma\": 7.0, \"wall_time\": 195.97774267196655}, {\"approx_grad\": true, \"beta\": [-0.8827942112640698, 0.6166268364806528, 0.5828977133369241, -0.2929787420823192, -0.6212947554007182, 0.9072455211672876, -0.2748304348371266, 0.060147878868150505, 0.5433172421187823, -1.1121212867854486], \"chunks\": 2000, \"d\": 10, \"max_iter\": 30, \"n\": 8000000, \"pointwise_loss\": 392228262.4553793, \"rel_error\": 0.06074344122218579, \"seed\": 3, \"sigma\": 7.0, \"wall_time\": 206.70327305793762}, {\"approx_grad\": true, \"beta\": [-0.8629285267307552, 0.5671430205245978, 0.5939363055152163, -0.3136346831101807, -0.48504145478632693, 0.9055869404266276, -0.3386675928486884, 0.024980442735857315, 0.49453810816927624, -1.0657688750768044], \"chunks\": 2000, \"d\": 10, \"max_iter\": 30, \"n\": 8000000, \"pointwise_loss\": 392227507.05134, \"rel_error\": 0.059750956746531476, \"seed\": 4, \"sigma\": 7.0, \"wall_time\": 202.11913299560547}, {\"approx_grad\": false, \"beta\": [-0.8840367015486226, 0.6155900769031977, 0.6429941913527001, -0.3229965972154067, -0.5412960062534135, 0.9374135532285174, -0.2980193731251264, 0.011222765961551154, 0.511669761456385, -1.1328107133068281], \"chunks\": 2000, \"d\": 10, \"max_iter\": 30, \"n\": 8000000, \"pointwise_loss\": 392098731.7829722, \"rel_error\": 0.003107145577421732, \"seed\": 0, \"sigma\": 7.0, \"wall_time\": 256.4530289173126}, {\"approx_grad\": false, \"beta\": [-0.8840367015486226, 0.6155900769031977, 0.6429941913527001, -0.3229965972154067, -0.5412960062534135, 0.9374135532285174, -0.2980193731251264, 0.011222765961551154, 0.511669761456385, -1.1328107133068281], \"chunks\": 2000, \"d\": 10, \"max_iter\": 30, \"n\": 8000000, \"pointwise_loss\": 392098731.7829722, \"rel_error\": 0.003107145577421732, \"seed\": 1, \"sigma\": 7.0, \"wall_time\": 271.3524467945099}, {\"approx_grad\": true, \"beta\": [-0.9144115552180049, 0.603015498404459, 0.5713064287707974, -0.34545575275615004, -0.6424885972652146, 0.8931970146541598, -0.2331797791375983, -0.02907522461562066, 0.46925773932688525, -1.1219401556826756], \"chunks\": 2000, \"d\": 10, \"max_iter\": 40, \"n\": 8000000, \"pointwise_loss\": 392312074.65794194, \"rel_error\": 0.0779374005493023, \"seed\": 0, \"sigma\": 7.0, \"wall_time\": 264.6048061847687}, {\"approx_grad\": true, \"beta\": [-0.8973887214484578, 0.625628029190906, 0.6978309024390227, -0.32447667279242576, -0.5497618115288487, 0.9206360070908075, -0.2452788563571229, -0.02033314724967066, 0.4922056424940327, -1.0984816585576087], \"chunks\": 2000, \"d\": 10, \"max_iter\": 40, \"n\": 8000000, \"pointwise_loss\": 392170503.4222042, \"rel_error\": 0.04541517197333465, \"seed\": 1, \"sigma\": 7.0, \"wall_time\": 269.0821568965912}, {\"approx_grad\": true, \"beta\": [-0.8823393235027647, 0.5492224550551256, 0.7019713370961441, -0.3435038812509715, -0.5853327207643598, 0.9835950273865081, -0.27844770423239806, -0.055921215247640724, 0.4887343295786643, -1.1194613758252137], \"chunks\": 2000, \"d\": 10, \"max_iter\": 40, \"n\": 8000000, \"pointwise_loss\": 392242635.91384065, \"rel_error\": 0.06373818373096356, \"seed\": 2, \"sigma\": 7.0, \"wall_time\": 