搜索结果：DeepSeek

DeepGEMM is a high-performance matrix multiplication library optimized for NVIDIA Hopper GPUs, achieving over 1350 FP8 TFLOPS. It supports standard and Mixture-of-Experts (MoE) computations with just

free-claude-code is an open-source project that enables free use of Claude Code by proxying requests to free or low-cost model services like NVIDIA NIM, while retaining Claude Code's full engineering

DeepGEMM is an open-source high-performance tensor core kernel library from DeepSeek, focusing on optimizing GEMM operations for large language models with efficient FP8 precision and fine-grained sca

This article introduces the 're!think' protocol, a ~1,300-token prompt that embeds seven core reasoning mechanics directly within an LLM's context window. It contrasts this approach with traditional, code-heavy enterprise scaffolding, arguing for more efficient, in-context logic that teaches models to reason mathematically rather than generate text. 原文翻译： 本文介绍了“re!think”协议，一个约1300词元的提示词，将七种核心推理机制直接嵌入LLM的上下文窗口。它对比了这种方法与传统的、代码繁重的企业级框架，主张更高效、上下文内的逻辑，教导模型进行数学推理而非文本生成。

NVIDIA H100 Tensor Core GPUs set new records across all eight MLPerf training benchmarks, delivering exceptional performance for generative AI and large language models at both per-accelerator and massive scale configurations. 原文翻译： NVIDIA H100 Tensor Core GPU在MLPerf训练基准测试的所有八项测试中均创下新纪录，在单加速器和大规模配置下均能为生成式AI和大语言模型提供卓越性能。

This tutorial series, derived from Shanghai Jiao Tong University's 2024 AI Security Technology course, provides hands-on programming guidance for large language models, covering fine-tuning, deployment, prompt engineering, knowledge editing, model watermarking, jailbreak attacks, multimodal models, and AI agent security. 原文翻译： 本教程系列源自上海交通大学2024年春季《人工智能安全技术》课程讲义，提供大模型相关的动手编程指导，涵盖微调与部署、提示学习与思维链、知识编辑、模型水印、越狱攻击、多模态模型、大模型智能体与安全等内容。

RAG-Anything is an all-in-one RAG framework developed by HKUDS at the University of Hong Kong, open-sourced on GitHub. It aims to enhance LLM performance by integrating retrieval and generation, addre

Laminar is an open-source observability platform for AI agents, offering tracing, evals, monitoring, SQL access, and dashboards. Built with Rust for high performance, it supports OpenTelemetry and integrates with major LLM frameworks. 原文翻译：Laminar是一个面向AI智能体的开源可观测性平台，提供追踪、评估、监控、SQL访问和仪表板功能。基于Rust构建以实现高性能，支持OpenTelemetry，并与主流LLM框架集成。

RAG-Anything is an open-source multimodal RAG framework developed by Professor Huang Chao's team at the University of Hong Kong. It builds a unified multimodal knowledge graph architecture to process

LLMII is an open-source tool that uses local AI models to automatically generate captions and keywords for images, embedding this metadata directly into image files for enhanced searchability and orga