Hugging Face Trending Papers
FeaTok: Feature-Loss Tokenization for Image Generation
Read the original on Hugging Face Trending Papers →The Flow has not summarised this story yet — read it at Hugging Face Trending Papers.
Related stories
arXiv Computer Vision
3d ago
FeaTok: Feature-Loss Tokenization for Image Generation
FeaTok introduces feature-loss tokenization for image generation, training discrete tokens by reconstructing pretrained features instead of pixels. This approach improves generative performance across multiple self-supervised models (SigLIP, SigLIP2, DINOv2, DINOv3). In class-conditional generation, FeaTok reduces gFDr6 by up to 29% at 256×256 resolution and outperforms the strongest unguided continuous baseline with 38× less sampling compute, while also achieving higher human-preference reward-model scores and surpassing text-conditional baselines after only 500 fine-tuning updates.
arXiv Machine Learning
4d ago
Generative Representation Learning for Scalable Latent Tokenization
The paper proposes a self‑supervised representation learning method that serves as a tokenizer for generative models. It supervises a representation loss by factorizing pointwise mutual information, estimated through differences between conditional predictions of a learned generative model. The authors demonstrate that this approach avoids dimensional collapse and feature suppression seen in methods like DINO and LeJEPA, scales predictably with compute, and achieves compute‑optimal performance over direct diffusion for pixel‑space generation, with strong scaling laws on ImageNet and GPIC.
arXiv Computer Vision
Aug 27
Efficient Training with Foresight: Multi-Token Auxiliary Supervision for Autoregressive Image Generation
The paper introduces MTAR, a training framework for autoregressive image generation that enhances performance through multi-token prediction, token-level contrastive regularization, and semantic dropping. These components address sparse supervision, improve representation discriminability, and accelerate training without affecting inference. On ImageNet, MTAR outperforms LlamaGen with lower FID and faster training, achieving comparable results in only a third of the iterations.
arXiv Computer Vision
Oct 8
QuadTok: Quadtree Visual Tokenizer for Autoregressive Image Generation
QuadTok introduces a quadtree-based visual tokenizer that replaces traditional 2D grids or 1D token sequences with a hierarchical structure, allowing dynamic allocation of tokens to visually complex regions while using coarser resolution for homogeneous areas. The tokenizer, trained on ImageNet, reduces token usage by about 10% compared to a fixed 256-token grid and retains comparable reconstruction fidelity. Leveraging the tree’s natural causality, a 947 M GPT-style model conditioned on a quadtree topology achieves a 2.08 gFID on ImageNet 256×256 and supports zero‑shot spatially controlled image generation.
arXiv Machine Learning
Jun 5
Balancing Image Compression and Generation with Bootstrapped Tokenization
arXiv:2606. 05552v1 Announce Type: new Abstract: Despite progress in image tokenization, standard methods encode redundant information by mixing all granularities within each token, thus redundancy persists between tokens.
Hugging Face Trending Papers
Sep 29
NesTok: Nested Self-Aligned 1D Tokenizer for Autoregressive Image Generation
One-dimensional (1D) variable-length visual tokenizers enable adaptive compression by varying the number of tokens, allowing downstream autoregressive (AR) models to flexibly trade off generation qual...