{"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/improved-transformer-for-high-resolution-gans","title":"Improved Transformer for High-Resolution GANs","arxiv_id":"2106.07631","date":"2021-06-14","proceeding":"NeurIPS 2021 12","authors":["Long Zhao","Zizhao Zhang","Ting Chen","Dimitris N. Metaxas","Han Zhang"],"abstract":"Attention-based models, exemplified by the Transformer, can effectively model long range dependency, but suffer from the quadratic complexity of self-attention operation, making them difficult to be adopted for high-resolution image generation based on Generative Adversarial Networks (GANs). In this paper, we introduce two key ingredients to Transformer to address this challenge. First, in low-resolution stages of the generative process, standard global self-attention is replaced with the proposed multi-axis blocked self-attention which allows efficient mixing of local and global attention. Second, in high-resolution stages, we drop self-attention while only keeping multi-layer perceptrons reminiscent of the implicit neural function. To further improve the performance, we introduce an additional self-modulation component based on cross-attention. The resulting model, denoted as HiT, has a nearly linear computational complexity with respect to the image size and thus directly scales to synthesizing high definition images. We show in the experiments that the proposed HiT achieves state-of-the-art FID scores of 30.83 and 2.95 on unconditional ImageNet $128 \\times 128$ and FFHQ $256 \\times 256$, respectively, with a reasonable throughput. We believe the proposed HiT is an important milestone for generators in GANs which are completely free of convolutions. Our code is made publicly available at https://github.com/google-research/hit-gan","url_abs":"https://arxiv.org/abs/2106.07631v3","url_pdf":"https://arxiv.org/pdf/2106.07631v3.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":"improved-transformer-for-high-resolution-gans","repo_url":"https://github.com/google-research/hit-gan","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"tf","reach":null}],"tasks":[{"task_slug":"image-generation","task_name":"Image Generation"},{"task_slug":"high","task_name":"Vocal Bursts Intensity Prediction"}],"methods":[{"method_slug":"absolute-position-encodings","method_name":"Absolute Position Encodings"},{"method_slug":"adam","method_name":"Adam"},{"method_slug":"attention","method_name":"Attention"},{"method_slug":"bpe","method_name":"BPE"},{"method_slug":"dense-connections","method_name":"Dense Connections"},{"method_slug":"dropout","method_name":"Dropout"},{"method_slug":"label-smoothing","method_name":"Label Smoothing"},{"method_slug":"layer-normalization","method_name":"Layer Normalization"},{"method_slug":"linear-layer","method_name":"Linear Layer"},{"method_slug":"multi-head-attention","method_name":"Multi-Head Attention"},{"method_slug":"position-wise-feed-forward-layer","method_name":"Position-Wise Feed-Forward Layer"},{"method_slug":"residual-connection","method_name":"Residual Connection"},{"method_slug":"softmax","method_name":"Softmax"},{"method_slug":"transformer","method_name":"Transformer"}],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/image-generation-on-celeba-256x256","task":"Image Generation","dataset":"CelebA 256x256","model":"HiT-B","rank_in_archive_order":12,"of":17,"metrics":{"FID":"3.39"},"uses_additional_data":false},{"leaderboard":"/sota/image-generation-on-celeba-hq-1024x1024","task":"Image Generation","dataset":"CelebA-HQ 1024x1024","model":"HiT-B","rank_in_archive_order":8,"of":10,"metrics":{"FID":"8.83"},"uses_additional_data":false},{"leaderboard":"/sota/image-generation-on-ffhq","task":"Image Generation","dataset":"FFHQ","model":"HiT-B","rank_in_archive_order":6,"of":12,"metrics":{"FID":"6.37"},"uses_additional_data":false},{"leaderboard":"/sota/image-generation-on-ffhq-1024-x-1024","task":"Image Generation","dataset":"FFHQ 1024 x 1024","model":"HiT-B","rank_in_archive_order":16,"of":20,"metrics":{"FID":"6.37"},"uses_additional_data":false},{"leaderboard":"/sota/image-generation-on-ffhq-256-x-256","task":"Image Generation","dataset":"FFHQ 256 x 256","model":"HiT-L","rank_in_archive_order":7,"of":51,"metrics":{"FID":"2.58"},"uses_additional_data":false},{"leaderboard":"/sota/image-generation-on-ffhq-256-x-256","task":"Image Generation","dataset":"FFHQ 256 x 256","model":"HiT-B","rank_in_archive_order":12,"of":51,"metrics":{"FID":"2.95"},"uses_additional_data":false},{"leaderboard":"/sota/image-generation-on-ffhq-256-x-256","task":"Image Generation","dataset":"FFHQ 256 x 256","model":"HiT-S","rank_in_archive_order":13,"of":51,"metrics":{"FID":"3.06"},"uses_additional_data":false},{"leaderboard":"/sota/image-generation-on-imagenet-128x128","task":"Image Generation","dataset":"ImageNet 128x128","model":"HiT","rank_in_archive_order":22,"of":23,"metrics":{"FID":"30.83"},"uses_additional_data":false}],"syntology":{"syntology_url":"https://syntology.ai/paper/2106.07631","atlas_url":"https://app.syntology.ai/?focus=2106.07631","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2106.07631"}},"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":"deterministic:regex_extraction","url":"https://github.com/google-research/hit-gan","reach":null}],"summary":{"ran":1,"ran_fixture":2,"unverified":1},"by_repo_kind":{"official":{"samples":4,"ran":3,"repositories":1}},"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":"6c10eab79338dde3","entry":"Block","repo":"google-research/hit-gan","repo_kind":"official","path":"models/generators.py","file_url":"https://github.com/google-research/hit-gan/blob/HEAD/models/generators.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"6c10eab79338dde3"}},{"code_sha256_prefix":"8826aa509040430b","entry":"block_images","repo":"google-research/hit-gan","repo_kind":"official","path":"models/generators.py","file_url":"https://github.com/google-research/hit-gan/blob/HEAD/models/generators.py","link_basis":"first_harvest_node","language":"python","status":"ran_fixture","verification_level":1,"contract_check":"RAISES","metamorphic_tier":null,"behaviour_fingerprint":true,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"8826aa509040430b"}},{"code_sha256_prefix":"5ec2d4992747d874","entry":"unblock_images","repo":"google-research/hit-gan","repo_kind":"official","path":"models/generators.py","file_url":"https://github.com/google-research/hit-gan/blob/HEAD/models/generators.py","link_basis":"first_harvest_node","language":"python","status":"ran_fixture","verification_level":1,"contract_check":"RAISES","metamorphic_tier":null,"behaviour_fingerprint":true,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"5ec2d4992747d874"}},{"code_sha256_prefix":"4130dcea05fdd401","entry":"HiTGenerator","repo":"google-research/hit-gan","repo_kind":"official","path":"models/generators.py","file_url":"https://github.com/google-research/hit-gan/blob/HEAD/models/generators.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"4130dcea05fdd401"}}]},"arxiv_metadata":null,"syntology_extracted_results":null}