256.9263689517975}, {\"approx_grad\": true, \"beta\": [-0.8279929314291558, 0.5785406092292175, 0.6888273335787916, -0.354480652670551, -0.5278944818518453, 0.9341898266525525, -0.28833210613646065, 0.12581324637761054, 0.47795866600785925, -1.159984780697032], \"chunks\": 2000, \"d\": 10, \"max_iter\": 40, \"n\": 8000000, \"pointwise_loss\": 392281900.7215971, \"rel_error\": 0.0729637904656608, \"seed\": 3, \"sigma\": 7.0, \"wall_time\": 243.45107984542847}, {\"approx_grad\": true, \"beta\": [-0.9078119382540301, 0.5449053559924391, 0.6857141065778634, -0.33039721408797773, -0.546340031592623, 0.9589346749302237, -0.2378234765841817, 0.01719609554537357, 0.4794984490149854, -1.1048310139934197], \"chunks\": 2000, \"d\": 10, \"max_iter\": 40, \"n\": 8000000, \"pointwise_loss\": 392205987.4079259, \"rel_error\": 0.05482705526360315, \"seed\": 4, \"sigma\": 7.0, \"wall_time\": 247.5673270225525}, {\"approx_grad\": false, \"beta\": [-0.8840367015486226, 0.6155900769031977, 0.6429941913527001, -0.3229965972154067, -0.5412960062534135, 0.9374135532285174, -0.2980193731251264, 0.011222765961551154, 0.511669761456385, -1.1328107133068281], \"chunks\": 2000, \"d\": 10, \"max_iter\": 40, \"n\": 8000000, \"pointwise_loss\": 392098731.7829722, \"rel_error\": 0.003107145577421732, \"seed\": 0, \"sigma\": 7.0, \"wall_time\": 261.3412778377533}, {\"approx_grad\": false, \"beta\": [-0.8840367015486226, 0.6155900769031977, 0.6429941913527001, -0.3229965972154067, -0.5412960062534135, 0.9374135532285174, -0.2980193731251264, 0.011222765961551154, 0.511669761456385, -1.1328107133068281], \"chunks\": 2000, \"d\": 10, \"max_iter\": 40, \"n\": 8000000, \"pointwise_loss\": 392098731.7829722, \"rel_error\": 0.003107145577421732, \"seed\": 1, \"sigma\": 7.0, \"wall_time\": 243.74875688552856}, {\"approx_grad\": true, \"beta\": [-0.8854073087268456, 0.6172178631773021, 0.6450145855969107, -0.32189737380866645, -0.5447108173125104, 0.9348775941459327, -0.29811605878941355, 0.013981237885834294, 0.5143839539544327, -1.1370906424701517], \"chunks\": 2000, \"d\": 10, \"max_iter\": 80, \"n\": 8000000, \"pointwise_loss\": 392099221.5401707, \"rel_error\": 0.005300432688336602, \"seed\": 0, \"sigma\": 7.0, \"wall_time\": 925.5044710636139}, {\"approx_grad\": true, \"beta\": [-0.8818837818078422, 0.613442176739526, 0.6419890655824715, -0.32382572208301896, -0.5415035086126159, 0.9344285595975396, -0.29847295104227245, 0.00894236421481264, 0.5093510115674734, -1.1237886908468033], \"chunks\": 2000, \"d\": 10, \"max_iter\": 80, \"n\": 8000000, \"pointwise_loss\": 392099628.5426975, \"rel_error\": 0.006657967894667196, \"seed\": 1, \"sigma\": 7.0, \"wall_time\": 1053.5675048828125}, {\"approx_grad\": true, \"beta\": [-0.8841524006885965, 0.6192790187917091, 0.6435924079837052, -0.3224913282377281, -0.5405742823931305, 0.9331562752015514, -0.303804688437872, 0.012429471065645424, 0.5122031575325743, -1.1330440943007851], \"chunks\": 2000, \"d\": 10, \"max_iter\": 80, \"n\": 8000000, \"pointwise_loss\": 392099277.1428425, \"rel_error\": 0.005527006304765571, \"seed\": 2, \"sigma\": 7.0, \"wall_time\": 1059.2264840602875}, {\"approx_grad\": true, \"beta\": [-0.8825335911850323, 