{"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/drop-connect","entry":"drop_connect","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":9,"n_papers":7,"n_buckets":2,"n_distinct_outputs":4,"n_class_bearing":0,"not_compared":{"not_run_on_shared_input":{"n":0,"by_error":{}},"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/drop-connect","buckets":[{"bucket":[[4,"float","float32"]],"n_implementations_compared":8,"n_papers":7,"n_not_digested":0,"not_digested_by_type":{},"clusters":[{"output_sha":"fae90c40e4854c2d","size":4,"n_papers":6,"shape":[2,3,4,4],"dtype":"float32","type":"Tensor","finite":true,"max_difference_among_recorded_values":0.0,"members_with_recorded_values":4,"torch_reference_conventions_with_this_digest":[],"values":[0.0,0.0,0.0,0.0,0.0,-0.0,-0.0,0.0,0.0,-0.0,-0.0,-0.0,-0.0,-0.0,-0.0,0.0,-0.0,0.0,0.0,-0.0,-0.0,0.0,-0.0,0.0,0.0,0.0,-0.0,-0.0,0.0,0.0,0.0,-0.0,-0.0,0.0,-0.0,-0.0,-0.0,0.0,0.0,-0.0,-0.0,-0.0,-0.0,0.0,0.0,-0.0,0.0,-0.0,2.0639593601226807,0.9513998627662659,4.2033185958862305,-0.03755061700940132,3.308645248413086,-1.2489731311798096,-4.816895484924316,-0.42870891094207764,-2.7740046977996826,-4.0130934715271,-2.7374205589294434,-0.052834078669548035,-5.59221887588501,-1.4770385026931763,-1.2697049379348755,0.4257344603538513,1.6881035566329956,-2.181506633758545,1.5447407960891724,4.188474178314209,0.9692551493644714,0.044237371534109116,6.2088446617126465,0.1472381055355072,1.2120463848114014,0.5684263706207275,1.0316386222839355,-1.3393075466156006,2.1854777336120605,0.53412264585495,1.3243402242660522,0.4742993116378784,-4.4336161613464355,-0.17499640583992004,0.33080628514289856,-1.6774734258651733,1.4334758520126343,2.460728168487549,-3.0427823066711426,2.1557228565216064,0.45674654841423035,0.8783701062202454,3.9451634883880615,-2.3373141288757324,0.9950191974639893,0.29571279883384705,-2.735473394393921,-0.5712951421737671],"values_recorded":96,"members":[{"code_sha256_prefix":"2dd5dfa94542911e","path":"models/MAT.py","papers":["2103.02406","2606.19184"],"paper_pages":[{"arxiv_id":"2103.02406","page":"/paper/multi-attentional-deepfake-detection"},{"arxiv_id":"2606.19184","page":"/paper/arxiv-2606-19184"}],"arg_sig_recorded":[["inputs",4,"float","float32"]],"scalar_args":{"p":"0.5","training":"True"},"class_bearing":false},{"code_sha256_prefix":"d3319e3d34ca90ca","path":"models/MI_Net.py","papers":["2403.01786","2402.02003"],"paper_pages":[{"arxiv_id":"2403.01786","page":"/paper/exposing-the-deception-uncovering-more"},{"arxiv_id":"2402.02003","page":"/paper/genface-a-large-scale-fine-grained-face"}],"arg_sig_recorded":[["inputs",4,"float","float32"]],"scalar_args":{"p":"0.5","training":"True"},"class_bearing":false},{"code_sha256_prefix":"1667bcdf49534735","path":"utils/efficientnet_pytorch/model.py","papers":["1905.11946"],"paper_pages":[{"arxiv_id":"1905.11946","page":"/paper/efficientnet-rethinking-model-scaling-for"}],"arg_sig_recorded":[["inputs",4,"float","float32"]],"scalar_args":{"p":"0.5","training":"True"},"class_bearing":false},{"code_sha256_prefix":"6053cb1369b46c73","path":"models/attentive_nas_dynamic_model.py","papers":["2102.07954"],"paper_pages":[{"arxiv_id":"2102.07954","page":"/paper/alphanet-improved-training-of-supernet-with"}],"arg_sig_recorded":[["inputs",4,"float","float32"]],"scalar_args":{"p":"0.5","training":"True"},"class_bearing":false}]},{"output_sha":"780f05c3265726f3","size":3,"n_papers":2,"shape":[2,3,4,4],"dtype":"float32","type":"Tensor","finite":true,"max_difference_among_recorded_values":0.0,"members_with_recorded_values":3,"torch_reference_conventions_with_this_digest":[],"values":[0.9038809537887573,0.8163491487503052,1.807607650756836,1.373175024986267,1.4216196537017822,-0.24990904331207275,-0.33935874700546265,1.0399739742279053,0.4291916489601135,-0.165479838848114,-1.4881962537765503,-0.015892446041107178,-0.8863425850868225,-0.5452536940574646,-3.332146406173706,1.0162031650543213,-0.5662907361984253,0.1569153517484665,0.7410518527030945,-1.8297579288482666,-0.9062730669975281,1.43929922580719,-0.49693870544433594,1.7919024229049683,0.27958348393440247,1.4808681011199951,-2.585092782974243,-0.21134039759635925,0.467806339263916,0.004892400931566954,1.1191266775131226,-0.6225588321685791,-0.526714563369751,1.2757916450500488,-0.6045889854431152,-0.606835126876831,-0.827340304851532,2.478611469268799,0.2657206058502197,-1.5050290822982788,-0.2650585472583771,-1.8492885828018188,-0.9282277226448059,1.521728277206421,0.02837824635207653,-0.20480617880821228,0.415722131729126,-1.5580719709396362,1.289974570274353,0.5946248769760132,2.6270740032196045,-0.02346913516521454,2.0679032802581787,-0.7806081771850586,-3.010559558868408,-0.26794305443763733,-1.733752965927124,-2.508183479309082,-1.7108877897262573,-0.03302129730582237,-3.4951367378234863,-0.923149049282074,-0.793565571308136,0.26608404517173767,1.0550646781921387,-1.3634415864944458,0.9654629826545715,2.6177964210510254,0.6057844758033752,0.027648357674479485,3.880527973175049,0.0920238122344017,0.7575289607048035,0.3552664816379547,0.6447741389274597,-0.837067186832428,1.365923523902893,0.3338266611099243,0.8277126550674438,0.2964370548725128,-2.771010160446167,-0.10937274992465973,0.2067539244890213,-1.0484209060668945,0.8959224224090576,1.5379550457000732,-1.9017388820648193,1.3473267555236816,0.28546658158302307,0.5489813089370728,2.4657270908355713,-1.460821270942688,0.6218869686126709,0.1848205029964447,-1.709670901298523,-0.35705944895744324],"values_recorded":96,"members":[{"code_sha256_prefix":"01ef2c316f547d5d","path":"classy_vision/models/efficientnet.py","papers":["1905.11946"],"paper_pages":[{"arxiv_id":"1905.11946","page":"/paper/efficientnet-rethinking-model-scaling-for"}],"arg_sig_recorded":[["inputs",4,"float","float32"]],"scalar_args":{"is_training":"True","drop_connect_rate":"0.2"},"class_bearing":false},{"code_sha256_prefix":"53d13d88994382df","path":"models/efficientnet.py","papers":["1905.11946"],"paper_pages":[{"arxiv_id":"1905.11946","page":"/paper/efficientnet-rethinking-model-scaling-for"}],"arg_sig_recorded":[["x",4,"float","float32"]],"scalar_args":{"drop_p":"0.2","training":"True"},"class_bearing":false},{"code_sha256_prefix":"ed953cd5fd3a44c8","path":"projected_gan.py","papers":["2111.01007"],"paper_pages":[{"arxiv_id":"2111.01007","page":"/paper/projected-gans-converge-faster"}],"arg_sig_recorded":[["x",4,"float","float32"]],"scalar_args":{"drop_connect_rate":"0.2","training":"True"},"class_bearing":false}]},{"output_sha":"a8c38328240cb5f9","size":1,"n_papers":1,"shape":[2,3,4,4],"dtype":"float32","type":"Tensor","finite":true,"max_difference_among_recorded_values":null,"members_with_recorded_values":1,"torch_reference_conventions_with_this_digest":[],"values":[0.0,0.0,0.0,0.0,0.0,-0.0,-0.0,0.0,0.0,-0.0,-0.0,-0.0,-0.0,-0.0,-0.0,0.0,-0.0,0.0,0.0,-0.0,-0.0,0.0,-0.0,0.0,0.0,0.0,-0.0,-0.0,0.0,0.0,0.0,-0.0,-0.0,0.0,-0.0,-0.0,-0.0,0.0,0.0,-0.0,-0.0,-0.0,-0.0,0.0,0.0,-0.0,0.0,-0.0,0.0,0.0,0.0,-0.0,0.0,-0.0,-0.0,-0.0,-0.0,-0.0,-0.0,-0.0,-0.0,-0.0,-0.0,0.0,0.0,-0