Official account of Tencent Technology. A platform for publishing and analyzing Tencent's technological innovations and cutting-edge developments.
This article explains how modern GPUs evolved from graphics rendering to general‑purpose computing, details the CPU‑GPU heterogenous architecture, walks through a CUDA demo that adds two billion‑element arrays, compares CPU and GPU performance, and describes the compilation, loading, and execution pipeline of CUDA kernels.
Tencent’s TEG team secured first place in the NTIRE 2025 UGC Video Enhancement competition by introducing a progressive, three‑stage AI framework that decomposes enhancement tasks into expert models for color correction, denoising, and temporal stability, incorporates advanced loss functions, extensive hardware‑level optimizations, INT8 quantization techniques, and outlines future diffusion‑based generative enhancements.
This article details how the Wukong AI Agent automatically audited the popular LLaMA‑Factory project, discovered a high‑severity remote code execution vulnerability (CVE‑2025‑53002) caused by unsafe torch.load usage, reported it to the maintainers, and demonstrated the official fix that adds a secure weights_only flag.
The IH‑VQA team’s iMatch solution clinched the CVPR2025 NTIRE Image‑Text Alignment champion by introducing dual‑model fusion, pseudo‑label data augmentation, Q‑Align probability mapping, and visual augmentations, and the paper also presents a comprehensive iMatch benchmark evaluating 23 state‑of‑the‑art text‑to‑image models across multiple resolutions.
This article explains how to reduce Unreal Engine Android package size and native library memory usage by applying a series of build‑time and resource‑trimming strategies—including native lib compression, code size reduction, relocation table compression, and selective asset packaging—while preserving runtime performance.
This comprehensive article explores the evolution of AI agents, explains key protocols like MCP and A2A, compares reasoning frameworks such as CoT, ReAct, and Plan‑and‑Execute, and demonstrates how Golang frameworks Eino and tRPC‑A2A‑Go enable elegant development, orchestration, and observability of complex multi‑agent systems with practical code examples and visual diagrams.
This comprehensive guide walks you through the fundamentals of Retrieval‑Augmented Generation (RAG) and AI agents, explains their inner workings, shares optimization tricks, provides ready‑to‑run code snippets, and demonstrates how to evaluate performance with metrics such as recall, faithfulness, and answer relevance.
This article critically examines the rise of Model Context Protocol (MCP) in AI, explaining its purpose as a unified tool‑calling standard, detailing its architecture, comparing it with traditional function calls, and evaluating the technical and market challenges that limit its universal applicability.
This article explores how a multi‑agent AI system dramatically improves 0‑day vulnerability detection by automating code audit, reducing false positives, and outperforming traditional static analysis tools in large‑scale real‑world benchmarks.
Kuikly, Tencent’s Kotlin Multiplatform‑based cross‑platform framework, has been extended to support HarmonyOS with native performance, offering Compose DSL, high‑performance rendering, Kotlin Native adaptation, debugging tools, and detailed optimization strategies that achieve near‑native speed and dynamic UI capabilities across Android, iOS, Web, and mini‑programs.
