Merge branch 'Support_Lyft' into 'master'

Add API to support Lyft

See merge request open-mmlab/mmdet.3d!106

data/lyft/test.txt

+host-a004-lidar0-1233944976297786786-1233945001198600096
+host-a004-lidar0-1233941213298388436-1233941238199278096
+host-a011-lidar0-1234644740299444586-1234644765198350636
+host-a004-lidar0-1233601648198462856-1233601673098488556
+host-a011-lidar0-1232746157199035666-1232746182098240026
+host-a009-lidar0-1236017375097801876-1236017399997624556
+host-a011-lidar0-1233962329198070906-1233962354099359636
+host-a007-lidar0-1233523259198624906-1233523284098400466
+host-a004-lidar0-1233427443198938856-1233427468098241556
+host-a007-lidar0-1233952760198014706-1233952785098360666
+host-a004-lidar0-1232831114198639196-1232831139098689906
+host-a011-lidar0-1233959097298631226-1233959122199133096
+host-a011-lidar0-1232736280198893346-1232736305098557556
+host-a008-lidar0-1235757362198327706-1235757387098420756
+host-a007-lidar0-1233947142198222706-1233947167098201666
+host-a009-lidar0-1234715568198453346-1234715593098413676
+host-a004-lidar0-1232911076298600176-1232911101199054856
+host-a011-lidar0-1236103955299212856-1236103980199733906
+host-a007-lidar0-1237572884098557116-1237572908997807546
+host-a011-lidar0-1233964240298899786-1233964265199453556
+host-a004-lidar0-1233079262298510786-1233079287198908466
+host-a004-lidar0-1233687421297771586-1233687446198032636
+host-a008-lidar0-1235780110998543436-1235780135897779096
+host-a011-lidar0-1232488772298383876-1232488797198160346
+host-a004-lidar0-1231878498298402636-1231878523198702316
+host-a015-lidar0-1235952792198484436-1235952817098084116
+host-a004-lidar0-1235945794298780666-1235945819198802346
+host-a011-lidar0-1233963833297804226-1233963858199264096
+host-a004-lidar0-1233506417198072656-1233506442098833106
+host-a009-lidar0-1231801138198800186-1231801163098659866
+host-a011-lidar0-1232751280197949666-1232751305099090996
+host-a004-lidar0-1233946658298833786-1233946683199182096
+host-a004-lidar0-1233687742297900586-1233687767199090986
+host-a004-lidar0-1232905595299117226-1232905620198562226
+host-a004-lidar0-1233961285198272466-1233961310098968226
+host-a011-lidar0-1233085141298793636-1233085166198948316
+host-a010-lidar0-1232314667198394906-1232314692099896986
+host-a007-lidar0-1230931253199029066-1230931278098162746
+host-a007-lidar0-1232995403098212786-1232995427998257586
+host-a011-lidar0-1236106341198519026-1236106366098688706
+host-a004-lidar0-1233962919198446116-1233962944098602196
+host-a011-lidar0-1232744874197271346-1232744899099026346
+host-a011-lidar0-1233957289298013416-1233957314197859536
+host-a011-lidar0-1236039862198450906-1236039887099020026
+host-a011-lidar0-1233956316299257226-1233956341199458096
+host-a004-lidar0-1233521279299019346-1233521304198708656
+host-a007-lidar0-1233005263198512026-1233005288098233756
+host-a007-lidar0-1232995508097871466-1232995532998255586
+host-a011-lidar0-1232486236299484666-1232486261197508346
+host-a011-lidar0-1233959639198614466-1233959664098178176
+host-a007-lidar0-1233015278098306756-1233015302998395786
+host-a004-lidar0-1235858020298070196-1235858045199172906
+host-a011-lidar0-1236105788197720026-1236105813099285026
+host-a004-lidar0-1233439235298936546-1233439260198396226
+host-a004-lidar0-1232987587298107736-1232987612198297786
+host-a011-lidar0-1236103100299832666-1236103125199414996
+host-a015-lidar0-1235952391197848436-1235952416098412116
+host-a004-lidar0-1233521391298537346-1233521416198295656
+host-a004-lidar0-1232991769198099466-1232991794097984146
+host-a004-lidar0-1233953140198673466-1233953165099150176
+host-a004-lidar0-1233081052299073106-1233081077199184436
+host-a007-lidar0-1233956133198909706-1233956158098315666
+host-a011-lidar0-1235868868199238666-1235868893097732316
+host-a007-lidar0-1233953460198410706-1233953485098842786
+host-a011-lidar0-1233961831198549906-1233961856098369176
+host-a011-lidar0-1233514529198419226-1233514554098449346
+host-a007-lidar0-1230939239197974066-1230939264099426746
+host-a011-lidar0-1233091237198666226-1233091262098974026
+host-a004-lidar0-1233442845299348546-1233442870197585226
+host-a009-lidar0-1236020549098474856-1236020573999334586
+host-a011-lidar0-1234024976198295026-1234025001098882226
+host-a011-lidar0-1232907883299065316-1232907908199299976
+host-a004-lidar0-1233601706199079226-1233601731099035556
+host-a011-lidar0-1233082653297382196-1233082678199113346
+host-a011-lidar0-1236094968298801346-1236094993199079346
+host-a007-lidar0-1233007278198865146-1233007303098536106
+host-a004-lidar0-1232740139298853106-1232740164198443466
+host-a007-lidar0-1232840098198209026-1232840123098175986
+host-a007-lidar0-1232491902199024026-1232491927099237736
+host-a004-lidar0-1233447945198501226-1233447970098212556
+host-a004-lidar0-1233963468197846466-1233963493098405196
+host-a011-lidar0-1232497134299369316-1232497159198333026
+host-a007-lidar0-1233683913198228346-1233683938097909026
+host-a004-lidar0-1233965315199029466-1233965340097937546
+host-a011-lidar0-1236106510198830536-1236106535098940676
+host-a011-lidar0-1234031888198360146-1234031913099866226
+host-a011-lidar0-1232483258298822666-1232483283199104026
+host-a011-lidar0-1233963883298333416-1233963908199031906
+host-a011-lidar0-1232411607198546106-1232411632099155466
+host-a009-lidar0-1236120835298099026-1236120860197794756
+host-a004-lidar0-1233089444197814786-1233089469098441116
+host-a004-lidar0-1233946716298797436-1233946741199280636
+host-a004-lidar0-1233443288299030176-1233443313198834856
+host-a004-lidar0-1233088880198387436-1233088905098238096
+host-a011-lidar0-1233956935299563906-1233956960199421666
+host-a009-lidar0-1236020134098279876-1236020158999209906
+host-a011-lidar0-1236094859299152316-1236094884199534466
+host-a008-lidar0-1235758174198256146-1235758199098282106
+host-a011-lidar0-1236121184299486346-1236121209198975996
+host-a011-lidar0-1233961523199115586-1233961548099271666
+host-a004-lidar0-1234732754197914906-1234732779097924636
+host-a009-lidar0-1236121511198446736-1236121536098500436
+host-a008-lidar0-1236034565298114876-1236034590198781906
+host-a006-lidar0-1236098659198274536-1236098684097992536
+host-a009-lidar0-1236013611198536176-1236013636098119856
+host-a015-lidar0-1233957342198356906-1233957367097686986
+host-a011-lidar0-1234028898199503706-1234028923098637226
+host-a007-lidar0-1234564891198141676-1234564916098497756
+host-a011-lidar0-1233512411200132666-1233512436099006556
+host-a011-lidar0-1232839888198187226-1232839913098323346
+host-a011-lidar0-1233959147299652876-1233959172199345586
+host-a011-lidar0-1233515019199540346-1233515044098381026
+host-a007-lidar0-1233621306298394226-1233621331197987026
+host-a009-lidar0-1236018631099309906-1236018655997712116
+host-a011-lidar0-1233956770298774786-1233956795198595906
+host-a011-lidar0-1234031012198940656-1234031037098235226
+host-a011-lidar0-1232834951198260666-1232834976099168996
+host-a004-lidar0-1231810077298582906-1231810102198764586
+host-a015-lidar0-1235432061197143666-1235432086098855666
+host-a004-lidar0-1233955370199325146-1233955395099287546
+host-a007-lidar0-1232739122099041026-1232739146998701786
+host-a008-lidar0-1231272360198562866-1231272385098213606
+host-a009-lidar0-1234043856197659756-1234043881098769786
+host-a011-lidar0-1236123850299533346-1236123875199183536
+host-a004-lidar0-1232829779198385176-1232829804099132906
+host-a004-lidar0-1234046078298238146-1234046103198417226
+host-a011-lidar0-1236118880299007316-1236118905198814556
+host-a011-lidar0-1232839522198475666-1232839547098176996
+host-a004-lidar0-1233683047198434616-1233683072097863296
+host-a011-lidar0-1232483134298411316-1232483159199017026
+host-a011-lidar0-1232835366199044316-1232835391098346996
+host-a011-lidar0-1236037587198526556-1236037612099015026
+host-a004-lidar0-1233514041198025676-1233514066098168106
+host-a011-lidar0-1233514916198928666-1233514941098966346
+host-a007-lidar0-1233952607199156666-1233952632098069666
+host-a011-lidar0-1234655115198110986-1234655140099130346
+host-a015-lidar0-1236124264998206536-1236124289898560636
+host-a011-lidar0-1234644690299314096-1234644715199118176
+host-a011-lidar0-1235952056198818296-1235952081099333786
+host-a011-lidar0-1232748626199298346-1232748651098789346
+host-a007-lidar0-1232475199297946856-1232475224198641556
+host-a007-lidar0-1233525441098373116-1233525465998053546
+host-a008-lidar0-1235761573098617226-1235761597998604906
+host-a004-lidar0-1232991714198772466-1232991739097967096
+host-a004-lidar0-1233687683299064556-1233687708198041986
+host-a007-lidar0-1233013003098382226-1233013027998697786
+host-a004-lidar0-1233439306298453176-1233439331197984226
+host-a004-lidar0-1233424178198144226-1233424203098692536
+host-a006-lidar0-1236097465198430226-1236097490097938026
+host-a011-lidar0-1232841048199446666-1232841073098839996
+host-a011-lidar0-1233516714199191856-1233516739099730116
+host-a007-lidar0-1234311713998886226-1234311738898096786
+host-a007-lidar0-1233536831198375786-1233536856097999586
+host-a004-lidar0-1233429274199024876-1233429299099306906
+host-a004-lidar0-1233959153198568116-1233959178098071546
+host-a011-lidar0-1234552220298342756-1234552245199496436
+host-a011-lidar0-1233521972198420856-1233521997098456536
+host-a006-lidar0-1236097911198275876-1236097936098518536
+host-a011-lidar0-1233689347298085226-1233689372198299556
+host-a004-lidar0-1232916227198731226-1232916252098964536
+host-a004-lidar0-1235943691298894636-1235943716198114296
+host-a011-lidar0-1232905354298416876-1232905379198827346
+host-a004-lidar0-1232825454199122176-1232825479098530856
+host-a004-lidar0-1235865310198314226-1235865335098810536
+host-a007-lidar0-1233511892098885096-1233511916998404176
+host-a004-lidar0-1235952483297492666-1235952508198691346
+host-a007-lidar0-1236123864198260676-1236123889098156106
+host-a011-lidar0-1232751443198845666-1232751468099206346
+host-a011-lidar0-1233078523199259666-1233078548098819996
+host-a004-lidar0-1233618447298368586-1233618472198344666
+host-a007-lidar0-1230678335199240106-1230678360099285186
+host-a004-lidar0-1233508848199005656-1233508873098993106
+host-a011-lidar0-1233958777298868906-1233958802198879556
+host-a007-lidar0-1233507949098749096-1233507973999200196
+host-a004-lidar0-1233953506198624096-1233953531097898546
+host-a015-lidar0-1236103405197509196-1236103430098109856
+host-a007-lidar0-1233620674297428346-1233620699199028906
+host-a012-lidar0-1235936434298098786-1235936459197995466
+host-a011-lidar0-1233514154199609666-1233514179099494586
+host-a009-lidar0-1231184014198521956-1231184039098791066
+host-a004-lidar0-1236019079298483436-1236019104198926466
+host-a006-lidar0-1236037423198601636-1236037448098691666
+host-a004-lidar0-1231888238197475346-1231888263098211346
+host-a010-lidar0-1232317276099120706-1232317300998122666
+host-a004-lidar0-1232815694198030546-1232815719097692906
+host-a007-lidar0-1233954463198147636-1233954488098173786
+host-a004-lidar0-1232923791198849226-1232923816098237586
+host-a011-lidar0-1236106196198429346-1236106221098810676
