Aligning Visual Foundation Encoders to Tokenizers for Diffusion Models

• URL: http://arxiv.org/abs/2509.25162v1
• Date: Mon, 29 Sep 2025 17:57:39 GMT
• Title: Aligning Visual Foundation Encoders to Tokenizers for Diffusion Models
• Authors: Bowei Chen, Sai Bi, Hao Tan, He Zhang, Tianyuan Zhang, Zhengqi Li, Yuanjun Xiong, Jianming Zhang, Kai Zhang,
• Abstract summary: We propose aligning pretrained visual encoders to serve as tokenizers for latent diffusion models in image generation.<n>On ImageNet 256$times$256, our tokenizer accelerates the convergence of diffusion models, reaching a gFID of 1.90 within just 64 epochs.<n>Our method is simple, scalable, and establishes a semantically grounded paradigm for continuous tokenizer design.
• Score: 37.59115132356727
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: In this work, we propose aligning pretrained visual encoders to serve as tokenizers for latent diffusion models in image generation. Unlike training a variational autoencoder (VAE) from scratch, which primarily emphasizes low-level details, our approach leverages the rich semantic structure of foundation encoders. We introduce a three-stage alignment strategy: (1) freeze the encoder and train an adapter and a decoder to establish a semantic latent space; (2) jointly optimize all components with an additional semantic preservation loss, enabling the encoder to capture perceptual details while retaining high-level semantics; and (3) refine the decoder for improved reconstruction quality. This alignment yields semantically rich image tokenizers that benefit diffusion models. On ImageNet 256$\times$256, our tokenizer accelerates the convergence of diffusion models, reaching a gFID of 1.90 within just 64 epochs, and improves generation both with and without classifier-free guidance. Scaling to LAION, a 2B-parameter text-to-image model trained with our tokenizer consistently outperforms FLUX VAE under the same training steps. Overall, our method is simple, scalable, and establishes a semantically grounded paradigm for continuous tokenizer design.

Related papers

• DINO-SAE: DINO Spherical Autoencoder for High-Fidelity Image Reconstruction and Generation [47.409626500688866]
  We present the DINO Spherical Autoencoder (DINO-SAE), a framework that bridges semantic representation and pixel-level reconstruction.<n>Our approach achieves state-of-the-art reconstruction quality, reaching 0.37 rFID and 26.2 dB PSNR, while maintaining strong semantic alignment to the pretrained VFM.
  arXiv  Detail & Related papers  (2026-01-30T12:25:34Z)
• Both Semantics and Reconstruction Matter: Making Representation Encoders Ready for Text-to-Image Generation and Editing [62.94394079771687]
  A burgeoning trend is to adopt high-dimensional features from representation encoders as generative latents.<n>We propose a systematic framework to adapt understanding-oriented encoder features for generative tasks.<n>We show that our approach achieves state-of-the-art reconstruction, faster convergence, and substantial performance gains in both Text-to-Image (T2I) and image editing tasks.
  arXiv  Detail & Related papers  (2025-12-19T18:59:57Z)
• VQRAE: Representation Quantization Autoencoders for Multimodal Understanding, Generation and Reconstruction [83.50898344094153]
  VQRAE produces Continuous semantic features for image understanding and tokens for visual generation within a unified tokenizer.<n>Design enables negligible semantic information for maintaining the ability of multimodal understanding, discrete tokens.<n>VQRAE presents competitive performance on several benchmarks of visual understanding, generation and reconstruction.
  arXiv  Detail & Related papers  (2025-11-28T17:26:34Z)
• Encoder-Decoder Diffusion Language Models for Efficient Training and Inference [27.027494484311948]
  We introduce a framework for Efficient Diffusion-Decoder (E2D2) consisting of an architecture with specialized training and sampling algorithms.<n>E2D2 achieves superior trade-offs between generation quality and inference on summarization, translation, and mathematical reasoning tasks.
  arXiv  Detail & Related papers  (2025-10-26T22:05:22Z)
• Vision Foundation Models as Effective Visual Tokenizers for Autoregressive Image Generation [66.73899356886652]
  We build an image tokenizer directly atop pre-trained vision foundation models.<n>Our proposed image tokenizer, VFMTok, achieves substantial improvements in image reconstruction and generation quality.<n>It further boosts autoregressive (AR) generation -- achieving a gFID of 2.07 on ImageNet benchmarks.
  arXiv  Detail & Related papers  (2025-07-11T09:32:45Z)
• One-Way Ticket:Time-Independent Unified Encoder for Distilling Text-to-Image Diffusion Models [65.96186414865747]
  Text-to-Image (T2I) diffusion models face a trade-off between inference speed and image quality.<n>We introduce the first Time-independent Unified TiUE for the student model UNet architecture.<n>Using a one-pass scheme, TiUE shares encoder features across multiple decoder time steps, enabling parallel sampling.
  arXiv  Detail & Related papers  (2025-05-28T04:23:22Z)
• Epsilon-VAE: Denoising as Visual Decoding [61.29255979767292]
  We propose denoising as decoding, shifting from single-step reconstruction to iterative refinement.<n>Specifically, we replace the decoder with a diffusion process that iteratively refines noise to recover the original image.<n>By adopting iterative reconstruction through diffusion, our autoencoder, namely Epsilon-VAE, achieves high reconstruction quality.
  arXiv  Detail & Related papers  (2024-10-05T08:27:53Z)
• Faster Diffusion: Rethinking the Role of the Encoder for Diffusion Model Inference [95.42299246592756]
  We study the UNet encoder and empirically analyze the encoder features. We find that encoder features change minimally, whereas the decoder features exhibit substantial variations across different time-steps. We validate our approach on other tasks: text-to-video, personalized generation and reference-guided generation.
  arXiv  Detail & Related papers  (2023-12-15T08:46:43Z)
• A Principled Hierarchical Deep Learning Approach to Joint Image Compression and Classification [27.934109301041595]
  This work proposes a three-step joint learning strategy to guide encoders to extract features that are compact, discriminative, and amenable to common augmentations/transformations. Tests show that our proposed method achieves accuracy improvement of up to 1.5% on CIFAR-10 and 3% on CIFAR-100 over conventional E2E cross-entropy training.
  arXiv  Detail & Related papers  (2023-10-30T15:52:18Z)
• Triple-View Knowledge Distillation for Semi-Supervised Semantic Segmentation [54.23510028456082]
  We propose a Triple-view Knowledge Distillation framework, termed TriKD, for semi-supervised semantic segmentation. The framework includes the triple-view encoder and the dual-frequency decoder.
  arXiv  Detail & Related papers  (2023-09-22T01:02:21Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.