{"url":"/sota/image-generation-on-textatlaseval","task":{"name":"Image Generation","url":"/task/image-generation","note":null},"dataset":{"name":"TextAtlasEval","url":"/dataset/textatlaseval"},"category":"Computer Vision","categories":["Computer Vision","Medical","Miscellaneous","Natural Language Processing"],"category_note":null,"description":"**Image Generation** (synthesis) is the task of generating new images from an existing dataset.\r\n\r\n- **Unconditional generation** refers to generating samples unconditionally from the dataset, i.e. $p(y)$\r\n- **[Conditional image generation](/task/conditional-image-generation)** (subtask) refers to generating samples conditionally from the dataset, based on a label, i.e. $p(y|x)$.\r\n\r\nIn this section, you can find state-of-the-art leaderboards for **unconditional generation**. For conditional  generation, and other types of image generations, refer to the subtasks.\r\n\r\n<span style=\"color:grey; opacity: 0.6\">( Image credit: [StyleGAN](https://github.com/NVlabs/stylegan) )</span>","description_from":"task","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","rank":"the archive's row order at snapshot; not re-ranked","rows_end_at":"2025-07-28","rows_withheld_as_spam":0,"metric_values":"the archive's strings, untouched"},"metrics":["TextVsionBlend OCR (F1 Score)","TextVisionBlend OCR (Accuracy)","TextVisionBlend OCR (Cer)","TextVisionBlend FID","TextVisionBlend Clip Score","StyledTextSynth OCR (F1 Score)","StyledTextSynth OCR (Accuracy)","StyledTextSynth OCR (Cer)","StyledTextSynth FID","StyledTextSynth Clip Score","TextScenesHQ OCR (F1 Score)","TextScenesHQ OCR (Accuracy)","TextScenesHQ OCR (Cer)","TextScenesHQ FID","TextScenesHQ Clip Score"],"metric_direction":{"note":"inferred from the metric name only (the archive records no direction); null = not inferred, chart draws points only","by_metric":{"TextVsionBlend OCR (F1 Score)":"higher","TextVisionBlend OCR (Accuracy)":"higher","TextVisionBlend OCR (Cer)":"lower","TextVisionBlend FID":"lower","TextVisionBlend Clip Score":"higher","StyledTextSynth OCR (F1 Score)":"higher","StyledTextSynth OCR (Accuracy)":"higher","StyledTextSynth OCR (Cer)":"lower","StyledTextSynth FID":"lower","StyledTextSynth Clip Score":"higher","TextScenesHQ OCR (F1 Score)":"higher","TextScenesHQ OCR (Accuracy)":"higher","TextScenesHQ OCR (Cer)":"lower","TextScenesHQ FID":"lower","TextScenesHQ Clip Score":"higher"}},"counts":{"rows":7,"rows_with_code":3,"rows_with_paper_page":4,"rows_dated":4,"rows_using_additional_data":0},"rows":[{"rank_in_archive_order":1,"model":"Grok3","metrics":{"StyledTextSynth Clip Score":"0.2938","StyledTextSynth FID":"80.33","StyledTextSynth OCR (Accuracy)":"15.82","StyledTextSynth OCR (Cer)":"0.73","StyledTextSynth OCR (F1 Score)":"21.40","TextScenesHQ Clip Score":"0.3197","TextScenesHQ FID":"-","TextScenesHQ OCR (Accuracy)":"35.07","TextScenesHQ OCR (Cer)":"0.57","TextScenesHQ OCR (F1 Score)":"37.94","TextVisionBlend Clip Score":"0.1697","TextVisionBlend FID":"-","TextVisionBlend OCR (Accuracy)":"41.54","TextVisionBlend OCR (Cer)":"0.57","TextVsionBlend OCR (F1 Score)":"44.22"},"uses_additional_data":false,"paper_date":null,"paper":null,"paper_url":null,"paper_title":"","code":null,"n_code_links":0,"syntology":null},{"rank_in_archive_order":2,"model":"SD3.5 