+host-a015-lidar0-1236124649998497586-1236124674897796616
+host-a004-lidar0-1232386904298052116-1232386929198958786
+host-a014-lidar0-1235764007298887586-1235764032198224636
+host-a011-lidar0-1233961953198251556-1233961978098299986
+host-a015-lidar0-1234646248197843296-1234646273098426346
+host-a004-lidar0-1232823719198617196-1232823744099030906
+host-a006-lidar0-1232910084198374756-1232910109098099436
+host-a008-lidar0-1231535639098399806-1231535663998755486
+host-a006-lidar0-1232909811198488106-1232909836098046436
+host-a011-lidar0-1234472721299135436-1234472746198770146
+host-a004-lidar0-1233941123298835466-1233941148198539096
+host-a004-lidar0-1232842293198603176-1232842318098888226
+host-a011-lidar0-1235503241198830666-1235503266098892346
+host-a007-lidar0-1233007421198689676-1233007446098306106
+host-a011-lidar0-1235942334298364116-1235942359197798196
+host-a011-lidar0-1233964524299020906-1233964549198793586
+host-a011-lidar0-1236094810299383666-1236094835198207026
+host-a004-lidar0-1233439181298702546-1233439206198541786
+host-a004-lidar0-1233429602198887906-1233429627098124906
+host-a011-lidar0-1233688997299519416-1233689022199633116
+host-a011-lidar0-1233088421199045856-1233088446099117556
+host-a011-lidar0-1235866237298160096-1235866262198355986
+host-a004-lidar0-1232825508197625546-1232825533098566906
+host-a007-lidar0-1233620829298440226-1233620854197980906
+host-a011-lidar0-1233962268199095556-1233962293098293666
+host-a009-lidar0-1236017846098562856-1236017870998139586
+host-a004-lidar0-1233521339297626996-1233521364198425656
+host-a011-lidar0-1233962379199267906-1233962404098719666
+host-a007-lidar0-1234043217198181466-1234043242098437196
+host-a007-lidar0-1233952679199347706-1233952704098573316
\ No newline at end of file

data/lyft/train.txt

+host-a101-lidar0-1241893239199111666-1241893264098084346
+host-a006-lidar0-1236037883198113706-1236037908098879296
+host-a011-lidar0-1235950297199142196-1235950322099405416
+host-a012-lidar0-1235937130198577346-1235937155098071026
+host-a101-lidar0-1240875136198305786-1240875161098795094
+host-a011-lidar0-1232752461198357666-1232752486099793996
+host-a102-lidar0-1241468916398562586-1241468941298742334
+host-a009-lidar0-1236013297198927176-1236013322098616226
+host-a008-lidar0-1235777217098625786-1235777241998473466
+host-a009-lidar0-1236015606098226876-1236015630998447586
+host-a011-lidar0-1236119823299280856-1236119848199397346
+host-a011-lidar0-1233963416198495906-1233963441098571986
+host-a102-lidar0-1241904536298706586-1241904561198322666
+host-a004-lidar0-1236021339298624436-1236021364198408146
+host-a011-lidar0-1233512833198873346-1233512858098831906
+host-a007-lidar0-1232470052198454586-1232470077098888666
+host-a004-lidar0-1233961181197891466-1233961206097713176
+host-a011-lidar0-1232755934199060666-1232755959099356536
+host-a011-lidar0-1235866707299065096-1235866732198912176
+host-a007-lidar0-1234740239998520226-1234740264899399906
+host-a011-lidar0-1234031754199586656-1234031779098211226
+host-a011-lidar0-1233090630199206666-1233090655098843996
+host-a004-lidar0-1233617933297688906-1233617958198483986
+host-a004-lidar0-1233693263298064536-1233693288197865986
+host-a007-lidar0-1233510590098435146-1233510614998778546
+host-a102-lidar0-1241548686398885894-1241548711298381586
+host-a015-lidar0-1233957265198932906-1233957290097795666
+host-a009-lidar0-1236118456298488636-1236118481198250736
+host-a007-lidar0-1230936221299185986-1230936246198612066
+host-a011-lidar0-1232835293199223316-1232835318097646346
+host-a102-lidar0-1242755400298847586-1242755425198579666
+host-a004-lidar0-1233422539197434856-1233422564099152556
+host-a101-lidar0-1242580003398722214-1242580028299473214
+host-a007-lidar0-1233954992197900986-1233955017097982666
+host-a004-lidar0-1233685315298191906-1233685340197986666
+host-a008-lidar0-1236015187198059026-1236015212098657616
+host-a011-lidar0-1232905783299194316-1232905808199600976
+host-a004-lidar0-1233682997198306636-1233683022098032666
+host-a006-lidar0-1236038131197892706-1236038156098552296
+host-a007-lidar0-1233955131199105986-1233955156098128666
+host-a007-lidar0-1234737958998212856-1234737983898305906
+host-a102-lidar0-1241878200398362906-1241878225298546586
+host-a011-lidar0-1232492192198952026-1232492217098860706
+host-a009-lidar0-1231200854198312986-1231200879098460066
+host-a011-lidar0-1232839939198224316-1232839964099010996
+host-a101-lidar0-1242493624298705334-1242493649198973302
+host-a101-lidar0-1243095610299140346-1243095635198749774
+host-a011-lidar0-1234553365299813296-1234553390199271786
+host-a011-lidar0-1233087918198597316-1233087943098472996
+host-a004-lidar0-1232923266198326856-1232923291098716906
+host-a007-lidar0-1233689791098884906-1233689815997978986
+host-a011-lidar0-1232753667198514346-1232753692099110026
+host-a004-lidar0-1232987652297797736-1232987677198272416
+host-a011-lidar0-1232841333199326666-1232841358099777556
+host-a011-lidar0-1235931114299585906-1235931139198414556
+host-a004-lidar0-1234051595199019546-1234051620099157876
+host-a004-lidar0-1232815252198642176-1232815277099387856
+host-a011-lidar0-1234025165199395146-1234025190098246226
+host-a004-lidar0-1233618009298519906-1233618034198134636
+host-a004-lidar0-1235947081298918616-1235947106198383666
+host-a004-lidar0-1233535955298751556-1233535980198120666
+host-a101-lidar0-1241889710198571346-1241889735098952214
+host-a004-lidar0-1233442991299115176-1233443016198370876
+host-a007-lidar0-1233007769198478676-1233007794098024226
+host-a011-lidar0-1233688931299349876-1233688956199708556
+host-a004-lidar0-1236018644297896466-1236018669198234096
+host-a004-lidar0-1233947108297817786-1233947133198765096
+host-a004-lidar0-1235944794298214636-1235944819199047666
+host-a007-lidar0-1233524852199250466-1233524877097605116
+host-a008-lidar0-1235776284099006226-1235776308998255906
+host-a101-lidar0-1242748817298870302-1242748842198675302
+host-a007-lidar0-1233529706297796756-1233529731198779786
+host-a011-lidar0-1232909940298708666-1232909965199312906
+host-a011-lidar0-1232412236198491106-1232412261098202466
+host-a011-lidar0-1236038911199051026-1236038936099738706
+host-a008-lidar0-1236013033198326026-1236013058097878706
+host-a101-lidar0-1242753236298794334-1242753261198702302
+host-a011-lidar0-1235933627299543026-1235933652198559196
+host-a004-lidar0-1233620260298411906-1233620285198333986
+host-a101-lidar0-1242144886399176654-1242144911299066654
+host-a102-lidar0-1242684244198410786-1242684269098866094
+host-a011-lidar0-1232738595197630316-1232738620098495346
+host-a007-lidar0-1233508020098603466-1233508044998550666
+host-a102-lidar0-1242510597398871466-1242510622298829226
+host-a101-lidar0-1243102866399012786-1243102891298922466
+host-a009-lidar0-1236020733098808906-1236020757997885536
+host-a011-lidar0-1232745770199275666-1232745795099558556
+host-a101-lidar0-1241886983298988182-1241887008198992182
+host-a004-lidar0-1233693191297468536-1233693216198791636
+host-a017-lidar0-1236119797198435536-1236119822098126616
+host-a007-lidar0-1233515286998507736-1233515311898052226
+host-a007-lidar0-1230672860198383106-1230672885099108186
+host-a004-lidar0-1233014343198383656-1233014368098267106
+host-a004-lidar0-1233516150198382706-1233516175098720106
+host-a007-lidar0-1230485630199365106-1230485655099030186
+host-a004-lidar0-1232838138198668196-1232838163098533856
+host-a011-lidar0-1236037921199248346-1236037946099430676
+host-a011-lidar0-1233081021198156856-1233081046098157026
+host-a004-lidar0-1232825386198046196-1232825411098056856
+host-a017-lidar0-1236118981198431906-1236119006097572636
+host-a015-lidar0-1236103725197932106-1236103750098792856
+host-a007-lidar0-1233515591998210666-1233515616898664876
+host-a007-lidar0-1232736726098319676-1232736750999473736
+host-a004-lidar0-1233011743198634026-1233011768099043756
+host-a101-lidar0-1242749258298976334-1242749283199254466
+host-a007-lidar0-1232490767197744146-1232490792098614756
+host-a011-lidar0-1234466278299425556-1234466303199121616
+host-a011-lidar0-1236122349298071346-1236122374198621346
+host-a101-lidar0-1241216089098610756-1241216113999079830
+host-a004-lidar0-1233602012198802906-1233602037098984906
+host-a004-lidar0-1233421984198960226-1233422009098905556
+host-a004-lidar0-1233685221298830296-1233685246198471636
+host-a017-lidar0-1236118873198607026-1236118898097847616
+host-a011-lidar0-1233522430198228856-1233522455098303536
+host-a101-lidar0-1242748985299274334-1242749010198891466
+host-a004-lidar0-1234047743298156466-1234047768199244736
+host-a015-lidar0-1235423635198474636-1235423660098038666
+host-a004-lidar0-1233955731199067146-1233955756098663196
+host-a009-lidar0-1236118555299408756-1236118580198077756
+host-a007-lidar0-1234551913098444106-1234551937998728436
+host-a101-lidar0-1241472407298206026-1241472432198409706
+host-a011-lidar0-1236039783198411996-1236039808098336676
+host-a015-lidar0-1236112601097782876-1236112625998366556
+host-a012-lidar0-1237329862198269106-1237329887099105436
+host-a007-lidar0-1231093036199514746-1231093061099651426
+host-a007-lidar0-1233960442199212986-1233960467099041666
+host-a009-lidar0-1236123717198611786-1236123742097892436
+host-a009-lidar0-1237581206198345466-1237581231098504546
+host-a011-lidar0-1233964282297830226-1233964307199768556
+host-a011-lidar0-1232752778198249666-1232752803099491536
+host-a102-lidar0-1242662270298972894-1242662295198395706
+host-a011-lidar0-1232485958298280666-1232485983200054996
+host-a005-lidar0-1231201437298603426-1231201462198815506
+host-a007-lidar0-1233510205099156466-1233510229998563196
+host-a102-lidar0-1242754954298696742-1242754979198120666
+host-a004-lidar0-1233444816298625546-1233444841198302226
+host-a004-lidar0-1232842166198181226-1232842191097390226
+host-a004-lidar0-1232833308197903226-1232833333099155226
+host-a011-lidar0-1232401360198078026-1232401385098379106
+host-a011-lidar0-1236123625299234316-1236123650197952996
+host-a004-lidar0-1233427004198119856-1233427029098998676
+host-a102-lidar0-1242749461398477906-1242749486298996742
+host-a102-lidar0-1242150795498255026-1242150820398693830
+host-a011-lidar0-1232731591298977986-1232731616197888346
+host-a011-lidar0-1233964369297973906-1233964394199186906
+host-a011-lidar0-1232837612197878316-1232837637099721996
+host-a101-lidar0-1241462203298815998-1241462228198805706
+host-a009-lidar0-1236014648198307196-1236014673098985906
+host-a007-lidar0-1233956183198377616-1233956208098469296
+host-a004-lidar0-1232817645198462196-1232817670098101226
\ No newline at end of file

data/lyft/val.txt

+host-a004-lidar0-1233080749298771736-1233080774198118416
+host-a004-lidar0-1232905197298264546-1232905222198133856
+host-a011-lidar0-1232732468299489666-1232732493199050666
+host-a101-lidar0-1241561147998866622-1241561172899320654
+host-a006-lidar0-1237322885198285226-1237322910098576786
+host-a004-lidar0-1233963848198981116-1233963873098642176
+host-a011-lidar0-1232752543198025666-1232752568099126026
+host-a004-lidar0-1232842367198056546-1232842392097783226
