https://huggingface.co/spaces/MCG-NJU/PixNerd

\n","updatedAt":"2025-08-04T02:16:30.149Z","author":{"_id":"66615c855fd9d736e670e0a9","avatarUrl":"/avatars/0ff3127b513552432a7c651e21d7f283.svg","fullname":"wangshuai","name":"wangsssssss","type":"user","isPro":false,"isHf":false,"isHfAdmin":false,"isMod":false,"followerCount":10}},"numEdits":0,"identifiedLanguage":{"language":"en","probability":0.6322138905525208},"editors":["wangsssssss"],"editorAvatarUrls":["/avatars/0ff3127b513552432a7c651e21d7f283.svg"],"reactions":[{"reaction":"🔥","users":["search-facility"],"count":1}],"isReport":false}},{"id":"68901d3a5c3f77ec2fd42886","author":{"_id":"66615c855fd9d736e670e0a9","avatarUrl":"/avatars/0ff3127b513552432a7c651e21d7f283.svg","fullname":"wangshuai","name":"wangsssssss","type":"user","isPro":false,"isHf":false,"isHfAdmin":false,"isMod":false,"followerCount":10},"createdAt":"2025-08-04T02:38:50.000Z","type":"comment","data":{"edited":false,"hidden":false,"latest":{"raw":"# Revision of the inference time statistics\n\n\n\n| Model | **Inference** | | | **Training** | |\n|-------------------|---------------|--------|----------|--------------|----------|\n| | 1 image | 1 step | Mem (GB) | Speed (s/it) | Mem (GB) |\n| SiT-L/2(VAE-f8) | 0.51s | **0.0097s**| 2.9 | 0.30 | 18.4 |\n| Baseline-L/16 | 0.48s | **0.0097s**| 2.1 | 0.18 | 18 |\n| PixNerd-L/16 | 0.51s | 0.010s | 2.1 | 0.19 | 22 |\n\nDeeply sorry for this mistake, the single-step inference time of SiT-L/2 and Baseline-L is missing a zero (0.097s vs 0.0097s). The single-step inference time of PixNerd and Baseline is close.","html":"

Revision of the inference time statistics

\n

\n\n

Deeply sorry for this mistake, the single-step inference time of SiT-L/2 and Baseline-L is missing a zero (0.097s vs 0.0097s). The single-step inference time of PixNerd and Baseline is close.

\n","updatedAt":"2025-08-04T02:38:50.005Z","author":{"_id":"66615c855fd9d736e670e0a9","avatarUrl":"/avatars/0ff3127b513552432a7c651e21d7f283.svg","fullname":"wangshuai","name":"wangsssssss","type":"user","isPro":false,"isHf":false,"isHfAdmin":false,"isMod":false,"followerCount":10}},"numEdits":0,"identifiedLanguage":{"language":"en","probability":0.601259708404541},"editors":["wangsssssss"],"editorAvatarUrls":["/avatars/0ff3127b513552432a7c651e21d7f283.svg"],"reactions":[{"reaction":"❤️","users":["edmond","eaglgenes101"],"count":2}],"isReport":false},"replies":[{"id":"68918dce46620722ba43a38f","author":{"_id":"66615c855fd9d736e670e0a9","avatarUrl":"/avatars/0ff3127b513552432a7c651e21d7f283.svg","fullname":"wangshuai","name":"wangsssssss","type":"user","isPro":false,"isHf":false,"isHfAdmin":false,"isMod":false,"followerCount":10},"createdAt":"2025-08-05T04:51:26.000Z","type":"comment","data":{"edited":false,"hidden":false,"latest":{"raw":"Since the arxiv paper has been updated, so i closed this issue! feel free to open!","html":"

Since the arxiv paper has been updated, so i closed this issue! feel free to open!

