{"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/soft-cross-entropy","entry":"soft_cross_entropy","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":8,"n_buckets":2,"n_distinct_outputs":3,"n_class_bearing":0,"not_compared":{"not_run_on_shared_input":{"n":1,"by_error":{"AssertionError":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/soft-cross-entropy","buckets":[{"bucket":[[2,"float","float32"],[2,"float","float32"]],"n_implementations_compared":5,"n_papers":6,"n_not_digested":0,"not_digested_by_type":{},"clusters":[{"output_sha":"ba10c57de7deee16","size":3,"n_papers":4,"shape":[],"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":[-9.508142471313477],"values_recorded":1,"members":[{"code_sha256_prefix":"49e2f0c554d81ea3","path":"lightning_subset/module.py","papers":["2203.12023","2312.10366"],"paper_pages":[{"arxiv_id":"2203.12023","page":"/paper/generative-modeling-helps-weak-supervision"},{"arxiv_id":"2312.10366","page":"/paper/fusing-conditional-submodular-gan-and"}],"arg_sig_recorded":[["pred",2,"float","float32"],["soft_targets",2,"float","float32"]],"scalar_args":{},"class_bearing":false},{"code_sha256_prefix":"7b40608d582dc51b","path":"label_distillation_or2.py","papers":["2006.08572"],"paper_pages":[{"arxiv_id":"2006.08572","page":"/paper/flexible-dataset-distillation-learn-labels"}],"arg_sig_recorded":[["pred",2,"float","float32"],["soft_targets",2,"float","float32"]],"scalar_args":{},"class_bearing":false},{"code_sha256_prefix":"cef161c1e3365a6a","path":"validate/train_fkd.py","papers":["2512.15647"],"paper_pages":[{"arxiv_id":"2512.15647","page":"/paper/arxiv-2512-15647"}],"arg_sig_recorded":[["logits",2,"float","float32"],["soft_targets",2,"float","float32"]],"scalar_args":{},"class_bearing":false}]},{"output_sha":"7e5824d63f9fbed5","size":2,"n_papers":2,"shape":[],"dtype":"float32","type":"Tensor","finite":true,"max_difference_among_recorded_values":0.0,"members_with_recorded_values":2,"torch_reference_conventions_with_this_digest":[],"values":[1.6227259635925293],"values_recorded":1,"members":[{"code_sha256_prefix":"32fee35165e8a309","path":"DynaBERT/run_glue.py","papers":["2004.04037"],"paper_pages":[{"arxiv_id":"2004.04037","page":"/paper/dynabert-dynamic-bert-with-adaptive-width-and"}],"arg_sig_recorded":[["predicts",2,"float","float32"],["targets",2,"float","float32"]],"scalar_args":{},"class_bearing":false},{"code_sha256_prefix":"5ea641b9e771d680","path":"minima/run_distillation_llama_ds.py","papers":["2311.07052"],"paper_pages":[{"arxiv_id":"2311.07052","page":"/paper/towards-the-law-of-capacity-gap-in-distilling"}],"arg_sig_recorded":[["input",2,"float","float32"],["target",2,"float","float32"]],"scalar_args":{"reduction":"'mean'"},"class_bearing":false}]}]},{"bucket":[[2,"float","float32"],[2,"float","float32"],[1,"float","float32"]],"n_implementations_compared":1,"n_papers":2,"n_not_digested":0,"not_digested_by_type":{},"clusters":[{"output_sha":"17b520c2acabfe0e","size":1,"n_papers":2,"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":[6.136525630950928],"values_recorded":1,"members":[{"code_sha256_prefix":"61492b343324d2ea","path":"scood/trainers/udg_trainer.py","papers":["2108.11941","2411.10794"],"paper_pages":[{"arxiv_id":"2108.11941","page":"/paper/semantically-coherent-out-of-distribution"},{"arxiv_id":"2411.10794","page":"/paper/going-beyond-conventional-ood-detection"}],"arg_sig_recorded":[["logit",2,"float","float32"],["label",2,"float","float32"],["weight",1,"float","float32"]],"scalar_args":{"reduce":"None","reduction":"'mean'"},"class_bearing":false}]}]}]}