+host-a004-lidar0-1233615989298293586-1233616014198854636
+host-a011-lidar0-1233965426299054906-1233965451199121906
+host-a011-lidar0-1236104034298928316-1236104059198988026
+host-a007-lidar0-1233946614199227636-1233946639098289666
+host-a015-lidar0-1235423696198069636-1235423721098551296
+host-a004-lidar0-1233014843199117706-1233014868098023786
+host-a011-lidar0-1236093962299300416-1236093987199363346
+host-a011-lidar0-1234639296198260986-1234639321099417316
+host-a011-lidar0-1233524871199389346-1233524896098591466
+host-a011-lidar0-1235933781298838116-1235933806199517736
+host-a011-lidar0-1233965312298542226-1233965337198958586
+host-a011-lidar0-1233090567199118316-1233090592098933996
+host-a007-lidar0-1233621256298511876-1233621281197988026
+host-a007-lidar0-1233079617197863906-1233079642098533586
+host-a015-lidar0-1236112516098396876-1236112540999028556
+host-a008-lidar0-1236016333197799906-1236016358099063636
+host-a101-lidar0-1240710366399037786-1240710391298976894
+host-a102-lidar0-1242755350298764586-1242755375198787666
+host-a101-lidar0-1240877587199107226-1240877612099413030
+host-a101-lidar0-1242583745399163026-1242583770298821706
+host-a011-lidar0-1232817034199342856-1232817059098800346
+host-a004-lidar0-1232905117299287546-1232905142198246226
\ No newline at end of file

docs/getting_started.md

@@ -31,6 +31,21 @@ mmdetection3d
 │   │   │   ├── image_2
 │   │   │   ├── label_2
 │   │   │   ├── velodyne
+│   ├── lyft
+│   │   ├── v1.01-train
+│   │   │   ├── v1.01-train (train_data)
+│   │   │   ├── lidar (train_lidar)
+│   │   │   ├── images (train_images)
+│   │   │   ├── maps (train_maps)
+│   │   ├── v1.01-test
+│   │   │   ├── v1.01-test (test_data)
+│   │   │   ├── lidar (test_lidar)
+│   │   │   ├── images (test_images)
+│   │   │   ├── maps (test_maps)
+│   │   ├── train.txt
+│   │   ├── val.txt
+│   │   ├── test.txt
+│   │   ├── sample_submission.csv
 │   ├── scannet
 │   │   ├── meta_data
 │   │   ├── scans
@@ -57,6 +72,12 @@ Download KITTI 3D detection data [HERE](http://www.cvlibs.net/datasets/kitti/eva
 python tools/create_data.py kitti --root-path ./data/kitti --out-dir ./data/kitti --extra-tag kitti
 ```
 
+Download Lyft 3D detection data [HERE](https://www.kaggle.com/c/3d-object-detection-for-autonomous-vehicles/data). Prepare Lyft data by running
+```bash
+python tools/create_data.py lyft --root-path ./data/lyft --out-dir ./data/lyft --extra-tag lyft --version v1.01
+```
+Note that we follow the original folder names for clear organization. Please rename the raw folders as shown above.
+
 To prepare scannet data, please see [scannet](../data/scannet/README.md).
 
 To prepare sunrgbd data, please see [sunrgbd](../data/sunrgbd/README.md).

mmdet3d/core/evaluation/__init__.py

 from .indoor_eval import indoor_eval
 from .kitti_utils import kitti_eval, kitti_eval_coco_style
+from .lyft_eval import lyft_eval
 
-__all__ = ['kitti_eval_coco_style', 'kitti_eval', 'indoor_eval']
+__all__ = ['kitti_eval_coco_style', 'kitti_eval', 'indoor_eval', 'lyft_eval']

mmdet3d/core/evaluation/lyft_eval.py

+import os.path as osp
+
+import mmcv
+import numpy as np
+from lyft_dataset_sdk.eval.detection.mAP_evaluation import (Box3D, get_ap,
+                                                            get_class_names,
+                                                            get_ious,
+                                                            group_by_key,
+                                                            wrap_in_box)
+from mmcv.utils import print_log
+from terminaltables import AsciiTable
+
+
+def load_lyft_gts(lyft, data_root, eval_split: str, logger=None) -> list:
+    """Loads ground truth boxes from database.
+
+    Args:
+        lyft (:obj:``LyftDataset``): Lyft class in the sdk.
+        data_root (str): Root of data for reading splits.
+        eval_split (str): Name of the split for evaluation.
+        logger (logging.Logger | str | None): Logger used for printing
+        related information during evaluation. Default: None.
+
+    Returns:
+        list[dict]: List of annotation dictionaries.
+    """
+    split_scenes = mmcv.list_from_file(
+        osp.join(data_root, f'{eval_split}.txt'))
+
+    # Read out all sample_tokens in DB.
+    sample_tokens_all = [s['token'] for s in lyft.sample]
+    assert len(sample_tokens_all) > 0, 'Error: Database has no samples!'
+
+    if eval_split == 'test':
+        # Check that you aren't trying to cheat :)
+        assert len(lyft.sample_annotation) > 0, \
+            'Error: You are trying to evaluate on the test set \
+             but you do not have the annotations!'
+
+    sample_tokens = []
+    for sample_token in sample_tokens_all:
+        scene_token = lyft.get('sample', sample_token)['scene_token']
+        scene_record = lyft.get('scene', scene_token)
+        if scene_record['name'] in split_scenes:
+            sample_tokens.append(sample_token)
+
+    all_annotations = []
+
+    print_log('Loading ground truth annotations...', logger=logger)
+    # Load annotations and filter predictions and annotations.
+    for sample_token in mmcv.track_iter_progress(sample_tokens):
+        sample = lyft.get('sample', sample_token)
+        sample_annotation_tokens = sample['anns']
+        for sample_annotation_token in sample_annotation_tokens:
+            # Get label name in detection task and filter unused labels.
+            sample_annotation = \
+                lyft.get('sample_annotation', sample_annotation_token)
+            detection_name = sample_annotation['category_name']
+            if detection_name is None:
+                continue
+            annotation = {
+                'sample_token': sample_token,
+                'translation': sample_annotation['translation'],
+                'size': sample_annotation['size'],
+                'rotation': sample_annotation['rotation'],
+                'name': detection_name,
+            }
+            all_annotations.append(annotation)
+
+    return all_annotations
+
+
+def load_lyft_predictions(res_path):
+    """Load Lyft predictions from json file.
+
+    Args:
+        res_path (str): Path of result json file recording detections.
+
+    Returns:
+        list[dict]: List of prediction dictionaries.
+    """
+    predictions = mmcv.load(res_path)
+    predictions = predictions['results']
+    all_preds = []
+    for sample_token in predictions.keys():
+        all_preds.extend(predictions[sample_token])
+    return all_preds
+
+
+def lyft_eval(lyft, data_root, res_path, eval_set, output_dir, logger=None):
+    """Evaluation API for Lyft dataset.
+
+    Args:
+        lyft (:obj:``LyftDataset``): Lyft class in the sdk.
+        data_root (str): Root of data for reading splits.
+        res_path (str): Path of result json file recording detections.
+        eval_set (str): Name of the split for evaluation.
+        output_dir (str): Output directory for output json files.
+        logger (logging.Logger | str | None): Logger used for printing
+                related information during evaluation. Default: None.
+
+    Returns:
+        dict: The metric dictionary recording the evaluation results.
+    """
+    # evaluate by lyft metrics
+    gts = load_lyft_gts(lyft, data_root, eval_set, logger)
+    predictions = load_lyft_predictions(res_path)
+
+    class_names = get_class_names(gts)
+    print_log('Evaluating...', logger=logger)
+    class_table = AsciiTable([class_names], title='Class Names')
+    print_log(class_table.table, logger=logger)
+
+    iou_thresholds = [0.5, 0.55, 0.6, 0.65, 0.7, 0.75, 0.8, 0.85, 0.9, 0.95]
+    metrics = {}
+    average_precisions = \
+        get_classwise_aps(gts, predictions, class_names, iou_thresholds)
+    APs_data = [['IOU', 0.5, 0.55, 0.6, 0.65, 0.7, 0.75, 0.8, 0.85, 0.9, 0.95]]
+
+    mAPs = np.mean(average_precisions, axis=0)
+    mAPs_cate = np.mean(average_precisions, axis=1)
+    final_mAP = np.mean(mAPs)
+
+    metrics['average_precisions'] = average_precisions.tolist()
+    metrics['mAPs'] = mAPs.tolist()
+    metrics['Final mAP'] = float(final_mAP)
+    metrics['class_names'] = class_names
+    metrics['mAPs_cate'] = mAPs_cate.tolist()
+
+    APs_data = [['class', 'mAP@0.5:0.95']]
+    for i in range(len(class_names)):
+        row = [class_names[i], round(mAPs_cate[i], 3)]
+        APs_data.append(row)
+    APs_data.append(['Overall', round(final_mAP, 3)])
+    APs_table = AsciiTable(APs_data, title='mAPs@0.5:0.95')
+    print_log(APs_table.table, logger=logger)
+
+    res_path = osp.join(output_dir, 'lyft_metrics.json')
+    mmcv.dump(metrics, res_path)
+    return metrics
+
+
+def get_classwise_aps(gt: list, predictions: list, class_names: list,
+                      iou_thresholds: list) -> np.array:
+    """Returns an array with an average precision per class.
+
+    Note: Ground truth and predictions should have the following format.
+    .. code-block::
+
+    gt = [{
+        'sample_token': '0f0e3ce89d2324d8b45aa55a7b4f8207
+                         fbb039a550991a5149214f98cec136ac',
+        'translation': [974.2811881299899, 1714.6815014457964,
+                        -23.689857123368846],
+        'size': [1.796, 4.488, 1.664],
+        'rotation': [0.14882026466054782, 0, 0, 0.9888642620837121],
+        'name': 'car'
+    }]
+
+    predictions = [{
+        'sample_token': '0f0e3ce89d2324d8b45aa55a7b4f8207
+                         fbb039a550991a5149214f98cec136ac',
+        'translation': [971.8343488872263, 1713.6816097857359,
+                        -25.82534357061308],
+        'size': [2.519726579986132, 7.810161372666739, 3.483438286096803],
+        'rotation': [0.10913582721095375, 0.04099572636992043,
+                     0.01927712319721745, 1.029328402625659],
+        'name': 'car',
+        'score': 0.3077029437237213
+    }]
+
+    Args:
+        gt (list[dict]): list of dictionaries in the format described below.
+        predictions (list[dict]): list of dictionaries in the format
+            described below.
+        class_names (list[str]): list of the class names.
+        iou_thresholds (list[float]): IOU thresholds used to calculate
+            TP / FN
+
+    Returns:
+        np.ndarray: an array with an average precision per class.
+    """
+    assert all([0 <= iou_th <= 1 for iou_th in iou_thresholds])
+
+    gt_by_class_name = group_by_key(gt, 'name')
+    pred_by_class_name = group_by_key(predictions, 'name')
+
+    average_precisions = np.zeros((len(class_names), len(iou_thresholds)))
+
+    for class_id, class_name in enumerate(class_names):
+        if class_name in pred_by_class_name:
+            recalls, precisions, average_precision = get_single_class_aps(
+                gt_by_class_name[class_name], pred_by_class_name[class_name],
+                iou_thresholds)
+            average_precisions[class_id, :] = average_precision
+
+    return average_precisions
+
+
+def get_single_class_aps(gt, predictions, iou_thresholds):
+    """Compute recall and precision for all iou thresholds.