\n","updatedAt":"2025-08-05T04:51:26.364Z","author":{"_id":"66615c855fd9d736e670e0a9","avatarUrl":"/avatars/0ff3127b513552432a7c651e21d7f283.svg","fullname":"wangshuai","name":"wangsssssss","type":"user","isPro":false,"isHf":false,"isHfAdmin":false,"isMod":false,"followerCount":10}},"numEdits":0,"identifiedLanguage":{"language":"en","probability":0.9719476103782654},"editors":["wangsssssss"],"editorAvatarUrls":["/avatars/0ff3127b513552432a7c651e21d7f283.svg"],"reactions":[],"isReport":false,"parentCommentId":"68901d3a5c3f77ec2fd42886"}}]},{"id":"68916041f9ee79541946e211","author":{"_id":"63d3e0e8ff1384ce6c5dd17d","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/1674830754237-63d3e0e8ff1384ce6c5dd17d.jpeg","fullname":"Librarian Bot (Bot)","name":"librarian-bot","type":"user","isPro":false,"isHf":false,"isHfAdmin":false,"isMod":false,"followerCount":264},"createdAt":"2025-08-05T01:37:05.000Z","type":"comment","data":{"edited":false,"hidden":false,"latest":{"raw":"This is an automated message from the [Librarian Bot](https://huggingface.co/librarian-bots). I found the following papers similar to this paper. \n\nThe following papers were recommended by the Semantic Scholar API \n\n* [Representation Entanglement for Generation:Training Diffusion Transformers Is Much Easier Than You Think](https://huggingface.co/papers/2507.01467) (2025)\n* [DGAE: Diffusion-Guided Autoencoder for Efficient Latent Representation Learning](https://huggingface.co/papers/2506.09644) (2025)\n* [Instella-T2I: Pushing the Limits of 1D Discrete Latent Space Image Generation](https://huggingface.co/papers/2506.21022) (2025)\n* [Taming Diffusion Transformer for Real-Time Mobile Video Generation](https://huggingface.co/papers/2507.13343) (2025)\n* [Diffusion Transformer-to-Mamba Distillation for High-Resolution Image Generation](https://huggingface.co/papers/2506.18999) (2025)\n* [Pyramidal Patchification Flow for Visual Generation](https://huggingface.co/papers/2506.23543) (2025)\n* [DC-AR: Efficient Masked Autoregressive Image Generation with Deep Compression Hybrid Tokenizer](https://huggingface.co/papers/2507.04947) (2025)\n\n\n Please give a thumbs up to this comment if you found it helpful!\n\n If you want recommendations for any Paper on Hugging Face checkout [this](https://huggingface.co/spaces/librarian-bots/recommend_similar_papers) Space\n\n You can directly ask Librarian Bot for paper recommendations by tagging it in a comment: `@librarian-bot recommend`","html":"

This is an automated message from the Librarian Bot. I found the following papers similar to this paper.

\n

The following papers were recommended by the Semantic Scholar API

\n

\n
• Representation Entanglement for Generation:Training Diffusion Transformers Is Much Easier Than You Think (2025)
\n
• DGAE: Diffusion-Guided Autoencoder for Efficient Latent Representation Learning (2025)
\n
• Instella-T2I: Pushing the Limits of 1D Discrete Latent Space Image Generation (2025)
\n
• Taming Diffusion Transformer for Real-Time Mobile Video Generation (2025)
\n
• Diffusion Transformer-to-Mamba Distillation for High-Resolution Image Generation (2025)
\n
• Pyramidal Patchification Flow for Visual Generation (2025)
\n
• DC-AR: Efficient Masked Autoregressive Image Generation with Deep Compression Hybrid Tokenizer (2025)
\n

\n

Please give a thumbs up to this comment if you found it helpful!

\n

If you want recommendations for any Paper on Hugging Face checkout this Space

\n

You can directly ask Librarian Bot for paper recommendations by tagging it in a comment: \n\n@librarian-bot\n\t recommend

