{"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":"/paper/categorical-reparameterization-with-gumbel","title":"Categorical Reparameterization with Gumbel-Softmax","arxiv_id":"1611.01144","date":"2016-11-03","proceeding":null,"authors":["Eric Jang","Shixiang Gu","Ben Poole"],"abstract":"Categorical variables are a natural choice for representing discrete\nstructure in the world. However, stochastic neural networks rarely use\ncategorical latent variables due to the inability to backpropagate through\nsamples. In this work, we present an efficient gradient estimator that replaces\nthe non-differentiable sample from a categorical distribution with a\ndifferentiable sample from a novel Gumbel-Softmax distribution. This\ndistribution has the essential property that it can be smoothly annealed into a\ncategorical distribution. We show that our Gumbel-Softmax estimator outperforms\nstate-of-the-art gradient estimators on structured output prediction and\nunsupervised generative modeling tasks with categorical latent variables, and\nenables large speedups on semi-supervised classification.","url_abs":"http://arxiv.org/abs/1611.01144v5","url_pdf":"http://arxiv.org/pdf/1611.01144v5.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"categorical-reparameterization-with-gumbel","repo_url":"https://github.com/AntixK/PyTorch-VAE","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok","spdx":"Apache-2.0"}},{"paper_slug":"categorical-reparameterization-with-gumbel","repo_url":"https://github.com/EddieCunningham/GraphLSSM","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":null},{"paper_slug":"categorical-reparameterization-with-gumbel","repo_url":"https://github.com/EdoardoBotta/Gaussian-Mixture-VAE","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null},{"paper_slug":"categorical-reparameterization-with-gumbel","repo_url":"https://github.com/GuyLor/direct_vae","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null},{"paper_slug":"categorical-reparameterization-with-gumbel","repo_url":"https://github.com/Jmkernes/PAR-Transformer-XL","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":null},{"paper_slug":"categorical-reparameterization-with-gumbel","repo_url":"https://github.com/OlivierAlgoet/Tensorflow2-GMM","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":null},{"paper_slug":"categorical-reparameterization-with-gumbel","repo_url":"https://github.com/caravagn/GDA","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":{"status":"unanswered"}},{"paper_slug":"categorical-reparameterization-with-gumbel","repo_url":"https://github.com/chijames/gst","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null},{"paper_slug":"categorical-reparameterization-with-gumbel","repo_url":"https://github.com/crowdflowTUe/stampnet","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"ok","spdx":"BSD-3-Clause"}},{"paper_slug":"categorical-reparameterization-with-gumbel","repo_url":"https://github.com/ericjang/gumbel-softmax","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":{"status":"unanswered"}},{"paper_slug":"categorical-reparameterization-with-gumbel","repo_url":"https://github.com/hellolzc/SBN_gumbel","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null},{"paper_slug":"categorical-reparameterization-with-gumbel","repo_url":"https://github.com/jariasf/gmvae","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null},{"paper_slug":"categorical-reparameterization-with-gumbel","repo_url":"https://github.com/kampta/pytorch-distributions","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null},{"paper_slug":"categorical-reparameterization-with-gumbel","repo_url":"https://github.com/karpathy/deep-vector-quantization","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null},{"paper_slug":"categorical-reparameterization-with-gumbel","repo_url":"https://github.com/shaabhishek/gumbel-softmax-pytorch","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"unanswered"}},{"paper_slug":"categorical-reparameterization-with-gumbel","repo_url":"https://github.com/stefanthaler/dl-papers-imlemented","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"unanswered"}},{"paper_slug":"categorical-reparameterization-with-gumbel","repo_url":"https://github.com/stefanthaler/gumbel-softmax-exploration","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"unanswered"}},{"paper_slug":"categorical-reparameterization-with-gumbel","repo_url":"https://github.com/tensorflow/models","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":null},{"paper_slug":"categorical-reparameterization-with-gumbel","repo_url":"https://github.com/tensorflow/models/tree/master/research/rebar","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":null},{"paper_slug":"categorical-reparameterization-with-gumbel","repo_url":"https://github.com/lvyufeng/gumbel_softmax_mindspore","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"mindspore","reach":{"status":"unanswered"}}],"tasks":[{"task_slug":"classification","task_name":"General