{"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/msg-gan-multi-scale-gradients-gan-for-more","title":"MSG-GAN: Multi-Scale Gradients for Generative Adversarial Networks","arxiv_id":"1903.06048","date":"2019-03-14","proceeding":"CVPR 2020 6","authors":["Animesh Karnewar","Oliver Wang"],"abstract":"While Generative Adversarial Networks (GANs) have seen huge successes in image synthesis tasks, they are notoriously difficult to adapt to different datasets, in part due to instability during training and sensitivity to hyperparameters. One commonly accepted reason for this instability is that gradients passing from the discriminator to the generator become uninformative when there isn't enough overlap in the supports of the real and fake distributions. In this work, we propose the Multi-Scale Gradient Generative Adversarial Network (MSG-GAN), a simple but effective technique for addressing this by allowing the flow of gradients from the discriminator to the generator at multiple scales. This technique provides a stable approach for high resolution image synthesis, and serves as an alternative to the commonly used progressive growing technique. We show that MSG-GAN converges stably on a variety of image datasets of different sizes, resolutions and domains, as well as different types of loss functions and architectures, all with the same set of fixed hyperparameters. When compared to state-of-the-art GANs, our approach matches or exceeds the performance in most of the cases we tried.","url_abs":"https://arxiv.org/abs/1903.06048v4","url_pdf":"https://arxiv.org/pdf/1903.06048v4.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":"msg-gan-multi-scale-gradients-gan-for-more","repo_url":"https://github.com/akanimax/msg-stylegan-tf","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"tf","reach":{"status":"ok","spdx":"NOASSERTION"}},{"paper_slug":"msg-gan-multi-scale-gradients-gan-for-more","repo_url":"https://github.com/Iceland-Leo/Style-MSG-GAN","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok"}},{"paper_slug":"msg-gan-multi-scale-gradients-gan-for-more","repo_url":"https://github.com/akanimax/BMSG-GAN","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok","spdx":"MIT"}},{"paper_slug":"msg-gan-multi-scale-gradients-gan-for-more","repo_url":"https://github.com/denizalperacar/MSG-GAN","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok","spdx":"MIT"}},{"paper_slug":"msg-gan-multi-scale-gradients-gan-for-more","repo_url":"https://github.com/manicman1999/StyleGAN-Tensorflow-2.0","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":null}],"tasks":[{"task_slug":null,"task_name":"Generative Adversarial Network"},{"task_slug":"image-generation","task_name":"Image Generation"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/image-generation-on-celeba-hq-1024x1024","task":"Image Generation","dataset":"CelebA-HQ 1024x1024","model":"MSG-StyleGAN","rank_in_archive_order":5,"of":10,"metrics":{"FID":"6.37"},"uses_additional_data":false},{"leaderboard":"/sota/image-generation-on-ffhq","task":"Image Generation","dataset":"FFHQ","model":"MSG-StyleGAN","rank_in_archive_order":8,"of":12,"metrics":{"Clean-FID (70k)":"6.51"},"uses_additional_data":false},{"leaderboard":"/sota/image-generation-on-ffhq-1024-x-1024","task":"Image Generation","dataset":"FFHQ 1024 x 1024","model":"MSG-StyleGAN","rank_in_archive_order":15,"of":20,"metrics":{"FID":"5.8"},"uses_additional_data":false},{"leaderboard":"/sota/image-generation-on-indian-celebs-256-x-256","task":"Image Generation","dataset":"Indian Celebs 256 x 256","model":"MSG-StyleGAN","rank_in_archive_order":1,"of":1,"metrics":{"FID":"28.44"},"uses_additional_data":false},{"leaderboard":"/sota/image-generation-on-lsun-churches-256-x-256","task":"Image Generation","dataset":"LSUN Churches 256 x 256","model":"MSG-StyleGAN","rank_in_archive_order":18,"of":27,"metrics":{"Clean-FID (trainfull)":"5.38 ± 0.03","FID":"5.2"},"uses_additional_data":false},{"leaderboard":"/sota/image-generation-on-oxford-102-flowers-256-x","task":"Image Generation","dataset":"Oxford 102 Flowers 256 x 256","model":"MSG-StyleGAN","rank_in_archive_order":2,"of":2,"metrics":{"FID":"19.60"},"uses_additional_data":false}],"syntology":{"syntology_url":null,"atlas_url":"https://app.syntology.ai/?focus=1903.06048","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"1903.06048"}},"developers":"https://syntology.ai/developers","read_at":"2026-09-24T18:15:14+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/akanimax/msg-stylegan-tf","reach":{"status":"ok","spdx":"NOASSERTION"}},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/manicman1999/StyleGAN-Tensorflow-2.0","reach":null},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/denizalperacar/MSG-GAN","reach":{"status":"ok","spdx":"MIT"}},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/akanimax/BMSG-GAN","reach":{"status":"ok","spdx":"MIT"}},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/Iceland-Leo/Style-MSG-GAN","reach":{"status":"ok"}}],"summary":{"ran_honours":3,"unverified":12},"by_repo_kind":{"listed":{"samples":15,"ran":3,"repositories":3}},"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":0,"samples":[{"code_sha256_prefix":"e443d34b4c89cb5e","entry":"mixedList","repo":"manicman1999/StyleGAN-Tensorflow-2.0","repo_kind":"listed","path":"stylegan.py","file_url":"https://github.com/manicman1999/StyleGAN-Tensorflow-2.0/blob/HEAD/stylegan.py","link_basis":"first_harvest_node","language":"python","status":"ran_honours","verification_level":1,"contract_check":"HONOURS","metamorphic_tier":"well_formed","behaviour_fingerprint":true,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"e443d34b4c89cb5e"}},{"code_sha256_prefix":"8af2f60798b76e8c","entry":"noise","repo":"manicman1999/StyleGAN-Tensorflow-2.0","repo_kind":"listed","path":"stylegan.py","file_url":"https://github.com/manicman1999/StyleGAN-Tensorflow-2.0/blob/HEAD/stylegan.py","link_basis":"first_harvest_node","language":"python","status":"ran_honours","verification_level":1,"contract_check":"HONOURS","metamorphic_tier":"well_formed","behaviour_fingerprint":true,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"8af2f60798b76e8c"}},{"code_sha256_prefix":"84deaa80c4d54b08","entry":"noiseList","repo":"manicman1999/StyleGAN-Tensorflow-2.0","repo_kind":"listed","path":"stylegan.py","file_url":"https://github.com/manicman1999/StyleGAN-Tensorflow-2.0/blob/HEAD/stylegan.py","link_basis":"first_harvest_node","language":"python","status":"ran_honours","verification_level":1,"contract_check":"HONOURS","metamorphic_tier":"well_formed","behaviour_fingerprint":true,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"84deaa80c4d54b08"}},{"code_sha256_prefix":"b4c8e1e1917daee0","entry":"WGANGP_loss","repo":"denizalperacar/MSG-GAN","repo_kind":"listed","path":"src/loss.py","file_url":"https://github.com/denizalperacar/MSG-GAN/blob/HEAD/src/loss.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"b4c8e1e1917daee0"}},{"code_sha256_prefix":"bd6383ad21a9d1b9","entry":"adjust_dynamic_range","repo":"akanimax/BMSG-GAN","repo_kind":"listed","path":"sourcecode/generate_multi_scale_samples.py","file_url":"https://github.com/akanimax/BMSG-GAN/blob/HEAD/sourcecode/generate_multi_scale_samples.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"bd6383ad21a9d1b9"}},{"code_sha256_prefix":"bbf23ae0401556fa","entry":"adjust_dynamic_range","repo":"akanimax/BMSG-GAN","repo_kind":"listed","path":"sourcecode/latent_space_interpolation.py","file_url":"https://github.com/akanimax/BMSG-GAN/blob/HEAD/sourcecode/latent_space_interpolation.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"bbf23ae0401556fa"}},{"code_sha256_prefix":"fae02fd84bdd8461","entry":"filter_func","repo":"denizalperacar/MSG-GAN","repo_kind":"listed","path":"src/data_loader.py","file_url":"https://github.com/denizalperacar/MSG-GAN/blob/HEAD/src/data_loader.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"fae02fd84bdd8461"}},{"code_sha256_prefix":"4c285e4ad37fd444","entry":"get_data_loader","repo":"akanimax/BMSG-GAN","repo_kind":"listed","path":"sourcecode/data_processing/DataLoader.py","file_url":"https://github.com/akanimax/BMSG-GAN/blob/HEAD/sourcecode/data_processing/DataLoader.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"4c285e4ad37fd444"}},{"code_sha256_prefix":"ec51937c7b4fbac6","entry":"get_latent_variable","repo":"denizalperacar/MSG-GAN","repo_kind":"listed","path":"src/custom.py","file_url":"https://github.com/denizalperacar/MSG-GAN/blob/HEAD/src/custom.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"ec51937c7b4fbac6"}},{"code_sha256_prefix":"d52b4c878ccdf30d","entry":"get_transform","repo":"akanimax/BMSG-GAN","repo_kind":"listed","path":"sourcecode/data_processing/DataLoader.py","file_url":"https://github.com/akanimax/BMSG-GAN/blob/HEAD/sourcecode/data_processing/DataLoader.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"d52b4c878ccdf30d"}},{"code_sha256_prefix":"4901d39068cad530","entry":"gradient_penalty_loss","repo":"denizalperacar/MSG-GAN","repo_kind":"listed","path":"src/loss.py","file_url":"https://github.com/denizalperacar/MSG-GAN/blob/HEAD/src/loss.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"4901d39068cad530"}},{"code_sha256_prefix":"e31aa75b8b0cc7c9","entry":"minbatchstd","repo":"denizalperacar/MSG-GAN","repo_kind":"listed","path":"src/custom.py","file_url":"https://github.com/denizalperacar/MSG-GAN/blob/HEAD/src/custom.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"e31aa75b8b0cc7c9"}},{"code_sha256_prefix":"968fdef6839bc97b","entry":"pixel_norm","repo":"denizalperacar/MSG-GAN","repo_kind":"listed","path":"src/custom.py","file_url":"https://github.com/denizalperacar/MSG-GAN/blob/HEAD/src/custom.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"968fdef6839bc97b"}},{"code_sha256_prefix":"56b98fbe693c2e6f","entry":"progressive_upscaling","repo":"akanimax/BMSG-GAN","repo_kind":"listed","path":"sourcecode/generate_multi_scale_samples.py","file_url":"https://github.com/akanimax/BMSG-GAN/blob/HEAD/sourcecode/generate_multi_scale_samples.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"56b98fbe693c2e6f"}},{"code_sha256_prefix":"d5151aa5668d7556","entry":"read_loss_log","repo":"akanimax/BMSG-GAN","repo_kind":"listed","path":"sourcecode/generate_loss_plots.py","file_url":"https://github.com/akanimax/BMSG-GAN/blob/HEAD/sourcecode/generate_loss_plots.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"d5151aa5668d7556"}}]},"arxiv_metadata":null,"syntology_extracted_results":null}