+    Adapted from LyftDatasetDevkit.
+
+    Args:
+        gt (list[dict]): list of dictionaries in the format described above.
+        predictions (list[dict]): list of dictionaries in the format
+            described below.
+        iou_thresholds (list[float]): IOU thresholds used to calculate
+            TP / FN
+
+    Returns:
+        tuple[np.ndarray]: returns (recalls, precisions, average precisions)
+            for each class.
+    """
+    num_gts = len(gt)
+    image_gts = group_by_key(gt, 'sample_token')
+    image_gts = wrap_in_box(image_gts)
+
+    sample_gt_checked = {
+        sample_token: np.zeros((len(boxes), len(iou_thresholds)))
+        for sample_token, boxes in image_gts.items()
+    }
+
+    predictions = sorted(predictions, key=lambda x: x['score'], reverse=True)
+
+    # go down dets and mark TPs and FPs
+    num_predictions = len(predictions)
+    tps = np.zeros((num_predictions, len(iou_thresholds)))
+    fps = np.zeros((num_predictions, len(iou_thresholds)))
+
+    for prediction_index, prediction in enumerate(predictions):
+        predicted_box = Box3D(**prediction)
+
+        sample_token = prediction['sample_token']
+
+        max_overlap = -np.inf
+        jmax = -1
+
+        if sample_token in image_gts:
+            gt_boxes = image_gts[sample_token]
+            # gt_boxes per sample
+            gt_checked = sample_gt_checked[sample_token]
+            # gt flags per sample
+        else:
+            gt_boxes = []
+            gt_checked = None
+
+        if len(gt_boxes) > 0:
+            overlaps = get_ious(gt_boxes, predicted_box)
+
+            max_overlap = np.max(overlaps)
+
+            jmax = np.argmax(overlaps)
+
+        for i, iou_threshold in enumerate(iou_thresholds):
+            if max_overlap > iou_threshold:
+                if gt_checked[jmax, i] == 0:
+                    tps[prediction_index, i] = 1.0
+                    gt_checked[jmax, i] = 1
+                else:
+                    fps[prediction_index, i] = 1.0
+            else:
+                fps[prediction_index, i] = 1.0
+
+    # compute precision recall
+    fps = np.cumsum(fps, axis=0)
+    tps = np.cumsum(tps, axis=0)
+
+    recalls = tps / float(num_gts)
+    # avoid divide by zero in case the first detection
+    # matches a difficult ground truth
+    precisions = tps / np.maximum(tps + fps, np.finfo(np.float64).eps)
+
+    aps = []
+    for i in range(len(iou_thresholds)):
+        recall = recalls[:, i]
+        precision = precisions[:, i]
+        assert np.all(0 <= recall) & np.all(recall <= 1)
+        assert np.all(0 <= precision) & np.all(precision <= 1)
+        ap = get_ap(recall, precision)
+        aps.append(ap)
+
+    aps = np.array(aps)
+
+    return recalls, precisions, aps

mmdet3d/datasets/__init__.py

@@ -2,6 +2,7 @@ from mmdet.datasets.builder import DATASETS, build_dataloader, build_dataset
 from .custom_3d import Custom3DDataset
 from .kitti2d_dataset import Kitti2DDataset
 from .kitti_dataset import KittiDataset
+from .lyft_dataset import LyftDataset
 from .nuscenes_dataset import NuScenesDataset
 from .pipelines import (GlobalRotScaleTrans, IndoorPointSample,
                         LoadAnnotations3D, LoadPointsFromFile,
@@ -14,9 +15,9 @@ from .sunrgbd_dataset import SUNRGBDDataset
 __all__ = [
     'KittiDataset', 'GroupSampler', 'DistributedGroupSampler',
     'build_dataloader', 'RepeatFactorDataset', 'DATASETS', 'build_dataset',
-    'CocoDataset', 'Kitti2DDataset', 'NuScenesDataset', 'ObjectSample',
-    'RandomFlip3D', 'ObjectNoise', 'GlobalRotScaleTrans', 'PointShuffle',
-    'ObjectRangeFilter', 'PointsRangeFilter', 'Collect3D',
+    'CocoDataset', 'Kitti2DDataset', 'NuScenesDataset', 'LyftDataset',
+    'ObjectSample', 'RandomFlip3D', 'ObjectNoise', 'GlobalRotScaleTrans',
+    'PointShuffle', 'ObjectRangeFilter', 'PointsRangeFilter', 'Collect3D',
     'LoadPointsFromFile', 'NormalizePointsColor', 'IndoorPointSample',
     'LoadAnnotations3D', 'SUNRGBDDataset', 'ScanNetDataset', 'Custom3DDataset'
 ]

mmdet3d/datasets/custom_3d.py

@@ -24,12 +24,13 @@ class Custom3DDataset(Dataset):
             Defaults to None.
         classes (tuple[str], optional): Classes used in the dataset.
             Defaults to None.
-        modality ([dict], optional): Modality to specify the sensor data used
+        modality (dict, optional): Modality to specify the sensor data used
             as input. Defaults to None.
         box_type_3d (str, optional): Type of 3D box of this dataset.
             Based on the `box_type_3d`, the dataset will encapsulate the box
             to its original format then converted them to `box_type_3d`.
             Defaults to 'LiDAR'. Available options includes
+
             - 'LiDAR': box in LiDAR coordinates
             - 'Depth': box in depth coordinates, usually for indoor dataset
             - 'Camera': box in camera coordinates

mmdet3d/datasets/kitti_dataset.py

@@ -16,7 +16,40 @@ from .custom_3d import Custom3DDataset
 
 @DATASETS.register_module()
 class KittiDataset(Custom3DDataset):
+    """KITTI Dataset
 
+    This class serves as the API for experiments on the KITTI Dataset.
+
+    Please refer to `<http://www.cvlibs.net/datasets/kitti/eval_object.php?
+    obj_benchmark=3d>`_for data downloading. It is recommended to symlink
+    the dataset root to $MMDETECTION3D/data and organize them as the doc
+    shows.
+
+    Args:
+        data_root (str): Path of dataset root.
+        ann_file (str): Path of annotation file.
+        split (str): Split of input data.
+        pts_prefix (str, optional): Prefix of points files.
+            Defaults to 'velodyne'.
+        pipeline (list[dict], optional): Pipeline used for data processing.
+            Defaults to None.
+        classes (tuple[str], optional): Classes used in the dataset.
+            Defaults to None.
+        modality (dict, optional): Modality to specify the sensor data used
+            as input. Defaults to None.
+        box_type_3d (str, optional): Type of 3D box of this dataset.
+            Based on the `box_type_3d`, the dataset will encapsulate the box
+            to its original format then converted them to `box_type_3d`.
+            Defaults to 'LiDAR' in this dataset. Available options includes
+
+            - 'LiDAR': box in LiDAR coordinates
+            - 'Depth': box in depth coordinates, usually for indoor dataset
+            - 'Camera': box in camera coordinates
+        filter_empty_gt (bool, optional): Whether to filter empty GT.
+            Defaults to True.
+        test_mode (bool, optional): Whether the dataset is in test mode.
+            Defaults to False.
+    """
     CLASSES = ('car', 'pedestrian', 'cyclist')
 
     def __init__(self,
@@ -189,7 +222,7 @@ class KittiDataset(Custom3DDataset):
         """Evaluation in KITTI protocol.
 
         Args:
-            results (list): Testing results of the dataset.
+            results (list[dict]): Testing results of the dataset.
             metric (str | list[str]): Metrics to be evaluated.
             logger (logging.Logger | str | None): Logger used for printing
                 related information during evaluation. Default: None.
@@ -352,7 +385,8 @@ class KittiDataset(Custom3DDataset):
         """Convert results to kitti format for evaluation and test submission.
 
         Args:
-            net_outputs (List[array]): list of array storing the bbox and score
+            net_outputs (List[np.ndarray]): list of array storing the
+                bbox and score
             class_nanes (List[String]): A list of class names
             pklfile_prefix (str | None): The prefix of pkl file.
             submission_prefix (str | None): The prefix of submission file.

mmdet3d/datasets/lyft_dataset.py

+import os.path as osp
+import tempfile
+
+import mmcv
+import numpy as np
+import pandas as pd
+from lyft_dataset_sdk.lyftdataset import LyftDataset as Lyft
+from lyft_dataset_sdk.utils.data_classes import Box as LyftBox
+from pyquaternion import Quaternion
+
+from mmdet3d.core.evaluation.lyft_eval import lyft_eval
+from mmdet.datasets import DATASETS
+from ..core.bbox import LiDARInstance3DBoxes
+from .custom_3d import Custom3DDataset
+
+
+@DATASETS.register_module()
+class LyftDataset(Custom3DDataset):
+    """Lyft Dataset
+
+    This class serves as the API for experiments on the Lyft Dataset.
+
+    Please refer to
+    `<https://www.kaggle.com/c/3d-object-detection-for-autonomous-vehicles
+    /data>`_for data downloading. It is recommended to symlink the dataset
+    root to $MMDETECTION3D/data and organize them as the doc shows.
+
+    Args:
+        ann_file (str): Path of annotation file.
+        pipeline (list[dict], optional): Pipeline used for data processing.
+            Defaults to None.
+        data_root (str): Path of dataset root.
+        classes (tuple[str], optional): Classes used in the dataset.
+            Defaults to None.
+        load_interval (int, optional): Interval of loading the dataset. It is
+            used to uniformly sample the dataset. Defaults to 1.
+        modality (dict, optional): Modality to specify the sensor data used
+            as input. Defaults to None.
+        box_type_3d (str, optional): Type of 3D box of this dataset.
+            Based on the `box_type_3d`, the dataset will encapsulate the box
+            to its original format then converted them to `box_type_3d`.
+            Defaults to 'LiDAR' in this dataset. Available options includes
+
+            - 'LiDAR': box in LiDAR coordinates
+            - 'Depth': box in depth coordinates, usually for indoor dataset
+            - 'Camera': box in camera coordinates
+        filter_empty_gt (bool, optional): Whether to filter empty GT.
+            Defaults to True.
+        test_mode (bool, optional): Whether the dataset is in test mode.
+            Defaults to False.
+    """
+    NameMapping = {
+        'bicycle': 'bicycle',
+        'bus': 'bus',
+        'car': 'car',
+        'emergency_vehicle': 'emergency_vehicle',
+        'motorcycle': 'motorcycle',
+        'other_vehicle': 'other_vehicle',
+        'pedestrian': 'pedestrian',
+        'truck': 'truck',
+        'animal': 'animal'
+    }
+    DefaultAttribute = {
+        'car': 'is_stationary',
+        'truck': 'is_stationary',
+        'bus': 'is_stationary',
+        'emergency_vehicle': 'is_stationary',
+        'other_vehicle': 'is_stationary',
+        'motorcycle': 'is_stationary',
+        'bicycle': 'is_stationary',
+        'pedestrian': 'is_stationary',
+        'animal': 'is_stationary'
+    }
+    CLASSES = ('car', 'truck', 'bus', 'emergency_vehicle', 'other_vehicle',
+               'motorcycle', 'bicycle', 'pedestrian', 'animal')
+
+    def __init__(self,
+                 ann_file,
+                 pipeline=None,
+                 data_root=None,
+                 classes=None,
+                 load_interval=1,
+                 modality=None,
+                 box_type_3d='LiDAR',
+                 filter_empty_gt=True,
+                 test_mode=False):
+        self.load_interval = load_interval
+        super().__init__(
+            data_root=data_root,
+            ann_file=ann_file,
+            pipeline=pipeline,
+            classes=classes,
+            modality=modality,
+            box_type_3d=box_type_3d,
+            filter_empty_gt=filter_empty_gt,
+            test_mode=test_mode)
+
+        if self.modality is None:
+            self.modality = dict(
+                use_camera=False,
+                use_lidar=True,
+                use_radar=False,
+                use_map=False,
+                use_external=False,
+            )
+
+    def load_annotations(self, ann_file):
+        """Load annotations from ann_file.
+
+        Args:
+            ann_file (str): Path of the annotation file.
+
+        Returns:
+            list[dict]: List of annotations sorted by timestamps.