0.6098598438138241, 0.6441103657737104, -0.3210903115661359, -0.5390939638365305, 0.933140324813863, -0.2976977255708058, 0.012605030831765042, 0.5152813665905337, -1.1273924354863225], \"chunks\": 2000, \"d\": 10, \"max_iter\": 80, \"n\": 8000000, \"pointwise_loss\": 392099592.01777464, \"rel_error\": 0.005867858009065959, \"seed\": 3, \"sigma\": 7.0, \"wall_time\": 1028.4987251758575}, {\"approx_grad\": true, \"beta\": [-0.8871790409057009, 0.6196366303695788, 0.6431369705966894, -0.3268682376868445, -0.5390371654462093, 0.9393368181590982, -0.294816062643927, 0.010144475140230992, 0.5120191693819195, -1.1346528130151012], \"chunks\": 2000, \"d\": 10, \"max_iter\": 80, \"n\": 8000000, \"pointwise_loss\": 392099251.88777584, \"rel_error\": 0.004409698151563556, \"seed\": 4, \"sigma\": 7.0, \"wall_time\": 1020.5811431407928}, {\"approx_grad\": false, \"beta\": [-0.8840367015486226, 0.6155900769031977, 0.6429941913527001, -0.3229965972154067, -0.5412960062534135, 0.9374135532285174, -0.2980193731251264, 0.011222765961551154, 0.511669761456385, -1.1328107133068281], \"chunks\": 2000, \"d\": 10, \"max_iter\": 80, \"n\": 8000000, \"pointwise_loss\": 392098731.7829722, \"rel_error\": 0.003107145577421732, \"seed\": 0, \"sigma\": 7.0, \"wall_time\": 273.7842438220978}, {\"approx_grad\": false, \"beta\": [-0.8840367015486226, 0.6155900769031977, 0.6429941913527001, -0.3229965972154067, -0.5412960062534135, 0.9374135532285174, -0.2980193731251264, 0.011222765961551154, 0.511669761456385, -1.1328107133068281], \"chunks\": 2000, \"d\": 10, \"max_iter\": 80, \"n\": 8000000, \"pointwise_loss\": 392098731.7829722, \"rel_error\": 0.003107145577421732, \"seed\": 1, \"sigma\": 7.0, \"wall_time\": 267.38317227363586}]}};\n",
"var selector = \"#7100bd77-4025-4247-acdd-2c798f08cbc7\";\n",
"var type = \"vega-lite\";\n",
"\n",
"var output_area = this;\n",
"require(['nbextensions/jupyter-vega/index'], function(vega) {\n",
" vega.render(selector, spec, type, output_area);\n",
"}, function (err) {\n",
" if (err.requireType !== 'scripterror') {\n",
" throw(err);\n",
" }\n",
"});\n"
]
},
"metadata": {
"jupyter-vega": "#7100bd77-4025-4247-acdd-2c798f08cbc7"
},
"output_type": "display_data"
},
{
"data": {
"image/png": "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"
},
"metadata": {
"jupyter-vega": "#7100bd77-4025-4247-acdd-2c798f08cbc7"
},
"output_type": "display_data"
}
],
"source": [
"Chart(df).mark_line().encode(\n",
" x=X('wall_time', scale=Scale(type='log')),\n",
" y=Y('pointwise_loss',\n",
" scale=Scale(domain=[getattr(df.pointwise_loss, f)()\n",
" for f in ['min', 'max']],\n",
" clamp=True)),\n",
" color='approx_grad')"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"collapsed": true
},
"outputs": [],
"source": []
}
],
"metadata": {
"kernelspec": {
"display_name": "Python [default]",
"language": "python",
"name": "python3"
},
"language_info": {
"codemirror_mode": {
"name": "ipython",
"version": 3
},
"file_extension": ".py",
"mimetype": "text/x-python",
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.6.1"
},
"notify_time": "5"
},
"nbformat": 4,
"nbformat_minor": 2
}
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment