{"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/batchify","entry":"batchify","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":6,"n_papers":6,"n_buckets":3,"n_distinct_outputs":3,"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/batchify","buckets":[{"bucket":[[2,"float","float32"]],"n_implementations_compared":4,"n_papers":4,"n_not_digested":0,"not_digested_by_type":{},"clusters":[{"output_sha":"7bd974ab2ba9e842","size":4,"n_papers":4,"shape":[16,2],"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":[1.8026286363601685,0.06639543920755386,-1.5337762832641602,-1.6228755712509155,0.47605323791503906,0.4321114122867584,-0.6908870339393616,-0.4061858654022217,-0.9028494954109192,-0.02259422466158867,-1.0304712057113647,-1.180782437324524,-0.6627609133720398,-0.5368682742118835,0.5728892087936401,0.983674168586731,0.5188759565353394,-0.7956128120422363,0.9591456651687622,0.43327441811561584,-0.3971995711326599,0.9449777603149414,-1.5683481693267822,1.0175182819366455,-0.28853508830070496,-1.225785732269287,2.322108745574951,1.274217963218689,-0.5147839784622192,-1.3222218751907349,0.5505285859107971,-0.5941091179847717],"values_recorded":32,"members":[{"code_sha256_prefix":"1f32d831b2c4c04b","path":"train_lstm.py","papers":["2105.13937"],"paper_pages":[{"arxiv_id":"2105.13937","page":"/paper/polygonal-unadjusted-langevin-algorithms"}],"arg_sig_recorded":[["data",2,"float","float32"]],"scalar_args":{"bsz":"2"},"class_bearing":false},{"code_sha256_prefix":"2943dadcb85e6e3c","path":"dine/utils.py","papers":["2003.04179"],"paper_pages":[{"arxiv_id":"2003.04179","page":"/paper/capacity-of-continuous-channels-with-memory"}],"arg_sig_recorded":[["data",2,"float","float32"]],"scalar_args":{"bsz":"2","args":"<fx_2943dadcb85e6e3c__batchify.Args object at 0xffff5b0057d0"},"class_bearing":false},{"code_sha256_prefix":"96d4db57e5c2248f","path":"main_lstm.py","papers":["2010.02519"],"paper_pages":[{"arxiv_id":"2010.02519","page":"/paper/improved-analysis-of-clipping-algorithms-for"}],"arg_sig_recorded":[["data",2,"float","float32"]],"scalar_args":{"bsz":"2","args":"<fx_96d4db57e5c2248f__batchify.Args object at 0xffff3e423e50"},"class_bearing":false},{"code_sha256_prefix":"c077f321f63362e6","path":"WT2/wt2_lstm.py","papers":["2208.05924"],"paper_pages":[{"arxiv_id":"2208.05924","page":"/paper/regularizing-deep-neural-networks-with-1"}],"arg_sig_recorded":[["data",2,"float","float32"]],"scalar_args":{"bsz":"2"},"class_bearing":false}]}]},{"bucket":[[1,"float","float32"]],"n_implementations_compared":1,"n_papers":1,"n_not_digested":0,"not_digested_by_type":{},"clusters":[{"output_sha":"ba9e3c49302ffd2c","size":1,"n_papers":1,"shape":[4,2],"dtype":"float32","type":"Tensor","finite":true,"max_difference_among_recorded_values":null,"members_with_recorded_values":1,"torch_reference_conventions_with_this_digest":[],"values":[-0.903889000415802,-0.17625556886196136,0.15713146328926086,0.05303172022104263,0.8976119756698608,-1.628440260887146,-0.7621955275535583,-0.1775333285331726],"values_recorded":8,"members":[{"code_sha256_prefix":"5edffb2dfb36c8c4","path":"finetune.py","papers":["1708.02182"],"paper_pages":[{"arxiv_id":"1708.02182","page":"/paper/regularizing-and-optimizing-lstm-language"}],"arg_sig_recorded":[["data",1,"float","float32"]],"scalar_args":{"batch_size":"2"},"class_bearing":false}]}]},{"bucket":[[2,"float","float32"],[2,"float","float32"]],"n_implementations_compared":1,"n_papers":1,"n_not_digested":0,"not_digested_by_type":{},"clusters":[{"output_sha":"c27acbe9e1ee7c4a","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":[3.605257272720337,-3.0675525665283203,0.9521064758300781,-1.3817740678787231,-1.8056989908218384,-2.0609424114227295,-1.3255218267440796,1.1457784175872803,1.0377519130706787,1.9182913303375244,-0.7943991422653198,-3.1366963386535645,-0.5770701766014099,4.644217491149902,-1.0295679569244385,1.1010571718215942,0.13279087841510773,-3.245751142501831,0.8642228245735168,-0.8123717308044434,-0.04518844932317734,-2.361564874649048,-1.073736548423767,1.967348337173462,-1.5912256240844727,0.8665488362312317,1.8899555206298828,2.035036563873291,-2.451571464538574,2.548435926437378,-2.6444437503814697,-1.1882182359695435],"values_recorded":32,"members":[{"code_sha256_prefix":"caddaff3835cd018","path":"eval/eval_retrieval.py","papers":["2503.19900"],"paper_pages":[{"arxiv_id":"2503.19900","page":"/paper/cafe-unifying-representation-and-generation"}],"arg_sig_recorded":[["X",2,"float","float32"],["Y",2,"float","float32"]],"scalar_args":{"func":"<function simple_func at 0xffff49026980>","batch_size":"2","device":"'cpu'"},"class_bearing":false}]}]}]}