+        """
+        data = mmcv.load(ann_file)
+        data_infos = list(sorted(data['infos'], key=lambda e: e['timestamp']))
+        data_infos = data_infos[::self.load_interval]
+        self.metadata = data['metadata']
+        self.version = self.metadata['version']
+        return data_infos
+
+    def get_data_info(self, index):
+        """Get data info according to the given index.
+
+        Args:
+            index (int): Index of the sample data to get.
+
+        Returns:
+            dict: Standard input_dict consists of the
+                data information.
+
+                - sample_idx (str): sample index
+                - pts_filename (str): filename of point clouds
+                - sweeps (list[dict]): infos of sweeps
+                - timestamp (float): sample timestamp
+                - img_filename (str, optional): image filename
+                - lidar2img (list[np.ndarray], optional): transformations from
+                    lidar to different cameras
+                - ann_info (dict): annotation info
+        """
+        info = self.data_infos[index]
+
+        # standard protocal modified from SECOND.Pytorch
+        input_dict = dict(
+            sample_idx=info['token'],
+            pts_filename=info['lidar_path'],
+            sweeps=info['sweeps'],
+            timestamp=info['timestamp'] / 1e6,
+        )
+
+        if self.modality['use_camera']:
+            image_paths = []
+            lidar2img_rts = []
+            for cam_type, cam_info in info['cams'].items():
+                image_paths.append(cam_info['data_path'])
+                # obtain lidar to image transformation matrix
+                lidar2cam_r = np.linalg.inv(cam_info['sensor2lidar_rotation'])
+                lidar2cam_t = cam_info[
+                    'sensor2lidar_translation'] @ lidar2cam_r.T
+                lidar2cam_rt = np.eye(4)
+                lidar2cam_rt[:3, :3] = lidar2cam_r.T
+                lidar2cam_rt[3, :3] = -lidar2cam_t
+                intrinsic = cam_info['cam_intrinsic']
+                viewpad = np.eye(4)
+                viewpad[:intrinsic.shape[0], :intrinsic.shape[1]] = intrinsic
+                lidar2img_rt = (viewpad @ lidar2cam_rt.T)
+                lidar2img_rts.append(lidar2img_rt)
+
+            input_dict.update(
+                dict(
+                    img_filename=image_paths,
+                    lidar2img=lidar2img_rts,
+                ))
+
+        if not self.test_mode:
+            annos = self.get_ann_info(index)
+            input_dict['ann_info'] = annos
+
+        return input_dict
+
+    def get_ann_info(self, index):
+        """Get annotation info according to the given index.
+
+        Args:
+            index (int): Index of the annotation data to get.
+
+        Returns:
+            dict: Standard annotation dictionary
+                consists of the data information.
+
+                - gt_bboxes_3d (:obj:``LiDARInstance3DBoxes``):
+                    3D ground truth bboxes
+                - gt_labels_3d (np.ndarray): labels of ground truths
+                - gt_names (list[str]): class names of ground truths
+        """
+        info = self.data_infos[index]
+        gt_bboxes_3d = info['gt_boxes']
+        gt_names_3d = info['gt_names']
+        gt_labels_3d = []
+        for cat in gt_names_3d:
+            if cat in self.CLASSES:
+                gt_labels_3d.append(self.CLASSES.index(cat))
+            else:
+                gt_labels_3d.append(-1)
+        gt_labels_3d = np.array(gt_labels_3d)
+
+        if 'gt_shape' in info:
+            gt_shape = info['gt_shape']
+            gt_bboxes_3d = np.concatenate([gt_bboxes_3d, gt_shape], axis=-1)
+
+        # the lyft box center is [0.5, 0.5, 0.5], we change it to be
+        # the same as KITTI (0.5, 0.5, 0)
+        gt_bboxes_3d = LiDARInstance3DBoxes(
+            gt_bboxes_3d,
+            box_dim=gt_bboxes_3d.shape[-1],
+            origin=(0.5, 0.5, 0.5)).convert_to(self.box_mode_3d)
+
+        anns_results = dict(
+            gt_bboxes_3d=gt_bboxes_3d,
+            gt_labels_3d=gt_labels_3d,
+        )
+        return anns_results
+
+    def _format_bbox(self, results, jsonfile_prefix=None):
+        """Convert the results to the standard format.
+
+        Args:
+            results (list[dict]): Testing results of the dataset.
+            jsonfile_prefix (str): The prefix of the output jsonfile.
+                You can specify the output directory/filename by
+                modifying the jsonfile_prefix. Default: None.
+
+        Returns:
+            str: Path of the output json file.
+        """
+        lyft_annos = {}
+        mapped_class_names = self.CLASSES
+
+        print('Start to convert detection format...')
+        for sample_id, det in enumerate(mmcv.track_iter_progress(results)):
+            annos = []
+            boxes = output_to_lyft_box(det)
+            sample_token = self.data_infos[sample_id]['token']
+            boxes = lidar_lyft_box_to_global(self.data_infos[sample_id], boxes)
+            for i, box in enumerate(boxes):
+                name = mapped_class_names[box.label]
+                lyft_anno = dict(
+                    sample_token=sample_token,
+                    translation=box.center.tolist(),
+                    size=box.wlh.tolist(),
+                    rotation=box.orientation.elements.tolist(),
+                    name=name,
+                    score=box.score)
+                annos.append(lyft_anno)
+            lyft_annos[sample_token] = annos
+        lyft_submissions = {
+            'meta': self.modality,
+            'results': lyft_annos,
+        }
+
+        mmcv.mkdir_or_exist(jsonfile_prefix)
+        res_path = osp.join(jsonfile_prefix, 'results_lyft.json')
+        print('Results writes to', res_path)
+        mmcv.dump(lyft_submissions, res_path)
+        return res_path
+
+    def _evaluate_single(self,
+                         result_path,
+                         logger=None,
+                         metric='bbox',
+                         result_name='pts_bbox'):
+        """Evaluation for a single model in Lyft protocol.
+
+        Args:
+            result_path (str): Path of the result file.
+            logger (logging.Logger | str | None): Logger used for printing
+                related information during evaluation. Default: None.
+            metric (str): Metric name used for evaluation. Default: 'bbox'.
+            result_name (str): Result name in the metric prefix.
+                Default: 'pts_bbox'.
+
+        Returns:
+            dict: Dictionary of evaluation details.
+        """
+
+        output_dir = osp.join(*osp.split(result_path)[:-1])
+        lyft = Lyft(
+            data_path=osp.join(self.data_root, self.version),
+            json_path=osp.join(self.data_root, self.version, self.version),
+            verbose=True)
+        eval_set_map = {
+            'v1.01-train': 'val',
+        }
+        metrics = lyft_eval(lyft, self.data_root, result_path,
+                            eval_set_map[self.version], output_dir, logger)
+
+        # record metrics
+        detail = dict()
+        metric_prefix = f'{result_name}_Lyft'
+
+        for i, name in enumerate(metrics['class_names']):
+            AP = float("f{round(metrics['mAPs_cate'][i], 3)}")
+            detail[f'{metric_prefix}/{name}_AP'] = AP
+
+        detail[f'{metric_prefix}/mAP'] = metrics['Final mAP']
+        return detail
+
+    def format_results(self, results, jsonfile_prefix=None, csv_savepath=None):
+        """Format the results to json (standard format for COCO evaluation).
+
+        Args:
+            results (list[dict]): Testing results of the dataset.
+            jsonfile_prefix (str | None): The prefix of json files. It includes
+                the file path and the prefix of filename, e.g., "a/b/prefix".
+                If not specified, a temp file will be created. Default: None.
+            csv_savepath (str | None): The path for saving csv files.
+                It includes the file path and the csv filename,
+                e.g., "a/b/filename.csv". If not specified,
+                the result will not be converted to csv file.
+
+        Returns:
+            tuple (dict, str): result_files is a dict containing the json
+                filepaths, tmp_dir is the temporal directory created for
+                saving json files when jsonfile_prefix is not specified.
+        """
+        assert isinstance(results, list), 'results must be a list'
+        assert len(results) == len(self), (
+            'The length of results is not equal to the dataset len: {} != {}'.
+            format(len(results), len(self)))
+
+        if jsonfile_prefix is None:
+            tmp_dir = tempfile.TemporaryDirectory()
+            jsonfile_prefix = osp.join(tmp_dir.name, 'results')
+        else:
+            tmp_dir = None
+
+        if not isinstance(results[0], dict):
+            result_files = self._format_bbox(results, jsonfile_prefix)
+        else:
+            result_files = dict()
+            for name in results[0]:
+                print(f'\nFormating bboxes of {name}')
+                results_ = [out[name] for out in results]
+                tmp_file_ = osp.join(jsonfile_prefix, name)
+                result_files.update(
+                    {name: self._format_bbox(results_, tmp_file_)})
+        if csv_savepath is not None:
+            self.json2csv(result_files['pts_bbox'], csv_savepath)
+        return result_files, tmp_dir
+
+    def evaluate(self,
+                 results,
+                 metric='bbox',
+                 logger=None,
+                 jsonfile_prefix=None,
+                 csv_savepath=None,
+                 result_names=['pts_bbox']):
+        """Evaluation in Lyft protocol.
+
+        Args:
+            results (list[dict]): Testing results of the dataset.
+            metric (str | list[str]): Metrics to be evaluated.
+            logger (logging.Logger | str | None): Logger used for printing
+                related information during evaluation. Default: None.
+            jsonfile_prefix (str | None): The prefix of json files. It includes
+                the file path and the prefix of filename, e.g., "a/b/prefix".
+                If not specified, a temp file will be created. Default: None.
+            csv_savepath (str | None): The path for saving csv files.
+                It includes the file path and the csv filename,
+                e.g., "a/b/filename.csv". If not specified,
+                the result will not be converted to csv file.
+
+        Returns:
+            dict[str: float]
+        """
+        result_files, tmp_dir = self.format_results(results, jsonfile_prefix,
+                                                    csv_savepath)
+
+        if isinstance(result_files, dict):
+            results_dict = dict()
+            for name in result_names:
+                print(f'Evaluating bboxes of {name}')
+                ret_dict = self._evaluate_single(result_files[name])
+            results_dict.update(ret_dict)
+        elif isinstance(result_files, str):
+            results_dict = self._evaluate_single(result_files)
+
+        if tmp_dir is not None:
+            tmp_dir.cleanup()
+        return results_dict
+
+    @staticmethod
+    def json2csv(json_path, csv_savepath):
+        """Convert the json file to csv format for submission.
+
+        Args:
+            json_path (str): Path of the result json file.
+            csv_savepath (str): Path to save the csv file.
+        """
+        with open(json_path, 'r') as f:
+            results = mmcv.load(f)['results']
+        csv_nopred = 'data/lyft/sample_submission.csv'
+        data = pd.read_csv(csv_nopred)
+        Id_list = list(data['Id'])
+        pred_list = list(data['PredictionString'])
+        cnt = 0
+        print('Converting the json to csv...')
+        for token in results.keys():
+            cnt += 1
+            predictions = results[token]
+            prediction_str = ''
+            for i in range(len(predictions)):
+                prediction_str += \
+                    str(predictions[i]['score']) + ' ' + \
+                    str(predictions[i]['translation'][0]) + ' ' + \
+                    str(predictions[i]['translation'][1]) + ' ' + \
+                    str(predictions[i]['translation'][2]) + ' ' + \
+                    str(predictions[i]['size'][0]) + ' ' + \
+                    str(predictions[i]['size'][1]) + ' ' + \
+                    str(predictions[i]['size'][2]) + ' ' + \
+                    str(Quaternion(list(predictions[i]['rotation']))
+                        .yaw_pitch_roll[0]) + ' ' + \
+                    predictions[i]['name'] + ' '
+            prediction_str = prediction_str[:-1]
+            idx = Id_list.index(token)
+            pred_list[idx] = prediction_str
+        df = pd.DataFrame({'Id': Id_list, 'PredictionString': pred_list})
+        df.to_csv(csv_savepath, index=False)
+
+
+def output_to_lyft_box(detection):
+    """Convert the output to the box class in the Lyft.
+
+    Args:
+        detection (dict): Detection results.
+
+    Returns:
+        list[:obj:``LyftBox``]: List of standard LyftBoxes.