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,-0.0,0.0,0.0,0.0,0.0,-0.0,-0.0,0.0,-0.0,0.0,0.0,-0.0,0.0,0.0,0.0,0.0,-0.0,0.0,0.0,-0.0,-0.0],"values_recorded":96,"members":[{"code_sha256_prefix":"79583b4154444ddf","path":"pycls/models/effnet.py","papers":["1905.11946"],"paper_pages":[{"arxiv_id":"1905.11946","page":"/paper/efficientnet-rethinking-model-scaling-for"}],"arg_sig_recorded":[["x",4,"float","float32"]],"scalar_args":{"drop_ratio":"0.5"},"class_bearing":false}]}]},{"bucket":[[5,"float","float32"]],"n_implementations_compared":1,"n_papers":1,"n_not_digested":0,"not_digested_by_type":{},"clusters":[{"output_sha":"9ee24930617c81cf","size":1,"n_papers":1,"shape":[2,2,3,4,4],"dtype":"float32","type":"Tensor","finite":true,"max_difference_among_recorded_values":null,"members_with_recorded_values":1,"torch_reference_conventions_with_this_digest":[],"values":[-0.0,-0.0,0.0,0.0,-0.0,-0.0,0.0,-0.0,0.0,-0.0,-0.0,-0.0,-0.0,0.0,-0.0,0.0,-0.0,-0.0,0.0,0.0,0.0,-0.0,0.0,-0.0,-0.0,-0.0,0.0,0.0,0.0,-0.0,-0.0,0.0,0.0,-0.0,-0.0,0.0,-0.0,-0.0,-0.0,-0.0,-0.0,-0.0,-0.0,-0.0,0.0,-0.0,0.0,-0.0,-0.0,0.0,-0.0,0.0,-0.0,-0.0,-0.0,0.0,0.0,0.0,0.0,-0.0,-0.0,-0.0,-0.0,-0.0,0.0,-0.0,0.0,-0.0,0.0,-0.0,0.0,-0.0,0.0,-0.0,0.0,0.0,-0.0,-0.0,0.0,-0.0,0.0,-0.0,-0.0,-0.0,-0.0,-0.0,0.0,0.0,-0.0,-0.0,-0.0,-0.0,-0.0,-0.0,0.0,-0.0,0.06441031396389008,0.6007662415504456,0.14865140616893768,4.448735237121582,2.459325075149536,-0.6932435035705566,-0.28722909092903137,4.165374279022217,0.17993491888046265,1.8237569332122803,0.9721740484237671,-1.8052058219909668,-0.9143292307853699,2.103243827819824,0.6775014400482178,3.3142647743225098,1.9914265871047974,-0.9509482979774475,-0.0831359401345253,-2.3738226890563965,2.2707324028015137,0.8995675444602966,-1.2713040113449097,1.3493634462356567,0.9074420928955078,1.4637775421142578,1.0660710334777832,0.04898466169834137,-2.4074432849884033,-0.3232475519180298,6.773853778839111,0.5426764488220215,-0.7959375381469727,-1.3189690113067627,-4.02977180480957,-2.2922449111938477,-0.04107870161533356,1.1144094467163086,-0.390829473733902,0.494262158870697,-1.176296591758728,-0.8357838988304138,2.635913372039795,2.9399800300598145,0.38253355026245117,-2.4592363834381104,0.723416268825531,1.1530719995498657,-4.180145263671875,0.05654565244913101,-2.4080212116241455,-3.936661958694458,2.128467559814453,1.6304186582565308,-1.1305887699127197,-1.0827977657318115,5.432962894439697,-0.2644973695278168,-2.7419021129608154,2.6128151416778564,3.3348472118377686,-2.9290919303894043,1.8370847702026367,0.8006091117858887,-0.9168629050254822,1.342897891998291,4.72020959854126,-3.373908519744873,-2.1549084186553955,-3.794008493423462,1.714280366897583,0.022738352417945862,0.05202436074614525,-1.2175214290618896,-2.1293137073516846,1.7599223852157593,0.31363219022750854,-0.7491605877876282,-1.6219093799591064,0.8541411757469177,3.211225986480713,3.1815099716186523,-0.20890863239765167,1.5249905586242676,-2.315267324447632,0.7002894282341003,-1.1612242460250854,0.0811651274561882,0.6036227941513062,-1.855951189994812,-0.5900571942329407,0.6964305639266968,0.048224229365587234,2.1174588203430176,-2.4959349632263184,1.288385033607483],"values_recorded":192,"members":[{"code_sha256_prefix":"4e82aac968b327d5","path":"efficientnet_pytorch_3d/model.py","papers":["1905.11946"],"paper_pages":[{"arxiv_id":"1905.11946","page":"/paper/efficientnet-rethinking-model-scaling-for"}],"arg_sig_recorded":[["inputs",5,"float","float32"]],"scalar_args":{"p":"0.5","training":"True"},"class_bearing":false}]}]}]}