paper-with-me

홈 › Papers

Soft Tail-dropping for Adaptive Visual Tokenization

2026-01-20 · Zeyuan Chen, Kai Zhang, Zhuowen Tu, Yuanjun Xiong arxiv

We present Soft Tail-dropping Adaptive Tokenizer (STAT), a 1D discrete visual tokenizer that adaptively chooses the number of output tokens per image according to its structural complexity and level of detail. STAT encodes an image into a sequence of discrete codes together with per-token keep probabilities. Beyond standard autoencoder objectives, we regularize these keep probabilities to be monotonically decreasing along the sequence and explicitly align their distribution with an image-level complexity measure. As a result, STAT produces length-adaptive 1D visual tokens that are naturally compatible with causal 1D autoregressive (AR) visual generative models. On ImageNet-1k, equipping vanilla causal AR models with STAT yields competitive or superior visual generation quality compared to other probabilistic model families, while also exhibiting favorable scaling behavior that has been elusive in prior vanilla AR visual generation attempts.

📄 PDF Abstract BibTeX arXiv:2601.14246

Code (0)

등록된 구현이 없습니다.

Similar Papers 제목 키워드 기반

Discrete Tokenization Unlocks Transformers for Calibrated Tabular Forecasting

2026-03-08 · Yael S. Elmatad arxiv

Gradient boosting still dominates Transformers on tabular benchmarks. Our tokenizer uses a deliberately simplistic discretized vocabulary so we can highlight how even basic tokenization unlocks the power of attention on …

ChannelTok: Efficient Flexible-Length Vision Tokenization

2026-06-03 · Sukriti Paul, Arpit Bansal, Tom Goldstein arxiv

Leading flexible vision tokenizers achieve SOTA quality at an extreme cost, relying on parameter-heavy backbones and slow, multi-step generative decoders. We depart from this complex, spatial-token paradigm and introduce…

Image Generation

SGAD: Soft-Guided Adaptively-Dropped Neural Network

2018-07-04 · Zhisheng Wang, Fangxuan Sun, Jun Lin, Zhongfeng Wang 외

Deep neural networks (DNNs) have been proven to have many redundancies. Hence, many efforts have been made to compress DNNs. However, the existing model compression methods treat all the input samples equally while ignor…

Model Compression

Semantic-Aware Prefix Learning for Token-Efficient Image Generation

2026-03-26 · Qingfeng Li, Haoxian Zhang, Xu He, Songlin Tang 외 arxiv

Visual tokenizers play a central role in latent image generation by bridging high-dimensional images and tractable generative modeling. However, most existing tokenizers are still trained with reconstruction-dominated ob…

Representation LearningImage Generation

Keep-or-Drop? Adaptive Tokenizer for Compact Video Representation

2026-08-25 · Yeonkyeong Lee, Hyunsung Go, Jongmin Kim, Sewoong Lim 외 arxiv

Latent diffusion models have emerged as a dominant framework for high-fidelity image and video synthesis, operating in compact latent spaces with variational autoencoders (VAEs) to enhance computational efficiency withou…

Computational Efficiency