From Tokens to Trajectories: Parameter-Efficient Continuous Adaptation of Unified Multimodal Models
cs.CV
Submitted: 2026-11-03
Updated: 2026-11-03
Terminology
Sources
- DRAGON: A Large-Scale Dataset of Realistic Images Generated by Diffusion Models
- Z-Image: An Efficient Image Generation Foundation Model with Single-Stream Diffusion Transformer
- Janus-Pro: Unified Multimodal Understanding and Generation with Data and Model Scaling
- OpenGPT-4o-Image: A Comprehensive Dataset for Advanced Image Generation and Editing
- Emerging Properties in Unified Multimodal Pretraining
- LoRA: Low-Rank Adaptation of Large Language Models
- One Editor, Many Edits: A Unified Training-Free Framework for Diverse Video Editing
- FLUX.1 Kontext: Flow Matching for In-Context Image Generation and Editing in Latent Space
- Discrete Tokenization for Multimodal LLMs: A Comprehensive Survey
- LAION-SG: An Enhanced Large-Scale Dataset for Training Complex Image-Text Models with Structural Annotations
- VTBench: Evaluating Visual Tokenizers for Autoregressive Image Generation
- Muon is Scalable for LLM Training
- Flow Straight and Fast: Learning to Generate and Transfer Data with Rectified Flow
- Direct Preference Optimization: Your Language Model is Secretly a Reward Model
- MammothModa2: A Unified AR-Diffusion Framework for Multimodal Understanding and Generation
- Muddit: Liberating Generation Beyond Text-to-Image with a Unified Discrete Diffusion Model
- Unified Multimodal Discrete Diffusion
- Chameleon: Mixed-Modal Early-Fusion Foundation Models
- Wan: Open and Advanced Large-Scale Video Generative Models
- Emu3: Next-Token Prediction is All You Need
Related papers
- TEA: Text Encoder Alignment for Robust Concept Erasure in Text-to-Image Models
- Do Egocentric Video-Language Models Capture Both Hand- and Object-Centric Cues?
- Aura: Consistent Multi-Subject Video Generation via VLM-Grounded Semantic Alignment
- PVRF: All-in-one Adverse Weather Removal via Prior-modulated and Velocity-constrained Rectified Flow
- DLWM: Diverse Latent World Models for Efficient Multimodal Reasoning
- Out of Sight, Still in Mind: Token Compression for Omni-LLMs