{"id":628796,"date":"2023-04-13T09:49:03","date_gmt":"2023-04-13T14:49:03","guid":{"rendered":"https:\/\/news.sellorbuyhomefast.com\/index.php\/2023\/04\/13\/openai-releases-consistency-model-for-one-step-generation\/"},"modified":"2023-04-13T09:49:03","modified_gmt":"2023-04-13T14:49:03","slug":"openai-releases-consistency-model-for-one-step-generation","status":"publish","type":"post","link":"https:\/\/newsycanuse.com\/index.php\/2023\/04\/13\/openai-releases-consistency-model-for-one-step-generation\/","title":{"rendered":"OpenAI releases Consistency Model for one-step generation"},"content":{"rendered":"<div data-target=\"readme-toc.content\">\n<article itemprop=\"text\">\n<h2 tabindex=\"-1\" dir=\"auto\">Consistency Models<\/h2>\n<p dir=\"auto\">This repository contains the codebase for <a href=\"https:\/\/arxiv.org\/abs\/2303.01469\" rel=\"nofollow\">Consistency Models<\/a>, implemented using PyTorch for conducting large-scale experiments on ImageNet-64, LSUN Bedroom-256, and LSUN Cat-256. We have based our repository on <a href=\"https:\/\/github.com\/openai\/guided-diffusion\">openai\/guided-diffusion<\/a>, which was initially released under the MIT license. Our modifications have enabled support for consistency distillation, consistency training, as well as several sampling and editing algorithms discussed in the paper.<\/p>\n<p dir=\"auto\">The repository for CIFAR-10 experiments is in JAX and will be released separately.<\/p>\n<h2 tabindex=\"-1\" dir=\"auto\">Pre-trained models<\/h2>\n<p dir=\"auto\">We have released checkpoints for the main models in the paper. Before using these models, please review the corresponding <a href=\"http:\/\/github.com\/openai\/consistency_models\/blob\/main\/model-card.md\">model card<\/a> to understand the intended use and limitations of these models.<\/p>\n<p dir=\"auto\">Here are the download links for each model checkpoint:<\/p>\n<ul dir=\"auto\">\n<li>EDM on ImageNet-64: <a href=\"https:\/\/openaipublic.blob.core.windows.net\/consistency\/edm_imagenet64_ema.pt\" rel=\"nofollow\">edm_imagenet64_ema.pt<\/a><\/li>\n<li>CD on ImageNet-64 with l2 metric: <a href=\"https:\/\/openaipublic.blob.core.windows.net\/consistency\/cd_imagenet64_l2.pt\" rel=\"nofollow\">cd_imagenet64_l2.pt<\/a><\/li>\n<li>CD on ImageNet-64 with LPIPS metric: <a href=\"https:\/\/openaipublic.blob.core.windows.net\/consistency\/cd_imagenet64_lpips.pt\" rel=\"nofollow\">cd_imagenet64_lpips.pt<\/a><\/li>\n<li>CT on ImageNet-64: <a href=\"https:\/\/openaipublic.blob.core.windows.net\/consistency\/ct_imagenet64.pt\" rel=\"nofollow\">ct_imagenet64.pt<\/a><\/li>\n<li>EDM on LSUN Bedroom-256: <a href=\"https:\/\/openaipublic.blob.core.windows.net\/consistency\/edm_bedroom256_ema.pt\" rel=\"nofollow\">edm_bedroom256_ema.pt<\/a><\/li>\n<li>CD on LSUN Bedroom-256 with l2 metric: <a href=\"https:\/\/openaipublic.blob.core.windows.net\/consistency\/cd_bedroom256_l2.pt\" rel=\"nofollow\">cd_bedroom256_l2.pt<\/a><\/li>\n<li>CD on LSUN Bedroom-256 with LPIPS metric: <a href=\"https:\/\/openaipublic.blob.core.windows.net\/consistency\/cd_bedroom256_lpips.pt\" rel=\"nofollow\">cd_bedroom256_lpips.pt<\/a><\/li>\n<li>CT on LSUN Bedroom-256: <a href=\"https:\/\/openaipublic.blob.core.windows.net\/consistency\/ct_bedroom256.pt\" rel=\"nofollow\">ct_bedroom256.pt<\/a><\/li>\n<li>EDM on LSUN Cat-256: <a href=\"https:\/\/openaipublic.blob.core.windows.net\/consistency\/edm_cat256_ema.pt\" rel=\"nofollow\">edm_cat256_ema.pt<\/a><\/li>\n<li>CD on LSUN Cat-256 with l2 metric: <a href=\"https:\/\/openaipublic.blob.core.windows.net\/consistency\/cd_cat256_l2.pt\" rel=\"nofollow\">cd_cat256_l2.pt<\/a><\/li>\n<li>CD on LSUN Cat-256 with LPIPS metric: <a href=\"https:\/\/openaipublic.blob.core.windows.net\/consistency\/cd_cat256_lpips.pt\" rel=\"nofollow\">cd_cat256_lpips.pt<\/a><\/li>\n<li>CT on LSUN Cat-256: <a href=\"https:\/\/openaipublic.blob.core.windows.net\/consistency\/ct_cat256.pt\" rel=\"nofollow\">ct_cat256.pt<\/a><\/li>\n<\/ul>\n<h2 tabindex=\"-1\" dir=\"auto\">Dependencies<\/h2>\n<p dir=\"auto\">To install all packages in this codebase along with their dependencies, run<\/p>\n<h2 tabindex=\"-1\" dir=\"auto\">Model training and sampling<\/h2>\n<p dir=\"auto\">We provide examples of EDM training, consistency distillation, consistency training, single-step generation, and multistep generation in <a href=\"http:\/\/github.com\/openai\/consistency_models\/blob\/main\/scripts\/launch.sh\">cm\/scripts\/launch.sh<\/a>.