{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/census/sinkhorn","entry":"sinkhorn","source":"Syntology differential census (groundwork/55, run v2_2026-09-22), per sample; not an archive number","census_date":"2026-09-22","battery_sha256":["b986f7e04d794a0d88ad4c5f32cf63ec3590b5deff0192150737bc5f1c0b4677"],"runner_sha256":["5a452d0e7c0da5b80771d1be2afe3572e253568cf5c5d59a0d08ebd663d00808"],"bucket_key":"positional (rank, kind, dtype) of each array argument; the argument name is not part of the key because the harness draws the shared array from (rank, kind) and casts it to the dtype, whatever the name","bucket_fields":["rank","kind","dtype"],"claim":"Implementations sharing this entry name were each run on one shared input fixed by the positional (rank, kind, dtype) of their array arguments (the bucket). A cluster is the set whose recorded output digest (sha256 of the output rounded to 6 decimals) is identical. Identical values to six decimals on the shared input are agreement on those inputs, not a statement about the whole domain and not a substitution claim.","n_implementations_compared":11,"n_papers":11,"n_buckets":2,"n_distinct_outputs":11,"n_class_bearing":0,"not_compared":{"not_run_on_shared_input":{"n":9,"by_error":{"RuntimeError":4,"ValueError":3,"AssertionError":1,"IndexError":1}},"no_array_argument_ran_on_own_fixture_arguments_only":{"n":0,"n_papers":0,"recorded_shared_digest_equals_own_fixture_digest":0},"output_not_digested_non_numeric":{"n":0,"by_type":{}}},"withdrawn_excluded":0,"code_page":"/code/sinkhorn","buckets":[{"bucket":[[2,"float","float32"]],"n_implementations_compared":10,"n_papers":10,"n_not_digested":0,"not_digested_by_type":{},"clusters":[{"output_sha":"51e63d44301d5cc2","size":1,"n_papers":1,"shape":[4,8],"dtype":"float32","type":"Tensor","finite":true,"max_difference_among_recorded_values":null,"members_with_recorded_values":1,"torch_reference_conventions_with_this_digest":[],"values":[0.09390107542276382,-0.1590421199798584,-0.05283842608332634,0.039471227675676346,0.9753065705299377,-0.07446111738681793,-0.15185116231441498,0.3295139968395233,0.02880423329770565,0.10598956793546677,0.04698201268911362,0.0954870730638504,0.3321642279624939,0.1788153052330017,-0.12569405138492584,0.33745163679122925,0.017986517399549484,-0.8751447796821594,-0.24942193925380707,0.12068213522434235,0.12693101167678833,-0.44371914863586426,-0.6396958231925964,2.9423818588256836,-0.06771592050790787,0.07340708374977112,-0.17137235403060913,-0.0949818342924118,2.163540840148926,0.15043962001800537,-0.494983047246933,-0.5583341717720032],"values_recorded":32,"members":[{"code_sha256_prefix":"c1dff37f3b19bb53","path":"lightly/loss/msn_loss.py","papers":["2204.07141"],"paper_pages":[{"arxiv_id":"2204.07141","page":"/paper/masked-siamese-networks-for-label-efficient"}],"arg_sig_recorded":[["probabilities",2,"float","float32"]],"scalar_args":{"iterations":"3","gather_distributed":"False"},"class_bearing":false}]},{"output_sha":"6b36e497308829af","size":1,"n_papers":1,"shape":[4,8],"dtype":"float32","type":"Tensor","finite":true,"max_difference_among_recorded_values":null,"members_with_recorded_values":1,"torch_reference_conventions_with_this_digest":[],"values":[0.8679903745651245,1.154862678829976e-21,0.00026541974511928856,4.5507560834077716e-15,9.982092308291612e-08,3.9414466266235207e-29,0.1305469572544098,0.001197113306261599,4.744660610581097e-13,0.40178975462913513,5.33554754197918e-13,8.40964458036045e-24,0.0016722894506528974,0.4018077254295349,0.19467103481292725,5.917485759709962e-05,6.888797860777547e-17,1.857718206874997e-23,1.053457071975572e-05,1.292183997607041e-13,0.4220641255378723,1.8639160346129956e-31,0.15474414825439453,0.4231811761856079,3.73598608255882e-24,2.2389509467757307e-05,0.49990609288215637,0.5000714659690857,2.489783552939162e-11,6.544257913354556e-10,3.887672406222009e-08,1.3920197900450532e-14],"values_recorded":32,"members":[{"code_sha256_prefix":"ede20b10087a299b","path":"model.py","papers":["2207.12795"],"paper_pages":[{"arxiv_id":"2207.12795","page":"/paper/static-and-dynamic-concepts-for-self"}],"arg_sig_recorded":[["scores",2,"float","float32"]],"scalar_args":{"eps":"0.05","niters":"3"},"class_bearing":false}]},{"output_sha":"93d803bcd4487dcb","size":1,"n_papers":1,"shape":[8,4],"dtype":"float32","type":"Tensor","finite":true,"max_difference_among_recorded_values":null,"members_with_recorded_values":1,"torch_reference_conventions_with_this_digest":[],"values":[20.173954010009766,-6.277158260345459,0.5947825908660889,-13.491578102111816,0.4773489534854889,0.32268089056015015,0.4042913019657135,-0.20432108640670776,0.17742322385311127,0.1600215584039688,0.12890979647636414,0.5336454510688782,-0.3110578954219818,0.7632932662963867,-0.14638419449329376,0.69414883852005,0.36606350541114807,-0.12646037340164185,0.007332865614444017,0.753063976764679,0.40330860018730164,0.9824239015579224,0.369919091463089,-0.755651593208313,0.26098188757896423,-0.21912570297718048,0.16922186315059662,0.7889219522476196,-4.242351055145264,4.4068732261657715,-5.830723285675049,6.666201114654541],"values_recorded":32,"members":[{"code_sha256_prefix":"0988ed5686f818dc","path":"time_tuning.py","papers":["2308.11796"],"paper_pages":[{"arxiv_id":"2308.11796","page":"/paper/time-does-tell-self-supervised-time-tuning-of"}],"arg_sig_recorded":[["Q",2,"float","float32"]],"scalar_args":{"nmb_iters":"2","world_size":"1"},"class_bearing":false}]},{"output_sha":"96f9141d2b84b842","size":1,"n_papers":1,"shape":[4,8],"dtype":"float32","type":"Tensor","finite":true,"max_difference_among_recorded_values":null,"members_with_recorded_values":1,"torch_reference_conventions_with_this_digest":[],"values":[1.0,1.0479876322326614e-29,3.0027109271246744e-12,2.195832826094861e-22,3.166042164360918e-24,2.4660864255001864e-25,3.8538907165116546e-22,2.082667371894331e-11,2.1743008854239809e-16,1.4502862886317525e-12,2.400982573157008e-24,1.6140680788717174e-34,2.1097770943495683e-23,1.0,2.2859352466509954e-25,4.0949725778692385e-16,1.0779815262651482e-08,2.2897579186862767e-23,1.6187505025300197e-05,8.46880779479392e-13,1.8182640948438689e-09,1.5840254113858178e-19,6.204820592872404e-14,0.9999838471412659,1.0435912233243307e-18,4.926167562757655e-08,0.0013712226646021008,0.0058504133485257626,1.9146785080707681e-22,0.9927783012390137,2.782674342097824e-23,5.871774003687679e-17],"values_recorded":32,"members":[{"code_sha256_prefix":"3e002910aaac8995","path":"models/layers.py","papers":["2404.01340"],"paper_pages":[{"arxiv_id":"2404.01340","page":"/paper/from-similarity-to-superiority-channel"}],"arg_sig_recorded":[["out",2,"float","float32"]],"scalar_args":{"epsilon":"0.05","sinkhorn_iterations":"3"},"class_bearing":false}]},{"output_sha":"ae8daa487fdd6089","size":1,"n_papers":1,"shape":[4,8],"dtype":"float32","type":"Tensor","finite":true,"max_difference_among_recorded_values":null,"members_with_recorded_values":1,"torch_reference_conventions_with_this_digest":[],"values":[0.11719024181365967,0.0012352935736998916,0.03362153843045235,0.005221362691372633,0.014974976889789104,0.00029215821996331215,0.04184277355670929,0.03441422060132027,0.003923194017261267,0.07886721193790436,0.002558124018833041,0.0003939422604162246,0.022323109209537506,0.1040959432721138,0.024544453248381615,0.014358780346810818,0.0034352957736700773,0.0009761694236658514,0.029080096632242203,0.00871400535106659,0.08224165439605713,0.00020427025447133929,0.050829414278268814,0.07390723377466202,0.0004512677842285484,0.0439213290810585,0.05974024534225464,0.11067069321870804,0.005460269749164581,0.020407624542713165,0.007783352863043547,0.0023197587579488754],"values_recorded":32,"members":[{"code_sha256_prefix":"bfbd162761e2e475","path":"switch_mlp.py","papers":["2101.03961"],"paper_pages":[{"arxiv_id":"2101.03961","page":"/paper