Large","metrics":{"StyledTextSynth Clip Score":"0.2849","StyledTextSynth FID":"71.09","StyledTextSynth OCR (Accuracy)":"27.21","StyledTextSynth OCR (Cer)":"0.73","StyledTextSynth OCR (F1 Score)":"33.86","TextScenesHQ Clip Score":"0.2363","TextScenesHQ FID":"64.44","TextScenesHQ OCR (Accuracy)":"19.03","TextScenesHQ OCR (Cer)":"0.73","TextScenesHQ OCR (F1 Score)":"24.45","TextVisionBlend Clip Score":"0.1846","TextVisionBlend FID":"118.85","TextVisionBlend OCR (Accuracy)":"14.55","TextVisionBlend OCR (Cer)":"0.88","TextVsionBlend OCR (F1 Score)":"16.25"},"uses_additional_data":false,"paper_date":null,"paper":null,"paper_url":null,"paper_title":"","code":null,"n_code_links":0,"syntology":null},{"rank_in_archive_order":3,"model":"Dalle3","metrics":{"StyledTextSynth Clip Score":"0.2938","StyledTextSynth FID":"90.70","StyledTextSynth OCR (Accuracy)":"30.58","StyledTextSynth OCR (Cer)":"0.78","StyledTextSynth OCR (F1 Score)":"38.25","TextScenesHQ Clip Score":"0.3367","TextScenesHQ FID":"86.73","TextScenesHQ OCR (Accuracy)":"69.26","TextScenesHQ OCR (Cer)":"-","TextScenesHQ OCR (F1 Score)":"51.63","TextVisionBlend Clip Score":"0.1938","TextVisionBlend FID":"153.21","TextVisionBlend OCR (Accuracy)":"8.38","TextVisionBlend OCR (Cer)":"0.93","TextVsionBlend OCR (F1 Score)":"7.94"},"uses_additional_data":false,"paper_date":null,"paper":null,"paper_url":null,"paper_title":"","code":null,"n_code_links":0,"syntology":null},{"rank_in_archive_order":4,"model":"Infinity-2B","metrics":{"StyledTextSynth Clip Score":"0.2727","StyledTextSynth FID":"84.95","StyledTextSynth OCR (Accuracy)":"0.80","StyledTextSynth OCR (Cer)":"0.93","StyledTextSynth OCR (F1 Score)":"1.42","TextScenesHQ Clip Score":"0.2346","TextScenesHQ FID":"71.59","TextScenesHQ OCR (Accuracy)":"1.06","TextScenesHQ OCR (Cer)":"0.88","TextScenesHQ OCR (F1 Score)":"1.74","TextVisionBlend Clip Score":"0.1979","TextVisionBlend FID":"95.69","TextVisionBlend OCR (Accuracy)":"2.98","TextVisionBlend OCR (Cer)":"0.83","TextVsionBlend OCR (F1 Score)":"3.44"},"uses_additional_data":false,"paper_date":"2024-10-24","paper":"/paper/infinity-mm-scaling-multimodal-performance","paper_url":"https://arxiv.org/abs/2410.18558v2","paper_title":"Infinity-MM: Scaling Multimodal Performance with Large-Scale and High-Quality Instruction Data","code":"https://github.com/LLaVA-VL/LLaVA-NeXT","n_code_links":4,"syntology":{"n_ran":0,"n_unverified":6,"n_samples":6,"n_pointer_only_licence":0}},{"rank_in_archive_order":5,"model":"PixArt-Sigma","metrics":{"StyledTextSynth Clip Score":"0.2764","StyledTextSynth FID":"82.83","StyledTextSynth OCR (Accuracy)":"0.42","StyledTextSynth OCR (Cer)":"0.90","StyledTextSynth OCR (F1 Score)":"0.62","TextScenesHQ Clip Score":"0.2347","TextScenesHQ FID":"72.62","TextScenesHQ OCR (Accuracy)":"0.34","TextScenesHQ OCR (Cer)":"0.91","TextScenesHQ OCR (F1 Score)":"0.53","TextVisionBlend Clip Score":"0.1891","TextVisionBlend FID":"81.29","TextVisionBlend OCR (Accuracy)":"2.40","TextVisionBlend OCR (Cer)":"0.83","TextVsionBlend OCR (F1 Score)":"1.57"},"uses_additional_data":false,"paper_date":"2024-03-07","paper":"/paper/pixart-s-weak-to-strong-training-of-diffusion","paper_url":"https://arxiv.org/abs/2403.04692v2","paper_title":"PixArt-Σ: