TokenFlow: Unified Image Tokenizer for Multimodal Understanding and Generation

• URL: http://arxiv.org/abs/2412.03069v1
• Date: Wed, 04 Dec 2024 06:46:55 GMT
• Title: TokenFlow: Unified Image Tokenizer for Multimodal Understanding and Generation
• Authors: Liao Qu, Huichao Zhang, Yiheng Liu, Xu Wang, Yi Jiang, Yiming Gao, Hu Ye, Daniel K. Du, Zehuan Yuan, Xinglong Wu,
• Abstract summary: TokenFlow is a novel unified image tokenizer that bridges the gap between multimodal understanding and generation. We demonstrate for the first time that discrete visual input can surpass LLaVA-1.5 13B in understanding performance. We also establish state-of-the-art performance in autoregressive image generation with a GenEval score of 0.55 at 256*256 resolution.
• Score: 26.29803524047736
• License:
• Abstract: We present TokenFlow, a novel unified image tokenizer that bridges the long-standing gap between multimodal understanding and generation. Prior research attempt to employ a single reconstruction-targeted Vector Quantization (VQ) encoder for unifying these two tasks. We observe that understanding and generation require fundamentally different granularities of visual information. This leads to a critical trade-off, particularly compromising performance in multimodal understanding tasks. TokenFlow addresses this challenge through an innovative dual-codebook architecture that decouples semantic and pixel-level feature learning while maintaining their alignment via a shared mapping mechanism. This design enables direct access to both high-level semantic representations crucial for understanding tasks and fine-grained visual features essential for generation through shared indices. Our extensive experiments demonstrate TokenFlow's superiority across multiple dimensions. Leveraging TokenFlow, we demonstrate for the first time that discrete visual input can surpass LLaVA-1.5 13B in understanding performance, achieving a 7.2\% average improvement. For image reconstruction, we achieve a strong FID score of 0.63 at 384*384 resolution. Moreover, TokenFlow establishes state-of-the-art performance in autoregressive image generation with a GenEval score of 0.55 at 256*256 resolution, achieving comparable results to SDXL.

Related papers

• Importance-based Token Merging for Diffusion Models [41.94334394794811]
  Diffusion models excel at high-quality image and video generation. A powerful way to speed them up is by merging similar tokens for faster computation. We show that preserving important tokens during merging significantly improves sample quality.
  arXiv  Detail & Related papers  (2024-11-23T02:01:49Z)
• ImageFolder: Autoregressive Image Generation with Folded Tokens [51.815319504939396]
  Increasing token length is a common approach to improve the image reconstruction quality. There exists a trade-off between reconstruction and generation quality regarding token length. We propose Image, a semantic tokenizer that provides spatially aligned image tokens that can be folded during autoregressive modeling.
  arXiv  Detail & Related papers  (2024-10-02T17:06:39Z)
• Matryoshka Multimodal Models [92.41824727506751]
  We propose M3: Matryoshka Multimodal Models, which learns to represent visual content as nested sets of visual tokens. We find that COCO-style benchmarks only need around 9 visual tokens to obtain accuracy similar to that of using all 576 tokens.
  arXiv  Detail & Related papers  (2024-05-27T17:59:56Z)
• Perceiving Longer Sequences With Bi-Directional Cross-Attention Transformers [13.480259378415505]
  BiXT scales linearly with input size in terms of computational cost and memory consumption. BiXT is inspired by the Perceiver architectures but replaces iterative attention with an efficient bi-directional cross-attention module. By combining efficiency with the generality and performance of a full Transformer architecture, BiXT can process longer sequences.
  arXiv  Detail & Related papers  (2024-02-19T13:38:15Z)
• Pink: Unveiling the Power of Referential Comprehension for Multi-modal LLMs [49.88461345825586]
  This paper proposes a new framework to enhance the fine-grained image understanding abilities of MLLMs. We present a new method for constructing the instruction tuning dataset at a low cost by leveraging annotations in existing datasets. We show that our model exhibits a 5.2% accuracy improvement over Qwen-VL and surpasses the accuracy of Kosmos-2 by 24.7%.
  arXiv  Detail & Related papers  (2023-10-01T05:53:15Z)
• ADDP: Learning General Representations for Image Recognition and Generation with Alternating Denoising Diffusion Process [94.41510903676837]
  We propose an Alternating Denoising Diffusion Process (ADDP) that integrates two spaces within a single representation learning framework. In each denoising step, our method first decodes pixels from previous VQ tokens, then generates new VQ tokens from the decoded pixels. The learned representations can be used to generate diverse high-fidelity images and also demonstrate excellent transfer performance on recognition tasks.
  arXiv  Detail & Related papers  (2023-06-08T17:59:32Z)
• How Image Generation Helps Visible-to-Infrared Person Re-Identification? [15.951145523749735]
  Flow2Flow is a framework that can jointly achieve training sample expansion and cross-modality image generation for V2I person ReID. For the purpose of identity alignment and modality alignment of generated images, we develop adversarial training strategies to train Flow2Flow. Experimental results on SYSU-MM01 and RegDB demonstrate that both training sample expansion and cross-modality image generation can significantly improve V2I ReID accuracy.
  arXiv  Detail & Related papers  (2022-10-04T13:09:29Z)
• Scalable Visual Transformers with Hierarchical Pooling [61.05787583247392]
  We propose a Hierarchical Visual Transformer (HVT) which progressively pools visual tokens to shrink the sequence length. It brings a great benefit by scaling dimensions of depth/width/resolution/patch size without introducing extra computational complexity. Our HVT outperforms the competitive baselines on ImageNet and CIFAR-100 datasets.
  arXiv  Detail & Related papers  (2021-03-19T03:55:58Z)
• ResNeSt: Split-Attention Networks [86.25490825631763]
  We present a modularized architecture, which applies the channel-wise attention on different network branches to leverage their success in capturing cross-feature interactions and learning diverse representations. Our model, named ResNeSt, outperforms EfficientNet in accuracy and latency trade-off on image classification.
  arXiv  Detail & Related papers  (2020-04-19T20:40:31Z)

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.