Multiscale Byte Language Models -- A Hierarchical Architecture for Causal Million-Length Sequence Modeling

• URL: http://arxiv.org/abs/2502.14553v1
• Date: Thu, 20 Feb 2025 13:31:50 GMT
• Title: Multiscale Byte Language Models -- A Hierarchical Architecture for Causal Million-Length Sequence Modeling
• Abstract summary: Byte Language Models (BLMs) have emerged to overcome tokenization, yet the excessive length of bytestreams requires new architectural paradigms.<n>We present the Multiscale Byte Language Model (MBLM), a model-agnostic hierarchical decoder stack that allows training with context windows of $5$M bytes on single GPU in full model precision.
• Score: 4.293921765264583
• License: http://creativecommons.org/licenses/by-sa/4.0/
• Abstract: Bytes form the basis of the digital world and thus are a promising building block for multimodal foundation models. Recently, Byte Language Models (BLMs) have emerged to overcome tokenization, yet the excessive length of bytestreams requires new architectural paradigms. Therefore, we present the Multiscale Byte Language Model (MBLM), a model-agnostic hierarchical decoder stack that allows training with context windows of $5$M bytes on single GPU in full model precision. We thoroughly examine MBLM's performance with Transformer and Mamba blocks on both unimodal and multimodal tasks. Our experiments demonstrate that hybrid architectures are efficient in handling extremely long byte sequences during training while achieving near-linear generational efficiency. To the best of our knowledge, we present the first evaluation of BLMs on visual Q\&A tasks and find that, despite serializing images and the absence of an encoder, a MBLM with pure next token prediction can match custom CNN-LSTM architectures with designated classification heads. We show that MBLMs exhibit strong adaptability in integrating diverse data representations, including pixel and image filestream bytes, underlining their potential toward omnimodal foundation models. Source code is publicly available at: https://github.com/ai4sd/multiscale-byte-lm

Related papers

• LLaDA2.0-Uni: Unifying Multimodal Understanding and Generation with Diffusion Large Language Model [43.908068133557215]
  We present LLaDA2.0-Uni, a unified discrete diffusion large language model (dLLM) that supports multimodal understanding and generation.<n>Its architecture combines a fully semantic discrete tokenizer, a MoE-based dLLM backbone, and a diffusion decoder.
  arXiv  Detail & Related papers  (2026-04-22T17:20:42Z)
• Cheers: Decoupling Patch Details from Semantic Representations Enables Unified Multimodal Comprehension and Generation [66.53544128707817]
  Cheers is a unified multimodal model that decouples patch-level details from semantic representations.<n> Cheers matches or surpasses advanced UMMs in both visual understanding and generation.
  arXiv  Detail & Related papers  (2026-03-13T08:55:27Z)
• ScreenCoder: Advancing Visual-to-Code Generation for Front-End Automation via Modular Multimodal Agents [40.697759330690815]
  ScreenCoder is a modular multi-agent framework that decomposes the task into three interpretable stages: grounding, planning, and generation.<n>By assigning these distinct responsibilities to specialized agents, our framework achieves significantly higher robustness and fidelity than end-to-end approaches.<n>Our approach achieves state-of-the-art performance in layout accuracy, structural coherence, and code correctness.
  arXiv  Detail & Related papers  (2025-07-30T16:41:21Z)
• OneLLM: One Framework to Align All Modalities with Language [86.8818857465443]
  We present OneLLM, an MLLM that aligns eight modalities to language using a unified framework.<n>OneLLM is evaluated on 25 diverse benchmarks, encompassing tasks such as multimodal captioning, question answering and reasoning.
  arXiv  Detail & Related papers  (2023-12-06T18:59:19Z)
• Reformulating Vision-Language Foundation Models and Datasets Towards Universal Multimodal Assistants [65.47222691674074]
  Muffin framework employs pre-trained vision-language models to act as providers of visual signals. UniMM-Chat dataset explores the complementarities of datasets to generate 1.1M high-quality and diverse multimodal instructions.
  arXiv  Detail & Related papers  (2023-10-01T12:35:18Z)
• Position-Enhanced Visual Instruction Tuning for Multimodal Large Language Models [50.07056960586183]
  We propose Position-enhanced Visual Instruction Tuning (PVIT) to extend the functionality of Multimodal Large Language Models (MLLMs) This integration promotes a more detailed comprehension of images for the MLLM. We present both quantitative experiments and qualitative analysis that demonstrate the superiority of the proposed model.
  arXiv  Detail & Related papers  (2023-08-25T15:33:47Z)
• MEGABYTE: Predicting Million-byte Sequences with Multiscale Transformers [78.85346970193518]
  Megabyte is a multi-scale decoder architecture that enables end-to-end differentiable modeling of sequences of over one million bytes. Experiments show that Megabyte allows byte-level models to perform competitively with subword models on long context language modeling. Results establish the viability of tokenization-free autoregressive sequence modeling at scale.
  arXiv  Detail & Related papers  (2023-05-12T00:55:41Z)
• BridgeTower: Building Bridges Between Encoders in Vision-Language Representation Learning [91.93547262073213]
  Vision-Language (VL) models with the Two-Tower architecture have dominated visual representation learning in recent years. We propose BridgeTower, which introduces multiple bridge layers that build a connection between the top layers of uni-modal encoders and each layer of the cross-modal encoder. BridgeTower achieves an accuracy of 78.73%, outperforming the previous state-of-the-art model METER by 1.09% with the same pre-training data.
  arXiv  Detail & Related papers  (2022-06-17T09:42:35Z)
• LAVENDER: Unifying Video-Language Understanding as Masked Language Modeling [102.42424022921243]
  Masked Language Modeling (MLM) is used as the common interface for all pre-training and downstream tasks. Experiments show that this unified framework achieves competitive performance on 14 VidL benchmarks.
  arXiv  Detail & Related papers  (2022-06-14T20:43:25Z)

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.