\n","updatedAt":"2025-08-05T01:37:05.973Z","author":{"_id":"63d3e0e8ff1384ce6c5dd17d","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/1674830754237-63d3e0e8ff1384ce6c5dd17d.jpeg","fullname":"Librarian Bot (Bot)","name":"librarian-bot","type":"user","isPro":false,"isHf":false,"isHfAdmin":false,"isMod":false,"followerCount":264}},"numEdits":0,"identifiedLanguage":{"language":"en","probability":0.6492796540260315},"editors":["librarian-bot"],"editorAvatarUrls":["https://cdn-avatars.huggingface.co/v1/production/uploads/1674830754237-63d3e0e8ff1384ce6c5dd17d.jpeg"],"reactions":[],"isReport":false}},{"id":"689bd66f09cb26ac337b47b4","author":{"_id":"65d9fc2a0e6ad24551d87a1e","avatarUrl":"/avatars/3aedb9522cc3cd08349d654f523fd792.svg","fullname":"Grant Singleton","name":"grantsing","type":"user","isPro":false,"isHf":false,"isHfAdmin":false,"isMod":false,"followerCount":1},"createdAt":"2025-08-13T00:03:59.000Z","type":"comment","data":{"edited":false,"hidden":false,"latest":{"raw":"arXiv explained breakdown of this paper 👉 https://arxivexplained.com/papers/pixnerd-pixel-neural-field-diffusion","html":"

arXiv explained breakdown of this paper 👉 https://arxivexplained.com/papers/pixnerd-pixel-neural-field-diffusion

