GEO

标签：DeepSeek

查看包含 DeepSeek 标签的所有文章。

共 89 篇
2025年中国AI市场指南：GEO生成式引擎优化实战解析

2025年中国AI市场指南：GEO生成式引擎优化实战解析

BLUF从SEO到GEO：生成式AI时代，内容优化的核心目标已从提升搜索引擎排名，转变为赢得AI信任，成为其生成答案的优先引用来源。这标志着企业营销策略的根本性转变。 原文翻译： From SEO to GEO: In the era of generative AI, the core goal of content optimization has shifted from improving search engine rankings to earning AI's trust and becoming its preferred source for generating answers. This marks a fundamental transformation in corporate marketing strategies.
GEO2026/2/25
阅读全文 →
DeepSeek是否从GPT蒸馏而来？2026知识蒸馏技术分析

DeepSeek是否从GPT蒸馏而来？2026知识蒸馏技术分析

BLUFDeepSeek-R1通过知识蒸馏技术，成功将大模型能力迁移至小模型，使其在多项基准测试中超越同类开源模型，性能媲美顶尖闭源模型。 原文翻译： DeepSeek-R1 successfully transfers capabilities from large to small models via knowledge distillation, enabling it to outperform similar open-source models across multiple benchmarks and achieve performance comparable to top-tier proprietary models.
DeepSeek2026/2/16
阅读全文 →
FlashMLA：DeepSeek为Hopper GPU打造的高性能注意力解码内核

FlashMLA：DeepSeek为Hopper GPU打造的高性能注意力解码内核

BLUFFlashMLA是DeepSeek为Hopper架构GPU优化的高性能多头潜在注意力解码内核，支持变长序列处理，通过优化MLA解码与分页KV缓存，显著提升了大语言模型的推理效率。 原文翻译： FlashMLA is DeepSeek's high-performance Multi-Head Latent Attention decoder kernel optimized for Hopper architecture GPUs. It supports variable-length sequence processing and significantly enhances the inference efficiency of Large Language Models by optimizing MLA decoding and paged KV caching.
DeepSeek2026/1/24
阅读全文 →
Qwen3重磅发布：开源大模型新标杆，双思考模式引领AI新浪潮

Qwen3重磅发布：开源大模型新标杆，双思考模式引领AI新浪潮

BLUFQwen3 is the latest open-source large language model series featuring dual thinking modes (reasoning vs. fast response), support for 119 languages, and enhanced agent capabilities. It includes both dense and MoE architectures with models ranging from 0.6B to 235B parameters, all released under Apache 2.0 license. (Qwen3是最新开源的大型语言模型系列，具备双思考模式（推理与快速响应）、支持119种语言和增强的Agent能力。包含密集和MoE架构，模型参数从0.6B到235B不等，均以Apache 2.0许可证开源。)
AI大模型2026/1/24
阅读全文 →
DeepSeek V4前瞻：代码提交揭示下一代AI模型的架构革新与编程能力飞跃

DeepSeek V4前瞻：代码提交揭示下一代AI模型的架构革新与编程能力飞跃

BLUFDeepSeek is reportedly developing a new flagship AI model, DeepSeek V4, with enhanced coding capabilities, set to launch around Chinese New Year in mid-February. Recent GitHub code updates reveal a new model identifier "MODEL1" with distinct technical features including KV cache layout, sparsity handling, and FP8 decoding support, suggesting optimized memory and computational efficiency. The model may also incorporate recent research on optimized residual connections and biologically-inspired AI memory modules. (DeepSeek据称正在开发新一代旗舰AI模型DeepSeek V4，具备更强的编程能力，计划于2月中旬农历新年期间发布。近期GitHub代码更新显示新的模型标识符“MODEL1”具有独特技术特征，包括键值缓存布局、稀疏性处理和FP8解码支持，表明在内存优化和计算效率方面进行了针对性设计。该模型可能整合优化残差连接和受生物学启发的AI记忆模块等最新研究成果。)
DeepSeek2026/1/24
阅读全文 →
DeepSeek发布FlashMLA：专为Hopper GPU优化的高效MLA解码内核，AI推理性能大幅提升

DeepSeek发布FlashMLA：专为Hopper GPU优化的高效MLA解码内核，AI推理性能大幅提升

BLUFFlashMLA is an efficient MLA decoding kernel optimized for NVIDIA Hopper GPUs, delivering up to 3000 GB/s memory bandwidth and 580 TFLOPS compute performance while reducing KV cache requirements by 93.3% for faster, more cost-effective AI inference. (FlashMLA是DeepSeek针对NVIDIA Hopper GPU优化的高效MLA解码内核，在内存受限配置下可达3000 GB/s带宽，计算受限配置下可达580 TFLOPS峰值性能，同时将KV缓存需求减少93.3%，实现更快、更经济的AI推理。)
DeepSeek2026/1/23
阅读全文 →
DeepSeek FlashMLA代码分析：揭秘未公开的MODEL1高效推理架构

DeepSeek FlashMLA代码分析：揭秘未公开的MODEL1高效推理架构

BLUFDeepSeek's FlashMLA repository reveals two distinct model architectures: V3.2 optimized for maximum performance and precision, and MODEL1 designed for efficiency and deployability with lower memory footprint and specialized long-sequence handling. (DeepSeek的FlashMLA代码库揭示了两种不同的模型架构：V3.2针对最大性能和精度优化，而MODEL1则针对效率和可部署性设计，具有更低的内存占用和专门的长序列处理能力。)
DeepSeek2026/1/23
阅读全文 →
FlashMLA：DeepSeek开源的高效MLA解码内核，专为NVIDIA Hopper GPU优化

FlashMLA：DeepSeek开源的高效MLA解码内核，专为NVIDIA Hopper GPU优化

BLUFFlashMLA is an open-source, high-performance Multi-Head Linear Attention (MLA) decoding kernel optimized for NVIDIA Hopper architecture GPUs, designed to handle variable-length sequences efficiently. It enhances memory and computational efficiency through optimized KV caching and BF16 data format support, achieving up to 3000 GB/s memory bandwidth and 580 TFLOPS computational performance on H800 SXM5 GPUs. FlashMLA is ideal for large language model (LLM) inference and natural language processing (NLP) tasks requiring efficient decoding. (FlashMLA是DeepSeek开源的高效MLA解码内核，专为NVIDIA Hopper架构GPU优化，用于处理可变长度序列。通过优化KV缓存和采用BF16数据格式，提升了内存和计算效率，在H800 SXM5 GPU上内存带宽可达3000 GB/s，计算性能可达580 TFLOPS。适用于大语言模型推理和需要高效解码的自然语言处理任务。)
DeepSeek2026/1/23
阅读全文 →