/switch-transformers-scaling-to-trillion"}],"arg_sig_recorded":[["cost",2,"float","float32"]],"scalar_args":{"tol":"0.0001"},"class_bearing":false}]},{"output_sha":"b1ec41a787384bfb","size":1,"n_papers":1,"shape":[8,4],"dtype":"float32","type":"Tensor","finite":true,"max_difference_among_recorded_values":null,"members_with_recorded_values":1,"torch_reference_conventions_with_this_digest":[],"values":[0.5331041216850281,0.4940335154533386,0.24756410717964172,-0.2747017443180084,0.08335258066654205,-0.16781413555145264,1.111951470375061,-0.02748994342982769,0.08281268179416656,-0.22245241701602936,0.9477212429046631,0.1919184774160385,0.07142145186662674,0.5219767689704895,0.5294073224067688,-0.12280546128749847,0.2544853985309601,0.26183828711509705,0.08029498159885406,0.4033813774585724,0.14818304777145386,-1.075058937072754,2.1408002376556396,-0.21392439305782318,0.06233302131295204,0.15587390959262848,0.6366088390350342,0.14518426358699799,0.04076406732201576,0.1261170208454132,0.8824734687805176,-0.04935452342033386],"values_recorded":32,"members":[{"code_sha256_prefix":"432bf71f80b71af8","path":"experiments/exp_mosic.py","papers":["2506.08694"],"paper_pages":[{"arxiv_id":"2506.08694","page":"/paper/mosic-optimal-transport-motion-trajectory-for"}],"arg_sig_recorded":[["Q",2,"float","float32"]],"scalar_args":{"nmb_iters":"5","world_size":"1"},"class_bearing":false}]},{"output_sha":"da8384339a925a0e","size":1,"n_papers":1,"shape":[4,8],"dtype":"float32","type":"Tensor","finite":true,"max_difference_among_recorded_values":null,"members_with_recorded_values":1,"torch_reference_conventions_with_this_digest":[],"values":[-2.994913101196289,0.21774928271770477,0.1260700225830078,-0.16157107055187225,0.5049204230308533,0.12352469563484192,0.3142646551132202,2.1924962997436523,2.440087080001831,0.38542795181274414,0.2977342903614044,1.0381567478179932,-0.4567411541938782,0.7878875732421875,-0.6909197568893433,-5.963647365570068,-0.22975113987922668,0.4798678159713745,0.2383381873369217,-0.19784408807754517,0.026317624375224113,0.2948012948036194,0.5302101969718933,7.8408098220825195,1.7845771312713623,-0.08304499089717865,0.33785751461982727,0.32125839591026306,0.925503134727478,-0.2062135636806488,0.8464449048042297,-3.0696585178375244],"values_recorded":32,"members":[{"code_sha256_prefix":"57f3995b8ea26d10","path":"model_symmetries/stitching/stitching.py","papers":["2205.14258"],"paper_pages":[{"arxiv_id":"2205.14258","page":"/paper/on-the-symmetries-of-deep-learning-models-and"}],"arg_sig_recorded":[["A",2,"float","float32"]],"scalar_args":{"iters":"16","verbose":"False"},"class_bearing":false}]},{"output_sha":"f6fc4ba3d136f539","size":1,"n_papers":1,"shape":[4,8],"dtype":"float32","type":"Tensor","finite":true,"max_difference_among_recorded_values":null,"members_with_recorded_values":1,"torch_reference_conventions_with_this_digest":[],"values":[0.8679905533790588,1.154863486623543e-21,0.00026541983243077993,4.550761589121929e-15,9.982066728753125e-08,3.9414484321828435e-29,0.13054686784744263,0.0011971142375841737,4.74466169478327e-13,0.40178972482681274,5.335552420888956e-13,8.409638269473208e-24,0.0016722874715924263,0.4018077254295349,0.1946711540222168,5.917486851103604e-05,6.888808448689388e-17,1.8577267265727734e-23,1.0534585271670949e-05,1.2921856239102997e-13,0.4220642149448395,1.8639212067881392e-31,0.15474408864974976,0.4231812357902527,3.735992787876515e-24,2.238949491584208e-05,0.49990609288215637,0.5000714659690857,2.489776267100563e-11,6.544259023577581e-10,3.88767560366432e-08,1.3920205523747058e-14],"values_recorded":32,"members":[{"code_sha256_prefix":"58db645a1b52ab84","path":"lightly/loss/swav_loss.py","papers":["2006.09882"],"paper_pages":[{"arxiv_id":"2006.09882","page":"/paper/unsupervised-learning-of-visual-features-by"}],"arg_sig_recorded":[["out",2,"float","float32"]],"scalar_args":{"iterations":"3","epsilon":"0.05","gather_distributed":"False"},"class_bearing":false}]},{"output_sha":"f973c3905903059d","size":1,"n_papers":1,"shape":[4,8],"dtype":"float32","type":"Tensor","finite":true,"max_difference_among_recorded_values":null,"members_with_recorded_values":1,"torch_reference_conventions_with_this_digest":[],"values":[0.2169976383447647,2.8871587165588574e-22,6.635495810769498e-05,1.1376903972804822e-15,2.4955166821882813e-08,9.853621080457109e-30,0.03263671696186066,0.0002992785593960434,1.1861654236958175e-13,0.10044743120670319,1.333888105222239e-13,2.102409567368302e-24,0.0004180718678981066,0.10045193135738373,0.0486677885055542,1.479371712775901e-05,1.722202112172347e-17,4.6443168164319335e-24,2.633646317917737e-06,3.230464059775749e-14,0.10551605373620987,4.659803016970348e-32,0.03868602216243744,0.10579530894756317,9.339981969691287e-25,5.59737372896052e-06,0.12497652322053909,0.12501786649227142,6.2244406677514075e-12,1.6360647558943953e-10,9.7191890091608e-09,3.4800513809367644e-15],"values_recorded":32,"members":[{"code_sha256_prefix":"2be40bf60ede15a3","path":"methods/uniot.py","papers":["1706.07522"],"paper_pages":[{"arxiv_id":"1706.07522","page":"/paper/deep-hashing-network-for-unsupervised-domain"}],"arg_sig_recorded":[["out",2,"float","float32"]],"scalar_args":{"epsilon":"0.05","sinkhorn_iterations":"3"},"class_bearing":false}]},{"output_sha":"fadf10afaef3dc75","size":1,"n_papers":1,"shape":[],"dtype":"float32","type":"Tensor","finite":true,"max_difference_among_recorded_values":null,"members_with_recorded_values":1,"torch_reference_conventions_with_this_digest":[],"values":[-0.1657937467098236],"values_recorded":1,"members":[{"code_sha256_prefix":"8cf6df4a3ab3e2a1","path":"loss.py","papers":["2410.03052"],"paper_pages":[{"arxiv_id":"2410.03052","page":"/paper/learning-structured-representations-by"}],"arg_sig_recorded":[["M",2,"float","float32"]],"scalar_args":{"reg":"10","numItermax":"10000"},"class_bearing":false}]}]},{"bucket":[[3,"float","float32"]],"n_implementations_compared":1,"n_papers":1,"n_not_digested":0,"not_digested_by_type":{},"clusters":[{"output_sha":"0d678eb6f4560989","size":1,"n_papers":1,"shape":[2,4,8],"dtype":"float32","type":"Tensor","finite":true,"max_difference_among_recorded_values":null,"members_with_recorded_values":1,"torch_reference_conventions_with_this_digest":[],"values":[0.0799625888466835,0.668081521987915,0.20390324294567108,0.29088467359542847,0.30556806921958923,0.052610624581575394,0.3021666407585144,0.09691191464662552,0.021596400067210197,0.11797633022069931,0.28845980763435364,0.16336455941200256,0.5695463418960571,0.17509019374847412,0.4639928340911865,0.20011630654335022,0.5959210991859436,0.027437176555395126,0.3063000440597534,0.3804064989089966,0.06915751099586487,0.31660038232803345,0.13205750286579132,0.1719985008239746,0.30251994729042053,0.1865050047636032,0.20133692026138306,0.1653442233800888,0.05572805181145668,0.4556988775730133,0.10178308188915253,0.5309733748435974,0.15373538434505463,0.5370232462882996,0.03032429702579975,0.17547428607940674,0.15969814360141754,0.2068093717098236,0.2200821340084076,0.5168307423591614,0.2067248821258545,0.057487040758132935,0.06095423921942711,0.3358665108680725,0.2672954797744751,0.1621871441602707,0.5289482474327087,0.38059666752815247,0.5257623791694641,0.13866601884365082,0.3852822780609131,0.26650041341781616,0.1838272660970688,0.2450800985097885,0.19117288291454315,0.06372252106666565,0.11377731710672379,0.2668237090110779,0.5234391689300537,0.2221587598323822,0.3891791105270386,0.38592347502708435,0.05979680269956589,0.038850054144859314],"values_recorded":64,"members":[{"code_sha256_prefix":"7819985853d82f66","path":"lib/learnable_sparsity.py","papers":["2510.10136"],"paper_pages":[{"arxiv_id":"2510.10136","page":"/paper/arxiv-2510-10136"}],"arg_sig_recorded":[["logits",3,"float","float32"]],"scalar_args":{"tau":"1.0","iter":"5"},"class_bearing":false}]}]}]}