Official Kuaishou tech account, providing real-time updates on the latest Kuaishou technology practices.
At the China Multimedia 2024 conference in Yinchuan, Kuaishou presented its latest AI‑driven large‑model technologies—including text‑to‑image, text‑to‑video, and audio models—alongside advances in intelligent video coding, a new research‑fund initiative, and recent industry awards.
The article describes Gundam, Kuaishou's front‑end engineering scaffold, its goals of unified standards and efficiency, the practical challenges faced during the New Year Eve project, and the future direction of engineering standardization and automation through a layered engineering funnel approach.
This paper introduces Parrot, a system that enhances large language models' (LLMs) multi-turn instruction following capabilities through context-aware preference optimization (CaPO) and synthetic data generation, achieving significant performance improvements with limited training data.
This article introduces the Multidimensional Preference Model (MPS), the first multi‑dimensional scoring system for evaluating text‑to‑image generation, built on the newly released MHP dataset with extensive human annotations across aesthetic, semantic alignment, detail quality, and overall preference dimensions, and demonstrates its superior performance through comprehensive experiments and RLHF integration.
The article details Kuaishou’s development of the 175B “Kuaiyi” multimodal large model, presenting eight novel technical innovations—from Temporal Scaling Law and MiLe Loss to MoE‑enhanced reward modeling—and describes how these advances enable high‑performance AI services such as the AI Xiao Kuai chatbot across diverse real‑world scenarios.
LivePortrait is an open‑source, controllable portrait video generation framework that transfers facial expressions and poses from a driving video to static or dynamic portraits in real time, leveraging a 69M‑frame mixed video‑image training set, stitching and retargeting modules, and achieving high quality with low latency.
Kuaishou has officially open-sourced Kolors, a state-of-the-art text-to-image diffusion model that leverages ChatGLM3 for advanced bilingual text understanding and employs a two-stage training strategy to achieve photographic image quality rivaling leading proprietary systems.
Kuaishou VP App, the first short video platform on Apple Vision Pro, offers immersive features like spatial gestures and 3D likes, leveraging visionOS for enhanced user interaction.
The CVPR NTIRE 2024 short-form video quality assessment competition introduced the KVQ dataset, attracted over 200 teams, evaluated submissions using SROCC and PLCC metrics, and highlighted the winning approaches of SJTU MMLab, IH‑VQA, and TVQE, showcasing advances in AI‑driven video quality evaluation.
At CVPR 2024 in Seattle, Kuaishou presented its latest AI research and applications, including a star‑studded gala, a short‑video quality competition, large‑model video generation demos, a multi‑dimensional text‑to‑image evaluation paper, and advanced video processing technologies, underscoring its strong ties with the academic community.