+    """
+    box3d = detection['boxes_3d']
+    scores = detection['scores_3d'].numpy()
+    labels = detection['labels_3d'].numpy()
+
+    box_gravity_center = box3d.gravity_center.numpy()
+    box_dims = box3d.dims.numpy()
+    box_yaw = box3d.yaw.numpy()
+    # TODO: check whether this is necessary
+    # with dir_offset & dir_limit in the head
+    box_yaw = -box_yaw - np.pi / 2
+
+    box_list = []
+    for i in range(len(box3d)):
+        quat = Quaternion(axis=[0, 0, 1], radians=box_yaw[i])
+        box = LyftBox(
+            box_gravity_center[i],
+            box_dims[i],
+            quat,
+            label=labels[i],
+            score=scores[i])
+        box_list.append(box)
+    return box_list
+
+
+def lidar_lyft_box_to_global(info, boxes):
+    """Convert the box from ego to global coordinate.
+
+    Args:
+        info (dict): Info for a specific sample data, including the
+            calibration information.
+        boxes (list[:obj:``LyftBox``]): List of predicted LyftBoxes.
+
+    Returns:
+        list: List of standard LyftBoxes in the global
+            coordinate.
+    """
+    box_list = []
+    for box in boxes:
+        # Move box to ego vehicle coord system
+        box.rotate(Quaternion(info['lidar2ego_rotation']))
+        box.translate(np.array(info['lidar2ego_translation']))
+        # Move box to global coord system
+        box.rotate(Quaternion(info['ego2global_rotation']))
+        box.translate(np.array(info['ego2global_translation']))
+        box_list.append(box)
+    return box_list

mmdet3d/datasets/nuscenes_dataset.py

@@ -14,6 +14,42 @@ from .custom_3d import Custom3DDataset
 
 @DATASETS.register_module()
 class NuScenesDataset(Custom3DDataset):
+    """NuScenes Dataset
+
+    This class serves as the API for experiments on the NuScenes Dataset.
+
+    Please refer to `<https://www.nuscenes.org/download>`_for data
+    downloading. It is recommended to symlink the dataset root to
+    $MMDETECTION3D/data and organize them as the doc shows.
+
+    Args:
+        ann_file (str): Path of annotation file.
+        pipeline (list[dict], optional): Pipeline used for data processing.
+            Defaults to None.
+        data_root (str): Path of dataset root.
+        classes (tuple[str], optional): Classes used in the dataset.
+            Defaults to None.
+        load_interval (int, optional): Interval of loading the dataset. It is
+            used to uniformly sample the dataset. Defaults to 1.
+        with_velocity (bool, optional): Whether include velocity prediction
+            into the experiments. Defaults to True.
+        modality (dict, optional): Modality to specify the sensor data used
+            as input. Defaults to None.
+        box_type_3d (str, optional): Type of 3D box of this dataset.
+            Based on the `box_type_3d`, the dataset will encapsulate the box
+            to its original format then converted them to `box_type_3d`.
+            Defaults to 'LiDAR' in this dataset. Available options includes
+
+            - 'LiDAR': box in LiDAR coordinates
+            - 'Depth': box in depth coordinates, usually for indoor dataset
+            - 'Camera': box in camera coordinates
+        filter_empty_gt (bool, optional): Whether to filter empty GT.
+            Defaults to True.
+        test_mode (bool, optional): Whether the dataset is in test mode.
+            Defaults to False.
+        eval_version (bool, optional): Configuration version of evaluation.
+            Defaults to  'detection_cvpr_2019'.
+    """
     NameMapping = {
         'movable_object.barrier': 'barrier',
         'vehicle.bicycle': 'bicycle',
@@ -172,7 +208,7 @@ class NuScenesDataset(Custom3DDataset):
             gt_velocity[nan_mask] = [0.0, 0.0]
             gt_bboxes_3d = np.concatenate([gt_bboxes_3d, gt_velocity], axis=-1)
 
-        # the nuscenes box center is [0.5, 0.5, 0.5], we keep it
+        # the nuscenes box center is [0.5, 0.5, 0.5], we change it to be
         # the same as KITTI (0.5, 0.5, 0)
         gt_bboxes_3d = LiDARInstance3DBoxes(
             gt_bboxes_3d,
@@ -270,7 +306,7 @@ class NuScenesDataset(Custom3DDataset):
         # record metrics
         metrics = mmcv.load(osp.join(output_dir, 'metrics_summary.json'))
         detail = dict()
-        metric_prefix = '{}_NuScenes'.format(result_name)
+        metric_prefix = f'{result_name}_NuScenes'
         for name in self.CLASSES:
             for k, v in metrics['label_aps'][name].items():
                 val = float('{:.4f}'.format(v))
@@ -287,15 +323,15 @@ class NuScenesDataset(Custom3DDataset):
         """Format the results to json (standard format for COCO evaluation).
 
         Args:
-            results (list): Testing results of the dataset.
+            results (list[dict]): Testing results of the dataset.
             jsonfile_prefix (str | None): The prefix of json files. It includes
                 the file path and the prefix of filename, e.g., "a/b/prefix".
                 If not specified, a temp file will be created. Default: None.
 
         Returns:
-            tuple: (result_files, tmp_dir), result_files is a dict containing
-                the json filepaths, tmp_dir is the temporal directory created
-                for saving json files when jsonfile_prefix is not specified.
+            tuple (dict, str): result_files is a dict containing the json
+                filepaths, tmp_dir is the temporal directory created for
+                saving json files when jsonfile_prefix is not specified.
         """
         assert isinstance(results, list), 'results must be a list'
         assert len(results) == len(self), (
@@ -331,7 +367,7 @@ class NuScenesDataset(Custom3DDataset):
         """Evaluation in nuScenes protocol.
 
         Args:
-            results (list): Testing results of the dataset.
+            results (list[dict]): Testing results of the dataset.
             metric (str | list[str]): Metrics to be evaluated.
             logger (logging.Logger | str | None): Logger used for printing
                 related information during evaluation. Default: None.

mmdet3d/datasets/scannet_dataset.py

@@ -10,7 +10,36 @@ from .custom_3d import Custom3DDataset
 
 @DATASETS.register_module()
 class ScanNetDataset(Custom3DDataset):
+    """ScanNet Dataset
 
+    This class serves as the API for experiments on the ScanNet Dataset.
+
+    Please refer to `<https://github.com/ScanNet/ScanNet>`_for data
+    downloading. It is recommended to symlink the dataset root to
+    $MMDETECTION3D/data and organize them as the doc shows.
+
+    Args:
+        data_root (str): Path of dataset root.
+        ann_file (str): Path of annotation file.
+        pipeline (list[dict], optional): Pipeline used for data processing.
+            Defaults to None.
+        classes (tuple[str], optional): Classes used in the dataset.
+            Defaults to None.
+        modality (dict, optional): Modality to specify the sensor data used
+            as input. Defaults to None.
+        box_type_3d (str, optional): Type of 3D box of this dataset.
+            Based on the `box_type_3d`, the dataset will encapsulate the box
+            to its original format then converted them to `box_type_3d`.
+            Defaults to 'Depth' in this dataset. Available options includes
+
+            - 'LiDAR': box in LiDAR coordinates
+            - 'Depth': box in depth coordinates, usually for indoor dataset
+            - 'Camera': box in camera coordinates
+        filter_empty_gt (bool, optional): Whether to filter empty GT.
+            Defaults to True.
+        test_mode (bool, optional): Whether the dataset is in test mode.
+            Defaults to False.
+    """
     CLASSES = ('cabinet', 'bed', 'chair', 'sofa', 'table', 'door', 'window',
                'bookshelf', 'picture', 'counter', 'desk', 'curtain',
                'refrigerator', 'showercurtrain', 'toilet', 'sink', 'bathtub',

mmdet3d/datasets/sunrgbd_dataset.py

@@ -10,7 +10,36 @@ from .custom_3d import Custom3DDataset
 
 @DATASETS.register_module()
 class SUNRGBDDataset(Custom3DDataset):
+    """SUNRGBD Dataset
 
+    This class serves as the API for experiments on the SUNRGBD Dataset.
+
+    Please refer to `<http://rgbd.cs.princeton.edu/challenge.html>`_for
+    data downloading. It is recommended to symlink the dataset root to
+    $MMDETECTION3D/data and organize them as the doc shows.
+
+    Args:
+        data_root (str): Path of dataset root.
+        ann_file (str): Path of annotation file.
+        pipeline (list[dict], optional): Pipeline used for data processing.
+            Defaults to None.
+        classes (tuple[str], optional): Classes used in the dataset.
+            Defaults to None.
+        modality (dict, optional): Modality to specify the sensor data used
+            as input. Defaults to None.
+        box_type_3d (str, optional): Type of 3D box of this dataset.
+            Based on the `box_type_3d`, the dataset will encapsulate the box
+            to its original format then converted them to `box_type_3d`.
+            Defaults to 'Depth' in this dataset. Available options includes
+
+            - 'LiDAR': box in LiDAR coordinates
+            - 'Depth': box in depth coordinates, usually for indoor dataset
+            - 'Camera': box in camera coordinates
+        filter_empty_gt (bool, optional): Whether to filter empty GT.
+            Defaults to True.
+        test_mode (bool, optional): Whether the dataset is in test mode.
+            Defaults to False.
+    """
     CLASSES = ('bed', 'table', 'sofa', 'chair', 'toilet', 'desk', 'dresser',
                'night_stand', 'bookshelf', 'bathtub')
 

requirements/runtime.txt

@@ -5,6 +5,7 @@ mmcv>=0.6.0
 numba==0.48.0
 numpy
 nuscenes-devkit==1.0.5
+lyft_dataset_sdk
 # need older pillow until torchvision is fixed
 Pillow<=6.2.2
 plyfile

tools/create_data.py

@@ -3,18 +3,30 @@ import os.path as osp
 
 import tools.data_converter.indoor_converter as indoor
 import tools.data_converter.kitti_converter as kitti
+import tools.data_converter.lyft_converter as lyft_converter
 import tools.data_converter.nuscenes_converter as nuscenes_converter
 from tools.data_converter.create_gt_database import create_groundtruth_database
 
 
 def kitti_data_prep(root_path, info_prefix, version, out_dir):
+    """Prepare data related to Kitti dataset.
+
+    Related data consists of '.pkl' files recording basic infos,
+    2D annotations and groundtruth database.
+
+    Args:
+        root_path (str): Path of dataset root.
+        info_prefix (str): The prefix of info filenames.
+        version (str): Dataset version.
+        out_dir (str): Output directory of the groundtruth database info.
+    """
     kitti.create_kitti_info_file(root_path, info_prefix)
     kitti.create_reduced_point_cloud(root_path, info_prefix)
     create_groundtruth_database(
         'KittiDataset',
         root_path,
         info_prefix,
-        '{}/{}_infos_train.pkl'.format(out_dir, info_prefix),
+        f'{out_dir}/{info_prefix}_infos_train.pkl',
         relative_path=False,
         mask_anno_path='instances_train.json',
         with_mask=(version == 'mask'))
@@ -26,30 +38,97 @@ def nuscenes_data_prep(root_path,
                        dataset_name,
                        out_dir,
                        max_sweeps=10):
+    """Prepare data related to nuScenes dataset.
+
+    Related data consists of '.pkl' files recording basic infos,
+    2D annotations and groundtruth database.
+
+    Args:
+        root_path (str): Path of dataset root.
+        info_prefix (str): The prefix of info filenames.
+        version (str): Dataset version.
+        dataset_name (str): The dataset class name.
+        out_dir (str): Output directory of the groundtruth database info.
+        max_sweeps (int): Number of input consecutive frames. Default: 10
+    """
     nuscenes_converter.create_nuscenes_infos(
         root_path, info_prefix, version=version, max_sweeps=max_sweeps)
 
     if version == 'v1.0-test':
         return
 
-    info_train_path = osp.join(root_path,
-                               '{}_infos_train.pkl'.format(info_prefix))
-    info_val_path = osp.join(root_path, '{}_infos_val.pkl'.format(info_prefix))
+    info_train_path = osp.join(root_path, f'{info_prefix}_infos_train.pkl')
+    info_val_path = osp.join(root_path, f'{info_prefix}_infos_val.pkl')
     nuscenes_converter.export_2d_annotation(
         root_path, info_train_path, version=version)
     nuscenes_converter.export_2d_annotation(
         root_path, info_val_path, version=version)
-    create_groundtruth_database(
-        dataset_name, root_path, info_prefix,
-        '{}/{}_infos_train.pkl'.format(out_dir, info_prefix))
+    create_groundtruth_database(dataset_name, root_path, info_prefix,
+                                f'{out_dir}/{info_prefix}_infos_train.pkl')
+
+
+def lyft_data_prep(root_path,
+                   info_prefix,
+                   version,
+                   dataset_name,
+                   out_dir,
+                   max_sweeps=10):
+    """Prepare data related to Lyft dataset.
