Tuna-2: Pixel Embeddings Beat Vision Encoders for Multimodal Understanding and Generation
Paper • 2604.24763 • Published • 64
Ujjwal Tyagi
Ujjwal-Tyagi
AI & ML interests
Chief Scientist at Shirova AI, focused on advancing open-source AI, Experienced in LLM fine-tuning, model architecture, and research, with a strong interest in building scalable and efficient models
Recent Activity
liked a dataset 1 day ago
Kassadin88/GLM-5.1-OpenThoughts3-Distill liked a model 1 day ago
ibm-granite/granite-4.1-30b liked a model 2 days ago
XiaomiMiMo/MiMo-V2.5Organizations
upvoted a paper 2 days ago
upvoted 7
papers 3 days ago
Enhancing Spatial Understanding in Image Generation via Reward Modeling
Paper • 2602.24233 • Published • 60
ScaleEdit-12M: Scaling Open-Source Image Editing Data Generation via Multi-Agent Framework
Paper • 2603.20644 • Published • 5
EditCaption: Human-Aligned Instruction Synthesis for Image Editing via Supervised Fine-Tuning and Direct Preference Optimization
Paper • 2604.08213 • Published • 1
ImageEdit-R1: Boosting Multi-Agent Image Editing via Reinforcement Learning
Paper • 2603.08059 • Published • 1
OARS: Process-Aware Online Alignment for Generative Real-World Image Super-Resolution
Paper • 2603.12811 • Published • 1
EditHF-1M: A Million-Scale Rich Human Preference Feedback for Image Editing
Paper • 2603.14916 • Published • 1
Trust Your Critic: Robust Reward Modeling and Reinforcement Learning for Faithful Image Editing and Generation
Paper • 2603.12247 • Published • 24
upvoted a collection 3 days ago
upvoted 11
papers 3 days ago
Diff-Aid: Inference-time Adaptive Interaction Denoising for Rectified Text-to-Image Generation
Paper • 2602.13585 • Published • 1
Omni-Video 2: Scaling MLLM-Conditioned Diffusion for Unified Video Generation and Editing
Paper • 2602.08820 • Published • 1
DeepGen 1.0: A Lightweight Unified Multimodal Model for Advancing Image Generation and Editing
Paper • 2602.12205 • Published • 82
Language-Free Generative Editing from One Visual Example
Paper • 2603.25441 • Published • 1
RPiAE: A Representation-Pivoted Autoencoder Enhancing Both Image Generation and Editing
Paper • 2603.19206 • Published • 1
RefAlign: Representation Alignment for Reference-to-Video Generation
Paper • 2603.25743 • Published • 2
EVLF: Early Vision-Language Fusion for Generative Dataset Distillation
Paper • 2603.07476 • Published • 1
Reflective Flow Sampling Enhancement
Paper • 2603.06165 • Published • 1
Rethinking UMM Visual Generation: Masked Modeling for Efficient Image-Only Pre-training
Paper • 2603.16139 • Published • 33
Towards Training-Free Scene Text Editing
Paper • 2603.24571 • Published • 1
OmniScript: Towards Audio-Visual Script Generation for Long-Form Cinematic Video
Paper • 2604.11102 • Published • 8