<\/p>\n<h2 tabindex=\"-1\" dir=\"auto\">Evaluations<\/h2>\n<p dir=\"auto\">To compare different generative models, we use FID, Precision, Recall, and Inception Score. These metrics can all be calculated using batches of samples stored in <code>.npz<\/code> (numpy) files. One can evaluate samples with <a href=\"http:\/\/github.com\/openai\/consistency_models\/blob\/main\/evaluations\/evaluator.py\">cm\/evaluations\/evaluator.py<\/a> in the same way as described in <a href=\"https:\/\/github.com\/openai\/guided-diffusion\">openai\/guided-diffusion<\/a>, with reference dataset batches provided therein.<\/p>\n<h2 tabindex=\"-1\" dir=\"auto\">Citation<\/h2>\n<p dir=\"auto\">If you find this method and\/or code useful, please consider citing<\/p>\n<div dir=\"auto\" data-snippet-clipboard-copy-content=\"@article{song2023consistency,\n  title={Consistency Models},\n  author={Song, Yang and Dhariwal, Prafulla and Chen, Mark and Sutskever, Ilya},\n  journal={arXiv preprint arXiv:2303.01469},\n  year={2023},\n}\"><\/p>\n<pre><span>@article<\/span>{<span>song2023consistency<\/span>,\n  <span>title<\/span>=<span><span>{<\/span>Consistency Models<span>}<\/span><\/span>,\n  <span>author<\/span>=<span><span>{<\/span>Song, Yang and Dhariwal, Prafulla and Chen, Mark and Sutskever, Ilya<span>}<\/span><\/span>,\n  <span>journal<\/span>=<span><span>{<\/span>arXiv preprint arXiv:2303.01469<span>}<\/span><\/span>,\n  <span>year<\/span>=<span><span>{<\/span>2023<span>}<\/span><\/span>,\n}<\/pre>\n<\/div><\/div>\n<p><a href=\"https:\/\/github.com\/openai\/consistency_models\" class=\"button purchase\" rel=\"nofollow noopener\" target=\"_blank\">Read More<\/a><br \/>\n Yuri Volkman<\/p>\n","protected":false},"excerpt":{"rendered":"<p>Consistency Models This repository contains the codebase for Consistency Models, implemented using PyTorch for conducting large-scale experiments on ImageNet-64, LSUN Bedroom-256, and LSUN Cat-256. We have based our repository on openai\/guided-diffusion, which was initially released under the MIT license. Our modifications have enabled support for consistency distillation, consistency training, as well as several sampling and<\/p>\n","protected":false},"author":1,"featured_media":628797,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[22088,1492,46],"tags":[],"class_list":{"0":"post-628796","1":"post","2":"type-post","3":"status-publish","4":"format-standard","5":"has-post-thumbnail","7":"category-openai","8":"category-releases","9":"category-technology"},"aioseo_notices":[],"_links":{"self":[{"href":"https:\/\/newsycanuse.com\/index.php\/wp-json\/wp\/v2\/posts\/628796","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/newsycanuse.com\/index.php\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/newsycanuse.com\/index.php\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/newsycanuse.com\/index.php\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/newsycanuse.com\/index.php\/wp-json\/wp\/v2\/comments?post=628796"}],"version-history":[{"count":0,"href":"https:\/\/newsycanuse.com\/index.php\/wp-json\/wp\/v2\/posts\/628796\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/newsycanuse.com\/index.php\/wp-json\/wp\/v2\/media\/628797"}],"wp:attachment":[{"href":"https:\/\/newsycanuse.com\/index.php\/wp-json\/wp\/v2\/media?parent=628796"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/newsycanuse.com\/index.php\/wp-json\/wp\/v2\/categories?post=628796"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/newsycanuse.com\/index.php\/wp-json\/wp\/v2\/tags?post=628796"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}