Classification"}],"methods":[{"method_slug":"gumbel-softmax","method_name":"Gumbel Softmax"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":"https://syntology.ai/paper/1611.01144","atlas_url":"https://app.syntology.ai/?focus=1611.01144","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"1611.01144"}},"developers":"https://syntology.ai/developers","read_at":"2026-09-25T09:33:49+00:00","read_at_is":"when the build read Syntology's graph, not when any sample ran","claim":"Per-sample execution status on synthesized fixtures; not a correctness claim about the paper. Samples come from repositories linked to the paper, official or community; repo_kind says which.","repos":[{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/tensorflow/models","reach":null},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/AntixK/PyTorch-VAE","reach":{"status":"ok","spdx":"Apache-2.0"}},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/caravagn/GDA","reach":{"status":"unanswered"}},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/shaabhishek/gumbel-softmax-pytorch","reach":{"status":"unanswered"}},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/stefanthaler/gumbel-softmax-exploration","reach":{"status":"unanswered"}},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/lvyufeng/gumbel_softmax_mindspore","reach":{"status":"unanswered"}},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/OlivierAlgoet/Tensorflow2-GMM","reach":null},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/crowdflowTUe/stampnet","reach":{"status":"ok","spdx":"BSD-3-Clause"}},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/Jmkernes/PAR-Transformer-XL","reach":null},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/tensorflow/models/tree/master/research/rebar","reach":null},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/ericjang/gumbel-softmax","reach":{"status":"unanswered"}},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/jariasf/gmvae","reach":null},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/hellolzc/SBN_gumbel","reach":null},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/GuyLor/direct_vae","reach":null},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/chijames/gst","reach":null},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/EddieCunningham/GraphLSSM","reach":null},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/karpathy/deep-vector-quantization","reach":null},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/stefanthaler/dl-papers-imlemented","reach":{"status":"unanswered"}},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/kampta/pytorch-distributions","reach":null},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/EdoardoBotta/Gaussian-Mixture-VAE","reach":null}],"summary":{"ran":4,"ran_draft_wrong":7,"ran_fixture":1,"unverified":2},"by_repo_kind":{"listed":{"samples":14,"ran":12,"repositories":8}},"repo_kind_vocabulary":{"official":"The archive marks this repository official for the paper","named_in_paper":"The archive records that the paper mentions this repository; it is not marked official","listed":"In the archive's code links for this paper, not marked official and not recorded as mentioned in the paper","found_in_text":"Syntology found this repository in the paper's own text; whether it is the authors' implementation is not asserted","community":"Not in the archive's code links for this paper; a community repository Syntology harvested"},"n_pointer_only_for_licence":8,"samples":[{"code_sha256_prefix":"dd897049a427c55d","entry":"GMM","repo":"OlivierAlgoet/Tensorflow2-GMM","repo_kind":"listed","path":"src/GMM.py","file_url":"https://github.com/OlivierAlgoet/Tensorflow2-GMM/blob/HEAD/src/GMM.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"dd897049a427c55d"}},{"code_sha256_prefix":"59b2d207600a31bf","entry":"GumbelQuantize","repo":"karpathy/deep-vector-quantization","repo_kind":"listed","path":"dvq/model/quantize.py","file_url":"https://github.com/karpathy/deep-vector-quantization/blob/HEAD/dvq/model/quantize.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":"deterministic","behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"59b2d207600a31bf"}},{"code_sha256_prefix":"3181ff255bf198ec","entry":"GumbelSoftmax","repo":"jariasf/gmvae","repo_kind":"listed","path":"pytorch/networks/Layers.py","file_url":"https://github.com/jariasf/gmvae/blob/HEAD/pytorch/networks/Layers.