[NeurIPS'25] One-Step Diffusion for Detail-Rich and Temporally Consistent Video Super-Resolution

Efficient Triton Kernels for LLM Training

A Distributed Attention Towards Linear Scalability for Ultra-Long Context, Heterogeneous Data Training

📰 Must-read papers and blogs on LLM based Long Context Modeling 🔥

Video datasets

Unified automatic quality assessment for speech, music, and sound.

VideoNIAH: A Flexible Synthetic Method for Benchmarking Video MLLMs

哔哩哔哩常用API调用。支持视频、番剧、用户、频道、音频等功能。原仓库地址：https://github.com/MoyuScript/bilibili-api

An open platform for training, serving, and evaluating large language models. Release repo for Vicuna and Chatbot Arena.

gpt-oss-120b and gpt-oss-20b are two open-weight language models by OpenAI

从YouTube上爬取视频

RePO: Replay-Enhanced Policy Optimization

video-SALMONN 2 is a powerful audio-visual large language model (LLM) that generates high-quality audio-visual video captions, which is developed by the Department of Electronic Engineering at Tsin…

[ArXiv] V2PE: Improving Multimodal Long-Context Capability of Vision-Language Models with Variable Visual Position Encoding

A paper list of some recent works about Token Compress for Vit and VLM

Kimi K2 is the large language model series developed by Moonshot AI team

[CVPR 2024 Highlight] Official GraCo: Granularity-Controllable Interactive Segmentation.

Narrative movie understanding benchmark

Summary about Video-to-Text datasets. This repository is part of the review paper *Bridging Vision and Language from the Video-to-Text Perspective: A Comprehensive Review*

[MathCoder, MathCoder-VL] Family of LLMs/LMMs for mathematical reasoning.

Official repository for "AM-RADIO: Reduce All Domains Into One"

DELTA: Dense Efficient Long-range 3D Tracking for Any video (ICLR 2025)

[CVPR 2023] Official implementation of the paper: Fine-grained Audible Video Description

Official repository of "Event-based Video Frame Interpolation with Cross-Modal Asymmetric Bidirectional Motion Fields", CVPR 2023 paper(highlight)

[ICLR 2025, Oral] EmbodiedSAM: Online Segment Any 3D Thing in Real Time

🐳 Efficient Triton implementations for "Native Sparse Attention: Hardware-Aligned and Natively Trainable Sparse Attention"

Domain-specific language designed to streamline the development of high-performance GPU/CPU/Accelerators kernels