+
+    Related data consists of '.pkl' files recording basic infos,
+    and 2D annotations.
+    Although the ground truth database is not used in Lyft, it can also be
+    generated like nuScenes.
+
+    Args:
+        root_path (str): Path of dataset root.
+        info_prefix (str): The prefix of info filenames.
+        version (str): Dataset version.
+        dataset_name (str): The dataset class name.
+        out_dir (str): Output directory of the groundtruth database info.
+            Not used here if the groundtruth database is not generated.
+        max_sweeps (int): Number of input consecutive frames. Default: 10
+    """
+    lyft_converter.create_lyft_infos(
+        root_path, info_prefix, version=version, max_sweeps=max_sweeps)
+
+    if version == 'v1.01-test':
+        return
+
+    train_info_name = f'{info_prefix}_infos_train'
+    val_info_name = f'{info_prefix}_infos_val'
+
+    info_train_path = osp.join(root_path, f'{train_info_name}.pkl')
+    info_val_path = osp.join(root_path, f'{val_info_name}.pkl')
+
+    lyft_converter.export_2d_annotation(
+        root_path, info_train_path, version=version)
+    lyft_converter.export_2d_annotation(
+        root_path, info_val_path, version=version)
 
 
 def scannet_data_prep(root_path, info_prefix, out_dir, workers):
+    """Prepare the info file for scannet dataset.
+
+    Args:
+        root_path (str): Path of dataset root.
+        info_prefix (str): The prefix of info filenames.
+        out_dir (str): Output directory of the generated info file.
+        workers (int): Number of threads to be used.
+    """
     indoor.create_indoor_info_file(
         root_path, info_prefix, out_dir, workers=workers)
 
 
 def sunrgbd_data_prep(root_path, info_prefix, out_dir, workers):
+    """Prepare the info file for sunrgbd dataset.
+
+    Args:
+        root_path (str): Path of dataset root.
+        info_prefix (str): The prefix of info filenames.
+        out_dir (str): Output directory of the generated info file.
+        workers (int): Number of threads to be used.
+    """
     indoor.create_indoor_info_file(
         root_path, info_prefix, out_dir, workers=workers)
 
@@ -117,6 +196,23 @@ if __name__ == '__main__':
             dataset_name='NuScenesDataset',
             out_dir=args.out_dir,
             max_sweeps=args.max_sweeps)
+    elif args.dataset == 'lyft':
+        train_version = f'{args.version}-train'
+        lyft_data_prep(
+            root_path=args.root_path,
+            info_prefix=args.extra_tag,
+            version=train_version,
+            dataset_name='LyftDataset',
+            out_dir=args.out_dir,
+            max_sweeps=args.max_sweeps)
+        test_version = f'{args.version}-test'
+        lyft_data_prep(
+            root_path=args.root_path,
+            info_prefix=args.extra_tag,
+            version=test_version,
+            dataset_name='LyftDataset',
+            out_dir=args.out_dir,
+            max_sweeps=args.max_sweeps)
     elif args.dataset == 'scannet':
         scannet_data_prep(
             root_path=args.root_path,

tools/data_converter/kitti_converter.py

@@ -13,6 +13,9 @@ def convert_to_kitti_info_version2(info):
 
     Args:
         info (dict): Info of the input kitti data.
+            - image (dict): image info
+            - calib (dict): calibration info
+            - point_cloud (dict): point cloud info
     """
     if 'image' not in info or 'calib' not in info or 'point_cloud' not in info:
         info['image'] = {
@@ -194,6 +197,20 @@ def create_reduced_point_cloud(data_path,
                                test_info_path=None,
                                save_path=None,
                                with_back=False):
+    """Create reduced point cloud info file.
+
+    Args:
+        data_path (str): Path of original infos.
+        pkl_prefix (str): Prefix of info files.
+        train_info_path (str | None): Path of training set info.
+            Default: None.
+        val_info_path (str | None): Path of validation set info.
+            Default: None.
+        test_info_path (str | None): Path of test set info.
+            Default: None.
+        save_path (str | None): Path to save reduced info.
+        with_back (bool | None): Whether to create backup info.
+    """
     if train_info_path is None:
         train_info_path = Path(data_path) / f'{pkl_prefix}_infos_train.pkl'
     if val_info_path is None:

tools/data_converter/lyft_converter.py

+import os.path as osp
+
+import mmcv
+import numpy as np
+from lyft_dataset_sdk.lyftdataset import LyftDataset as Lyft
+from pyquaternion import Quaternion
+
+from mmdet3d.datasets import LyftDataset
+from .nuscenes_converter import (get_2d_boxes, get_available_scenes,
+                                 obtain_sensor2top)
+
+lyft_categories = ('car', 'truck', 'bus', 'emergency_vehicle', 'other_vehicle',
+                   'motorcycle', 'bicycle', 'pedestrian', 'animal')
+
+
+def create_lyft_infos(root_path,
+                      info_prefix,
+                      version='v1.01-train',
+                      max_sweeps=10):
+    """Create info file of lyft dataset.
+
+    Given the raw data, generate its related info file in pkl format.
+
+    Args:
+        root_path (str): Path of the data root.
+        info_prefix (str): Prefix of the info file to be generated.
+        version (str): Version of the data.
+            Default: 'v1.01-train'
+        max_sweeps (int): Max number of sweeps.
+            Default: 10
+    """
+    lyft = Lyft(
+        data_path=osp.join(root_path, version),
+        json_path=osp.join(root_path, version, version),
+        verbose=True)
+    available_vers = ['v1.01-train', 'v1.01-test']
+    assert version in available_vers
+    if version == 'v1.01-train':
+        train_scenes = mmcv.list_from_file('data/lyft/train.txt')
+        val_scenes = mmcv.list_from_file('data/lyft/val.txt')
+    elif version == 'v1.01-test':
+        train_scenes = mmcv.list_from_file('data/lyft/test.txt')
+        val_scenes = []
+    else:
+        raise ValueError('unknown')
+
+    # filter existing scenes.
+    available_scenes = get_available_scenes(lyft)
+    available_scene_names = [s['name'] for s in available_scenes]
+    train_scenes = list(
+        filter(lambda x: x in available_scene_names, train_scenes))
+    val_scenes = list(filter(lambda x: x in available_scene_names, val_scenes))
+    train_scenes = set([
+        available_scenes[available_scene_names.index(s)]['token']
+        for s in train_scenes
+    ])
+    val_scenes = set([
+        available_scenes[available_scene_names.index(s)]['token']
+        for s in val_scenes
+    ])
+
+    test = 'test' in version
+    if test:
+        print(f'test scene: {len(train_scenes)}')
+    else:
+        print(f'train scene: {len(train_scenes)}, \
+                val scene: {len(val_scenes)}')
+    train_lyft_infos, val_lyft_infos = _fill_trainval_infos(
+        lyft, train_scenes, val_scenes, test, max_sweeps=max_sweeps)
+
+    metadata = dict(version=version)
+    if test:
+        print(f'test sample: {len(train_lyft_infos)}')
+        data = dict(infos=train_lyft_infos, metadata=metadata)
+        info_name = f'{info_prefix}_infos_test'
+        info_path = osp.join(root_path, f'{info_name}.pkl')
+        mmcv.dump(data, info_path)
+    else:
+        print(f'train sample: {len(train_lyft_infos)}, \
+                val sample: {len(val_lyft_infos)}')
+        data = dict(infos=train_lyft_infos, metadata=metadata)
+        train_info_name = f'{info_prefix}_infos_train'
+        info_path = osp.join(root_path, f'{train_info_name}.pkl')
+        mmcv.dump(data, info_path)
+        data['infos'] = val_lyft_infos
+        val_info_name = f'{info_prefix}_infos_val'
+        info_val_path = osp.join(root_path, f'{val_info_name}.pkl')
+        mmcv.dump(data, info_val_path)
+
+
+def _fill_trainval_infos(lyft,
+                         train_scenes,
+                         val_scenes,
+                         test=False,
+                         max_sweeps=10):
+    """Generate the train/val infos from the raw data.
+
+    Args:
+        lyft (:obj:``LyftDataset``): Dataset class in the Lyft dataset.
+        train_scenes (list[str]): Basic information of training scenes.
+        val_scenes (list[str]): Basic information of validation scenes.
+        test (bool): Whether use the test mode. In the test mode, no
+            annotations can be accessed. Default: False.
+        max_sweeps (int): Max number of sweeps. Default: 10.
+
+    Returns:
+        tuple[list[dict]]: Information of training set and
+            validation set that will be saved to the info file.
+    """
+    train_lyft_infos = []
+    val_lyft_infos = []
+
+    for sample in mmcv.track_iter_progress(lyft.sample):
+        lidar_token = sample['data']['LIDAR_TOP']
+        sd_rec = lyft.get('sample_data', sample['data']['LIDAR_TOP'])
+        cs_record = lyft.get('calibrated_sensor',
+                             sd_rec['calibrated_sensor_token'])
+        pose_record = lyft.get('ego_pose', sd_rec['ego_pose_token'])
+        lidar_path, boxes, _ = lyft.get_sample_data(lidar_token)
+        lidar_path = str(lidar_path)
+
+        mmcv.check_file_exist(lidar_path)
+
+        info = {
+            'lidar_path': lidar_path,
+            'token': sample['token'],
+            'sweeps': [],
+            'cams': dict(),
+            'lidar2ego_translation': cs_record['translation'],
+            'lidar2ego_rotation': cs_record['rotation'],
+            'ego2global_translation': pose_record['translation'],
+            'ego2global_rotation': pose_record['rotation'],
+            'timestamp': sample['timestamp'],
+        }
+
+        l2e_r = info['lidar2ego_rotation']
+        l2e_t = info['lidar2ego_translation']
+        e2g_r = info['ego2global_rotation']
+        e2g_t = info['ego2global_translation']
+        l2e_r_mat = Quaternion(l2e_r).rotation_matrix
+        e2g_r_mat = Quaternion(e2g_r).rotation_matrix
+
+        # obtain 6 image's information per frame
+        camera_types = [
+            'CAM_FRONT',
+            'CAM_FRONT_RIGHT',
+            'CAM_FRONT_LEFT',
+            'CAM_BACK',
+            'CAM_BACK_LEFT',
+            'CAM_BACK_RIGHT',
+        ]
+        for cam in camera_types:
+            cam_token = sample['data'][cam]
+            cam_path, _, cam_intrinsic = lyft.get_sample_data(cam_token)
+            cam_info = obtain_sensor2top(lyft, cam_token, l2e_t, l2e_r_mat,
+                                         e2g_t, e2g_r_mat, cam)
+            cam_info.update(cam_intrinsic=cam_intrinsic)
+            info['cams'].update({cam: cam_info})
+
+        # obtain sweeps for a single key-frame
+        sd_rec = lyft.get('sample_data', sample['data']['LIDAR_TOP'])
+        sweeps = []
+        while len(sweeps) < max_sweeps:
+            if not sd_rec['prev'] == '':
+                sweep = obtain_sensor2top(lyft, sd_rec['prev'], l2e_t,
+                                          l2e_r_mat, e2g_t, e2g_r_mat, 'lidar')
+                sweeps.append(sweep)
+                sd_rec = lyft.get('sample_data', sd_rec['prev'])
+            else:
+                break
+        info['sweeps'] = sweeps
+        # obtain annotation
+        if not test:
+            annotations = [
+                lyft.get('sample_annotation', token)
+                for token in sample['anns']
+            ]
+            locs = np.array([b.center for b in boxes]).reshape(-1, 3)
+            dims = np.array([b.wlh for b in boxes]).reshape(-1, 3)
+            rots = np.array([b.orientation.yaw_pitch_roll[0]
+                             for b in boxes]).reshape(-1, 1)
+
+            names = [b.name for b in boxes]
+            for i in range(len(names)):
+                if names[i] in LyftDataset.NameMapping:
+                    names[i] = LyftDataset.NameMapping[names[i]]
+            names = np.array(names)
+
+            # we need to convert rot to SECOND format.
