-
NJU AALab
- Nanjing, China
Stars
[CVPR 2025] OVO-Bench: How Far is Your Video-LLMs from Real-World Online Video Understanding?
A simple video streaming baseline that outperforms SOTAs.
Evidence-grounded CVPR paper skills for research framing, writing, experimental design, figures, pre-submission review, and rebuttals.
See, hear, and interact on your local devices.
VoxCanvas-41M: A Large-Scale Bilingual Dataset for Instruction-Based Speech Editing and Its Learning Dynamics
[NeurIPS 2025] YOLOv12: Attention-Centric Real-Time Object Detectors
Fast, accurate OCR with correctness assessment.
ZwZ-v2: RL-trained vision-language models (4B/9B/27B) that connect fine-grained visual evidence across image regions, trained with GRPO on 20.4K verified multi-region VQA examples.
A research framework for principled agent self-improvement under frozen evaluators and declared mutation boundaries, recording verifiable lineage to make it reproducible and auditable.
An open-source model for understanding speech, environmental sounds, and music through captioning, question answering, and reasoning
A Survey on Video Understanding Agents
This is the latest survey for the Agentic System for Long-Form Video Understanding
Video-Arena: A Scalable Agentic Video Benchmark for Spatiotemporal Capability. Videos and GT on Hugging Face.
[ICML 2026] ZwZ model family: SOTA fine-grained perception performace; ZoomBench: a new challenging perception benchmark
Ming-3.0-tts is an open-source TTS model family with 0.5B dense and 7.9B-A1.3B MoE variants, supporting multilingual speech, voice cloning, voice design, and natural-language control.
Vision-OPD is a regional-to-global on-policy self-distillation framework that transfers a model's own privileged crop-conditioned perception to its full-image policy, enabling fine-grained visual u…
VideoJev: System One Model for Video Understanding and Interaction
Open multimodal decision models, 0.8B to 27B, for text, images and video. Decide with probabilities over your own candidates and locate with bounding boxes, in place of generated text. Free API, ON…
Make any agent harness multimodal-native.
[EMNLP 2026] CSAVocoder: A Causal Spatial Audio Vocoder Towards Real-Time Spatial Audio Generation