MANZANO: A Simple and Scalable Unified Multimodal Model with a Hybrid Vision Tokenizer
This addresses the challenge of balancing visual understanding and generation in multimodal AI, offering a scalable solution for applications requiring both capabilities, though it is incremental in improving existing frameworks.
The authors tackled the performance trade-off between understanding and generating visual content in unified multimodal LLMs by introducing Manzano, a framework with a hybrid vision tokenizer and unified training recipe, achieving state-of-the-art results among unified models and competitive performance with specialist models, especially on text-rich evaluations.
Unified multimodal Large Language Models (LLMs) that can both understand and generate visual content hold immense potential. However, existing open-source models often suffer from a performance trade-off between these capabilities. We present Manzano, a simple and scalable unified framework that substantially reduces this tension by coupling a hybrid image tokenizer with a well-curated training recipe. A single shared vision encoder feeds two lightweight adapters that produce continuous embeddings for image-to-text understanding and discrete tokens for text-to-image generation within a common semantic space. A unified autoregressive LLM predicts high-level semantics in the form of text and image tokens, with an auxiliary diffusion decoder subsequently translating the image tokens into pixels. The architecture, together with a unified training recipe over understanding and generation data, enables scalable joint learning of both capabilities. Manzano achieves state-of-the-art results among unified models, and is competitive with specialist models, particularly on text-rich evaluation. Our studies show minimal task conflicts and consistent gains from scaling model size, validating our design choice of a hybrid tokenizer.