+            gt_boxes = np.concatenate([locs, dims, -rots - np.pi / 2], axis=1)
+            assert len(gt_boxes) == len(
+                annotations), f'{len(gt_boxes)}, {len(annotations)}'
+            info['gt_boxes'] = gt_boxes
+            info['gt_names'] = names
+            info['num_lidar_pts'] = np.array(
+                [a['num_lidar_pts'] for a in annotations])
+            info['num_radar_pts'] = np.array(
+                [a['num_radar_pts'] for a in annotations])
+
+        if sample['scene_token'] in train_scenes:
+            train_lyft_infos.append(info)
+        else:
+            val_lyft_infos.append(info)
+
+    return train_lyft_infos, val_lyft_infos
+
+
+def export_2d_annotation(root_path, info_path, version):
+    """Export 2d annotation from the info file and raw data.
+
+    Args:
+        root_path (str): Root path of the raw data.
+        info_path (str): Path of the info file.
+        version (str): Dataset version.
+    """
+    # get bbox annotations for camera
+    camera_types = [
+        'CAM_FRONT',
+        'CAM_FRONT_RIGHT',
+        'CAM_FRONT_LEFT',
+        'CAM_BACK',
+        'CAM_BACK_LEFT',
+        'CAM_BACK_RIGHT',
+    ]
+    lyft_infos = mmcv.load(info_path)['infos']
+    lyft = Lyft(
+        data_path=osp.join(root_path, version),
+        json_path=osp.join(root_path, version, version),
+        verbose=True)
+    # info_2d_list = []
+    cat2Ids = [
+        dict(id=lyft_categories.index(cat_name), name=cat_name)
+        for cat_name in lyft_categories
+    ]
+    coco_ann_id = 0
+    coco_2d_dict = dict(annotations=[], images=[], categories=cat2Ids)
+    for info in mmcv.track_iter_progress(lyft_infos):
+        for cam in camera_types:
+            cam_info = info['cams'][cam]
+            coco_infos = get_2d_boxes(
+                lyft,
+                cam_info['sample_data_token'],
+                visibilities=['', '1', '2', '3', '4'])
+            (height, width, _) = mmcv.imread(cam_info['data_path']).shape
+            coco_2d_dict['images'].append(
+                dict(
+                    file_name=cam_info['data_path'],
+                    id=cam_info['sample_data_token'],
+                    width=width,
+                    height=height))
+            for coco_info in coco_infos:
+                if coco_info is None:
+                    continue
+                # add an empty key for coco format
+                coco_info['segmentation'] = []
+                coco_info['id'] = coco_ann_id
+                coco_2d_dict['annotations'].append(coco_info)
+                coco_ann_id += 1
+    mmcv.dump(coco_2d_dict, f'{info_path[:-4]}.coco.json')

tools/data_converter/nuscenes_converter.py

@@ -50,7 +50,7 @@ def create_nuscenes_infos(root_path,
         raise ValueError('unknown')
 
     # filter existing scenes.
-    available_scenes = _get_available_scenes(nusc)
+    available_scenes = get_available_scenes(nusc)
     available_scene_names = [s['name'] for s in available_scenes]
     train_scenes = list(
         filter(lambda x: x in available_scene_names, train_scenes))
@@ -93,7 +93,19 @@ def create_nuscenes_infos(root_path,
         mmcv.dump(data, info_val_path)
 
 
-def _get_available_scenes(nusc):
+def get_available_scenes(nusc):
+    """Get available scenes from the input nuscenes class.
+
+    Given the raw data, get the information of available scenes for
+    further info generation.
+
+    Args:
+        nusc (class): Dataset class in the nuScenes dataset.
+
+    Returns:
+        available_scenes (list[dict]): List of basic information for the
+            available scenes.
+    """
     available_scenes = []
     print('total scene num: {}'.format(len(nusc.scene)))
     for scene in nusc.scene:
@@ -105,6 +117,7 @@ def _get_available_scenes(nusc):
         scene_not_exist = False
         while has_more_frames:
             lidar_path, boxes, _ = nusc.get_sample_data(sd_rec['token'])
+            lidar_path = str(lidar_path)
             if not mmcv.is_filepath(lidar_path):
                 scene_not_exist = True
                 break
@@ -126,6 +139,20 @@ def _fill_trainval_infos(nusc,
                          val_scenes,
                          test=False,
                          max_sweeps=10):
+    """Generate the train/val infos from the raw data.
+
+    Args:
+        nusc (:obj:``NuScenes``): Dataset class in the nuScenes dataset.
+        train_scenes (list[str]): Basic information of training scenes.
+        val_scenes (list[str]): Basic information of validation scenes.
+        test (bool): Whether use the test mode. In the test mode, no
+            annotations can be accessed. Default: False.
+        max_sweeps (int): Max number of sweeps. Default: 10.
+
+    Returns:
+        tuple[list[dict]]: Information of training set and validation set
+            that will be saved to the info file.
+    """
     train_nusc_infos = []
     val_nusc_infos = []
 
@@ -137,7 +164,7 @@ def _fill_trainval_infos(nusc,
         pose_record = nusc.get('ego_pose', sd_rec['ego_pose_token'])
         lidar_path, boxes, _ = nusc.get_sample_data(lidar_token)
 
-        mmcv.check_file_exist(lidar_path, msg_tmpl='file "{}" does not exist.')
+        mmcv.check_file_exist(lidar_path)
 
         info = {
             'lidar_path': lidar_path,
@@ -238,13 +265,28 @@ def obtain_sensor2top(nusc,
                       e2g_t,
                       e2g_r_mat,
                       sensor_type='lidar'):
-    """Obtain the info with RT matric from general sensor to Top LiDAR
+    """Obtain the info with RT matric from general sensor to Top LiDAR.
+
+    Args:
+        nusc (class): Dataset class in the nuScenes dataset.
+        sensor_token (str): Sample data token corresponding to the
+            specific sensor type.
+        l2e_t (np.ndarray): Translation from lidar to ego in shape (1, 3).
+        l2e_r_mat (np.ndarray): Rotation matrix from lidar to ego
+            in shape (3, 3).
+        e2g_t (np.ndarray): Translation from ego to global in shape (1, 3).
+        e2g_r_mat (np.ndarray): Rotation matrix from ego to global
+            in shape (3, 3).
+        sensor_type (str): Sensor to calibrate. Default: 'lidar'.
+
+    Returns:
+        sweep (dict): Sweep information after transformation.
     """
     sd_rec = nusc.get('sample_data', sensor_token)
     cs_record = nusc.get('calibrated_sensor',
                          sd_rec['calibrated_sensor_token'])
     pose_record = nusc.get('ego_pose', sd_rec['ego_pose_token'])
-    data_path = nusc.get_sample_data_path(sd_rec['token'])
+    data_path = str(nusc.get_sample_data_path(sd_rec['token']))
     sweep = {
         'data_path': data_path,
         'type': sensor_type,
@@ -276,6 +318,13 @@ def obtain_sensor2top(nusc,
 
 
 def export_2d_annotation(root_path, info_path, version):
+    """Export 2d annotation from the info file and raw data.
+
+    Args:
+        root_path (str): Root path of the raw data.
+        info_path (str): Path of the info file.
+        version (str): Dataset version.
+    """
     # get bbox annotations for camera
     camera_types = [
         'CAM_FRONT',
@@ -295,9 +344,6 @@ def export_2d_annotation(root_path, info_path, version):
     coco_ann_id = 0
     coco_2d_dict = dict(annotations=[], images=[], categories=cat2Ids)
     for info in mmcv.track_iter_progress(nusc_infos):
-        # info_2d = dict(token=info['token'],
-        #                timestamp=info['timestamp'],
-        #                cams=dict())
         for cam in camera_types:
             cam_info = info['cams'][cam]
             coco_infos = get_2d_boxes(
@@ -319,27 +365,7 @@ def export_2d_annotation(root_path, info_path, version):
                 coco_info['id'] = coco_ann_id
                 coco_2d_dict['annotations'].append(coco_info)
                 coco_ann_id += 1
-            # gt_bbox_2d = [res['bbox_corners'] for res in anno_info]
-            # gt_names_2d = [res['category_name'] for res in anno_info]
-            # for i in range(len(gt_names_2d)):
-            #     if gt_names_2d[i] in NuScenesDataset.NameMapping:
-            #         gt_names_2d[i] = NuScenesDataset.NameMapping[
-            #               gt_names_2d[i]]
-            # assert len(gt_bbox_2d) == len(gt_names_2d)
-            # gt_bbox_2d = np.array(gt_bbox_2d, dtype=np.float32)
-            # gt_names_2d = np.array(gt_names_2d)
-            # info_2d['cams'][cam] = dict(
-            #     data_path=info['cams'][cam]['data_path'],
-            #     type=info['cams'][cam]['type'],
-            #     token=info['cams'][cam]['sample_data_token'],
-            #     gt_boxes=gt_bbox_2d,
-            #     gt_names=gt_names_2d)
-        # info_2d_list.append(info_2d)
-    # mmcv.dump(
-    #     info_2d_list,
-    #     osp.join(root_path,
-    #     '{}_2d_infos_train.pkl'.format(info_prefix)))
-    mmcv.dump(coco_2d_dict, '{}.coco.json'.format(info_path[:-4]))
+    mmcv.dump(coco_2d_dict, f'{info_path[:-4]}.coco.json')
 
 
 def get_2d_boxes(nusc, sample_data_token: str,
@@ -351,7 +377,7 @@ def get_2d_boxes(nusc, sample_data_token: str,
         visibilities: Visibility filter.
 
     Return:
-        list: List of 2D annotation record that belongs to the input
+        list[dict]: List of 2D annotation record that belongs to the input
             `sample_data_token`.
     """
 
@@ -436,12 +462,14 @@ def post_process_coords(
     bbox corners and the image canvas, return None if no
     intersection.
 
-    corner_coords: Corner coordinates of reprojected bounding box.
-    imsize: Size of the image canvas.
+    Args:
+        corner_coords (list[int]): Corner coordinates of reprojected
+            bounding box.
+        imsize (tuple[int]): Size of the image canvas.
 
     Return:
-        Intersection of the convex hull of the 2D box corners and the image
-        canvas.
+        tuple [float]: Intersection of the convex hull of the 2D box
+            corners and the image canvas.
     """
     polygon_from_2d_box = MultiPoint(corner_coords).convex_hull
     img_canvas = box(0, 0, imsize[0], imsize[1])
@@ -466,15 +494,26 @@ def generate_record(ann_rec: dict, x1: float, y1: float, x2: float, y2: float,
     """
     Generate one 2D annotation record given various informations on
     top of the 2D bounding box coordinates.
-    :param ann_rec: Original 3d annotation record.
-    :param x1: Minimum value of the x coordinate.
-    :param y1: Minimum value of the y coordinate.
-    :param x2: Maximum value of the x coordinate.
-    :param y2: Maximum value of the y coordinate.
-    :param sample_data_token: Sample data token.
-    :param filename:The corresponding image file where the annotation
+
+    Args:
+        ann_rec (dict): Original 3d annotation record.
+        x1 (float): Minimum value of the x coordinate.
+        y1 (float): Minimum value of the y coordinate.
+        x2 (float): Maximum value of the x coordinate.
+        y2 (float): Maximum value of the y coordinate.
+        sample_data_token (str): Sample data token.
+        filename (str):The corresponding image file where the annotation
             is present.
-    :return: A sample 2D annotation record.
+
+    Returns:
+        dict: A sample 2D annotation record.
+            - file_name (str): flie name
+            - image_id (str): sample data token
+            - area (float): 2d box area
+            - category_name (str): category name
+            - category_id (int): category id
+            - bbox (list[float]): left x, top y, dx, dy of 2d box
+            - iscrowd (int): whether the area is crowd
     """
     repro_rec = OrderedDict()
     repro_rec['sample_data_token'] = sample_data_token