\n","updatedAt":"2025-08-13T00:03:59.773Z","author":{"_id":"65d9fc2a0e6ad24551d87a1e","avatarUrl":"/avatars/3aedb9522cc3cd08349d654f523fd792.svg","fullname":"Grant Singleton","name":"grantsing","type":"user","isPro":false,"isHf":false,"isHfAdmin":false,"isMod":false,"followerCount":1}},"numEdits":0,"identifiedLanguage":{"language":"en","probability":0.6850689649581909},"editors":["grantsing"],"editorAvatarUrls":["/avatars/3aedb9522cc3cd08349d654f523fd792.svg"],"reactions":[],"isReport":false}}],"primaryEmailConfirmed":false,"paper":{"id":"2507.23268","authors":[{"_id":"688c1ec68c434640078cc386","user":{"_id":"66615c855fd9d736e670e0a9","avatarUrl":"/avatars/0ff3127b513552432a7c651e21d7f283.svg","isPro":false,"fullname":"wangshuai","user":"wangsssssss","type":"user"},"name":"Shuai Wang","status":"claimed_verified","statusLastChangedAt":"2025-08-06T19:30:56.415Z","hidden":false},{"_id":"688c1ec68c434640078cc387","name":"Ziteng Gao","hidden":false},{"_id":"688c1ec68c434640078cc388","user":{"_id":"65e55ca4a0681de63022843e","avatarUrl":"/avatars/5b1ac4a81f0c38fda6f47b392f7474c8.svg","isPro":false,"fullname":"zhu chenhui","user":"flateon","type":"user"},"name":"Chenhui Zhu","status":"claimed_verified","statusLastChangedAt":"2025-08-06T19:30:51.704Z","hidden":false},{"_id":"688c1ec68c434640078cc389","name":"Weilin Huang","hidden":false},{"_id":"688c1ec68c434640078cc38a","name":"Limin Wang","hidden":false}],"publishedAt":"2025-07-31T06:07:20.000Z","submittedOnDailyAt":"2025-08-04T00:45:14.702Z","title":"PixNerd: Pixel Neural Field Diffusion","submittedOnDailyBy":{"_id":"66615c855fd9d736e670e0a9","avatarUrl":"/avatars/0ff3127b513552432a7c651e21d7f283.svg","isPro":false,"fullname":"wangshuai","user":"wangsssssss","type":"user"},"summary":"The current success of diffusion transformers heavily depends on the\ncompressed latent space shaped by the pre-trained variational autoencoder(VAE).\nHowever, this two-stage training paradigm inevitably introduces accumulated\nerrors and decoding artifacts. To address the aforementioned problems,\nresearchers return to pixel space at the cost of complicated cascade pipelines\nand increased token complexity. In contrast to their efforts, we propose to\nmodel the patch-wise decoding with neural field and present a single-scale,\nsingle-stage, efficient, end-to-end solution, coined as pixel neural field\ndiffusion~(PixelNerd). Thanks to the efficient neural field representation in\nPixNerd, we directly achieved 2.15 FID on ImageNet 256times256 and 2.84 FID\non ImageNet 512times512 without any complex cascade pipeline or VAE. We also\nextend our PixNerd framework to text-to-image applications. Our PixNerd-XXL/16\nachieved a competitive 0.73 overall score on the GenEval benchmark and 80.9\noverall score on the DPG benchmark.","upvotes":51,"discussionId":"688c1ec68c434640078cc38b","projectPage":"https://huggingface.co/spaces/MCG-NJU/PixNerd","githubRepo":"https://github.com/MCG-NJU/PixNerd","ai_summary":"Pixel Neural Field Diffusion (PixNerd) achieves high-quality image generation in a single-scale, single-stage process without VAEs or complex pipelines, and extends to text-to-image applications with competitive performance.","ai_keywords":["diffusion transformers","compressed latent space","pre-trained variational autoencoder","pixel space","patch-wise decoding","neural field","single-scale","single-stage","end-to-end solution","pixel neural field diffusion","PixNerd","FID","ImageNet","text-to-image","GenEval benchmark","DPG benchmark"],"githubStars":118},"canReadDatabase":false,"canManagePapers":false,"canSubmit":false,"hasHfLevelAccess":false,"upvoted":false,"upvoters":[{"_id":"6427e08288215cee63b1c44d","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/6427e08288215cee63b1c44d/rzaG978FF-ywzicWNl_xl.jpeg","isPro":false,"fullname":"yao teng","user":"tytyt","type":"user"},{"_id":"66615c855fd9d736e670e0a9","avatarUrl":"/avatars/0ff3127b513552432a7c651e21d7f283.svg","isPro":false,"fullname":"wangshuai","user":"wangsssssss","type":"user"},{"_id":"648ac3d53470b17ccc90deaf","avatarUrl":"/avatars/35b678baa7a91959f40f704c5de2a3e1.svg","isPro":false,"fullname":"Ziteng Gao","user":"sebgao","type":"user"},{"_id":"6441438acea37249a0fd8775","avatarUrl":"/avatars/2cc696926e2ef9ce7d58210c45875a82.svg","isPro":false,"fullname":"SHEN","user":"Hiyoung","type":"user"},{"_id":"65d851096769b3a9c9376134","avatarUrl":"/avatars/e4ca32cf96fa8bcba44f1f1dc8e0ec9b.svg","isPro":false,"fullname":"ZehongMa","user":"zehongma","type":"user"},{"_id":"664ee978e24897b66cac36b0","avatarUrl":"/avatars/1757f533b08e6d85ae8bd4faa5496ef2.svg","isPro":false,"fullname":"Chunxu Liu","user":"lcxrocks","type":"user"},{"_id":"65768087e58db15694d1edbe","avatarUrl":"/avatars/94aa8526d1fa9686788512be697d9a08.svg","isPro":false,"fullname":"Yifan Xu","user":"geekifan","type":"user"},{"_id":"64ad086c5d48838462e2eee1","avatarUrl":"/avatars/735b29d01bae05599e7d5ce61b223153.svg","isPro":false,"fullname":"Du","user":"xiaobiaodu","type":"user"},{"_id":"633e6f07309a99325095dd42","avatarUrl":"/avatars/57b91a488ac1745b3c0509c04eb6ad93.svg","isPro":false,"fullname":"Hoigi Seo","user":"Agorium","type":"user"},{"_id":"6513030fb3a463e17df56edd","avatarUrl":"/avatars/867bd4316b2de758654ad3a84ea868c1.svg","isPro":false,"fullname":"Hyun, Jeongseok","user":"js-hyun","type":"user"},{"_id":"6342796a0875f2c99cfd313b","avatarUrl":"/avatars/98575092404c4197b20c929a6499a015.svg","isPro":false,"fullname":"Yuseung \"Phillip\" Lee","user":"phillipinseoul","type":"user"},{"_id":"664920999e82dfd4a3fe1914","avatarUrl":"/avatars/fafbaa3a94fb25832f411c6b2c7b3aa8.svg","isPro":false,"fullname":"larouri","user":"hachim1","type":"user"}],"acceptLanguages":["*"],"dailyPaperRank":3}">