Weak-to-Strong Training of Diffusion Transformer for 4K Text-to-Image Generation","code":"https://github.com/PixArt-alpha/PixArt-sigma","n_code_links":2,"syntology":{"n_ran":6,"n_unverified":3,"n_samples":9,"n_pointer_only_licence":2}},{"rank_in_archive_order":6,"model":"TextDiffuser2","metrics":{"StyledTextSynth Clip Score":"0.2510","StyledTextSynth FID":"114.31","StyledTextSynth OCR (Accuracy)":"0.76","StyledTextSynth OCR (Cer)":"0.99","StyledTextSynth OCR (F1 Score)":"1.46","TextScenesHQ Clip Score":"0.2252","TextScenesHQ FID":"84.10","TextScenesHQ OCR (Accuracy)":"0.66","TextScenesHQ OCR (Cer)":"0.96","TextScenesHQ OCR (F1 Score)":"1.25","TextVisionBlend Clip Score":"-","TextVisionBlend FID":"-","TextVisionBlend OCR (Accuracy)":"-","TextVisionBlend OCR (Cer)":"-","TextVsionBlend OCR (F1 Score)":"-"},"uses_additional_data":false,"paper_date":"2023-11-28","paper":"/paper/textdiffuser-2-unleashing-the-power-of","paper_url":"https://arxiv.org/abs/2311.16465v1","paper_title":"TextDiffuser-2: Unleashing the Power of Language Models for Text Rendering","code":null,"n_code_links":0,"syntology":null},{"rank_in_archive_order":7,"model":"Anytext","metrics":{"StyledTextSynth Clip Score":"0.2501","StyledTextSynth FID":"117.71","StyledTextSynth OCR (Accuracy)":"0.35","StyledTextSynth OCR (Cer)":"0.98","StyledTextSynth OCR (F1 Score)":"0.66","TextScenesHQ Clip Score":"0.2174","TextScenesHQ FID":"101.32","TextScenesHQ OCR (Accuracy)":"0.42","TextScenesHQ OCR (Cer)":"0.95","TextScenesHQ OCR (F1 Score)":"0.8","TextVisionBlend Clip Score":"-","TextVisionBlend FID":"-","TextVisionBlend OCR (Accuracy)":"-","TextVisionBlend OCR (Cer)":"-","TextVsionBlend OCR (F1 Score)":"-"},"uses_additional_data":false,"paper_date":"2023-11-06","paper":"/paper/anytext-multilingual-visual-text-generation","paper_url":"https://arxiv.org/abs/2311.03054v5","paper_title":"AnyText: Multilingual Visual Text Generation And Editing","code":"https://github.com/tyxsspa/anytext","n_code_links":1,"syntology":{"n_ran":4,"n_unverified":15,"n_samples":19,"n_pointer_only_licence":0}}],"since_archive":{"present":false,"note":"No Syntology-extracted rows are published in this build."},"syntology":{"read_at":"2026-09-24T18:15:14+00:00","claim":"Per row: N of M harvested code samples from that row's paper executed on a synthesized fixture; the other M-N are unverified. Not a reproduction of the row's number; not a correctness claim. n_pointer_only_licence counts samples the site points at rather than redistributes (a licence axis, independent of ran/unverified).","rows_with_graph_line":3,"rows_with_any_sample_ran":2,"distinct_papers_with_graph_line":3,"distinct_papers_with_any_sample_ran":2,"samples_over_distinct_papers":{"n_ran":10,"n_unverified":24,"n_samples":34,"n_pointer_only_licence":2,"note":"each paper (arXiv id) counted once, however many rows it is behind; this is the page-level figure"},"samples_row_weighted":{"n_ran":10,"n_unverified":24,"n_samples":34,"n_pointer_only_licence":2,"note":"row-weighted: a paper behind several rows is counted once per row; inflated relative to samples_over_distinct_papers by design, kept for readers summing the per-row syntology blocks"}}}