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":"deterministic","behaviour_fingerprint":true,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"3181ff255bf198ec"}},{"code_sha256_prefix":"e87691ecef7118db","entry":"VAE","repo":"kampta/pytorch-distributions","repo_kind":"listed","path":"concrete_vae.py","file_url":"https://github.com/kampta/pytorch-distributions/blob/HEAD/concrete_vae.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"e87691ecef7118db"}},{"code_sha256_prefix":"a41fecd8dcdc493a","entry":"_gumbel_softmax_sample","repo":"hellolzc/SBN_gumbel","repo_kind":"listed","path":"gumbel_sigmoid_softmax.py","file_url":"https://github.com/hellolzc/SBN_gumbel/blob/HEAD/gumbel_sigmoid_softmax.py","link_basis":"first_harvest_node","language":"python","status":"ran_draft_wrong","verification_level":1,"contract_check":"MISDECLARED","metamorphic_tier":null,"behaviour_fingerprint":true,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"a41fecd8dcdc493a"}},{"code_sha256_prefix":"82f3a84259b22b4c","entry":"_sample_gumbel","repo":"hellolzc/SBN_gumbel","repo_kind":"listed","path":"gumbel_sigmoid_softmax.py","file_url":"https://github.com/hellolzc/SBN_gumbel/blob/HEAD/gumbel_sigmoid_softmax.py","link_basis":"first_harvest_node","language":"python","status":"ran_draft_wrong","verification_level":1,"contract_check":"MISDECLARED","metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"82f3a84259b22b4c"}},{"code_sha256_prefix":"ddba371fd54afdd0","entry":"convert_to_one_hot","repo":"chijames/gst","repo_kind":"listed","path":"model/basic.py","file_url":"https://github.com/chijames/gst/blob/HEAD/model/basic.py","link_basis":"first_harvest_node","language":"python","status":"ran_draft_wrong","verification_level":1,"contract_check":"MISDECLARED","metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"ddba371fd54afdd0"}},{"code_sha256_prefix":"ae5d3a400d0e7968","entry":"gumbel_softmax","repo":"hellolzc/SBN_gumbel","repo_kind":"listed","path":"gumbel_sigmoid_softmax.py","file_url":"https://github.com/hellolzc/SBN_gumbel/blob/HEAD/gumbel_sigmoid_softmax.py","link_basis":"first_harvest_node","language":"python","status":"ran_fixture","verification_level":1,"contract_check":"RAISES","metamorphic_tier":null,"behaviour_fingerprint":true,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"ae5d3a400d0e7968"}},{"code_sha256_prefix":"d069f664a7cf326c","entry":"gumbel_softmax_sample","repo":"EdoardoBotta/Gaussian-Mixture-VAE","repo_kind":"listed","path":"distributions/gumbel.py","file_url":"https://github.com/EdoardoBotta/Gaussian-Mixture-VAE/blob/HEAD/distributions/gumbel.py","link_basis":"first_harvest_node","language":"python","status":"ran_draft_wrong","verification_level":1,"contract_check":"MISDECLARED","metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"d069f664a7cf326c"}},{"code_sha256_prefix":"c8b051f5575732d3","entry":"masked_softmax","repo":"chijames/gst","repo_kind":"listed","path":"model/basic.py","file_url":"https://github.com/chijames/gst/blob/HEAD/model/basic.py","link_basis":"first_harvest_node","language":"python","status":"ran_draft_wrong","verification_level":1,"contract_check":"MISDECLARED","metamorphic_tier":null,"behaviour_fingerprint":true,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"c8b051f5575732d3"}},{"code_sha256_prefix":"0df9e0f36c21548a","entry":"sample_gumbel","repo":"EdoardoBotta/Gaussian-Mixture-VAE","repo_kind":"listed","path":"distributions/gumbel.py","file_url":"https://github.com/EdoardoBotta/Gaussian-Mixture-VAE/blob/HEAD/distributions/gumbel.py","link_basis":"first_harvest_node","language":"python","status":"ran_draft_wrong","verification_level":1,"contract_check":"MISDECLARED","metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"0df9e0f36c21548a"}},{"code_sha256_prefix":"878b0d58fc1d164a","entry":"st_gumbel_softmax","repo":"chijames/gst","repo_kind":"listed","path":"model/basic.py","file_url":"https://github.com/chijames/gst/blob/HEAD/model/basic.py","link_basis":"first_harvest_node","language":"python","status":"ran_draft_wrong","verification_level":1,"contract_check":"MISDECLARED","metamorphic_tier":"invariant","behaviour_fingerprint":true,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"878b0d58fc1d164a"}},{"code_sha256_prefix":"97c545c491659a35","entry":"GSM_VAE","repo":"GuyLor/direct_vae","repo_kind":"listed","path":"discrete_vae/GSM.py","file_url":"https://github.com/GuyLor/direct_vae/blob/HEAD/discrete_vae/GSM.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"97c545c491659a35"}},{"code_sha256_prefix":"1e08ae9c81b34087","entry":"VAE","repo":"GuyLor/direct_vae","repo_kind":"listed","path":"discrete_vae/GSM.py","file_url":"https://github.com/GuyLor/direct_vae/blob/HEAD/discrete_vae/GSM.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"1e08ae9c81b34087"}}]},"arxiv_metadata":null,"syntology_extracted_results":null}