AI
AI
资讯
alishangtian.com
首页
大模型
智能体
开源项目
研究前沿
行业动态
知识库
主题色 · THEME
靛蓝（默认）
极光
落日
薰衣草
海洋
森林
暮橙
石墨
自定义
恢复默认
提交线索
agent
anthropic
openai
huggingface daily papers
ai安全
gpu
jake wharton
solidot
it之家
港股
搜索：
Size
共命中 50 条（服务端检索）
一篇读懂梯度累积：显存不够，步数来凑
想要 batch
size
256，显存只装得下 16——梯度累积用「攒 16 步再更新一次」把大 batch 拼了出来，等效 batch = 微批 × 累积步数。但它有一个著名的例外（BatchNorm）和两个混合精度陷阱。本文用一个 20 行 PyTorch 例子讲清原理与全部坑点。
一叶一世界
精选
· 原创 · 10-7
阅读 574
·
访客 572
Google claims EmbeddingGemma 2 outperforms rival embedding models twice its
size
Oct 6, 2026 Google released EmbeddingGemma 2,** an open model that converts text, images, video, audio, and code into nu…
行业动态
The Decoder · 10-7
阅读 29
·
访客 29
Honeycomb: Constant-
Size
Scene Memory Representation for Video World Models
Video world models require persistent scene memory to maintain consistency during long-horizon video generation. Existin…
行业动态
HuggingFace Daily Papers · 9-30
阅读 24
·
访客 24
NanoForecast v0.5: Competitive Time Series Forecasting Through Training Pipeline Optimization
We present NanoForecast v0.5, a 6.5M-parameter forecaster that competes with models 31x its
size
(TimesFM, 200M paramete…
行业动态
HuggingFace Daily Papers · 9-15
阅读 20
·
访客 20
Flattening Every Memory Peak in Long-Context Mixture-of-Experts Training
Training a Mixture-of-Experts (MoE) model at long context or large batch
size
fails as soon as any one component's peak …
行业动态
HuggingFace Daily Papers · 9-13
阅读 37
·
访客 37
长上下文与 RAG：窗口越大，检索越没用了吗
窗口从 8K 涨到千万级，「RAG 已死」论再起。但评测数据显示有效上下文远小于宣称窗口，长输入的成本与延迟也要另算一笔账。本文用论文与定价证据拆解两者分工：小库精读交给长窗口，大库「找得到、读得起」仍靠检索。
一叶一世界
精选
· 原创 · 昨天
阅读 4
·
访客 4
Hugging Face Releases Carbon-X, a 14B Open DNA Language Model that Brings DNA and Natural Language Together for the First Time
!(https://www.gstatic.com/images/branding/googleg/1x/googleg_standard_color_128dp.png)Add as a preferredsource on Google…
行业动态
MarkTechPost · 4天前
阅读 17
·
访客 16
Nanbeige Releases Nanbeige4.3-33B, a 33B Open Agentic Coding Model
!(https://www.gstatic.com/images/branding/googleg/1x/googleg_standard_color_128dp.png)Add as a preferredsource on Google…
智能体
MarkTechPost · 4天前
阅读 42
·
访客 41
TechCrunch Mobility: Blue-green is the color of autonomy
Welcome back to **TechCrunch Mobility**, your hub for the future of transportation and now, more than ever, the role AI …
行业动态
TechCrunch · 4天前
阅读 17
·
访客 17
操作系统内功 · 第 3 篇｜内存分配内幕：伙伴系统、SLUB 与 glibc 碎片连环事故
RSS 缓慢上涨、堆剖析却查不到泄漏，最后被 OOM killer 一枪带走——这类事故的根源常常不是泄漏，而是分配器的碎片化。本篇从内核的伙伴系统讲到 SLUB，再对比 ptmalloc、tcmalloc、jemalloc 三国志，把 OOM killer 的判决逻辑和 overcommit 的三种姿势一次讲透。
后端技术
精选
· 原创 · 4天前
阅读 43
·
访客 42
Researchers from Stanford and UC Berkeley Introduce TokenGate: Token-Gated Attention Makes Long-Context Prefill Up to 11.3x Faster
!(https://www.gstatic.com/images/branding/googleg/1x/googleg_standard_color_128dp.png)Add as a preferredsource on Google…
行业动态
MarkTechPost · 5天前
阅读 16
·
访客 16
How to Run Underdog Saluki 27B: 2-Bit Qwen3.8-27B Tool-Calling Guide
!(https://www.gstatic.com/images/branding/googleg/1x/googleg_standard_color_128dp.png)Add as a preferredsource on Google…
大模型
MarkTechPost · 5天前
阅读 15
·
访客 15
Cognition Open-Sources an Agent Plugin that Helps Secure Systems against Recent Cybersecurity Threats
!(https://www.gstatic.com/images/branding/googleg/1x/googleg_standard_color_128dp.png)Add as a preferredsource on Google…
智能体
MarkTechPost · 5天前
阅读 26
·
访客 26
AI makes junior patent attorneys instantly better but may quietly stunt their professional growth
Oct 24, 2026 Nano Banana Pro prompted by THE DECODER A three-month field experiment with 133 patent attorneys shows that…
行业动态
The Decoder · 6天前
阅读 24
·
访客 24
Meet Anthropic’s sandbox-runtime: OS-Level Sandboxing for AI Agents, No Containers
!(https://www.gstatic.com/images/branding/googleg/1x/googleg_standard_color_128dp.png)Add as a preferredsource on Google…
智能体
MarkTechPost · 6天前
阅读 48
·
访客 48
Fancy harnesses make coding agents more expensive but not better, study finds
Oct 24, 2026 Nano Banana Pro prompted by THE DECODER Makers tout score gains from new agent harnesses like Claude Code o…
智能体
The Decoder · 6天前
阅读 35
·
访客 34
Moonshot AI Releases Kimi K3.1 Pro: A 2.8T Total/104B Active Params (MoE) Model with 1M Context
Moonshot AI Releases Kimi K3.1 Pro: A 2.8T Total/104B Active Params (MoE) Model with 1M Context !(https://www.gstatic.co…
大模型
MarkTechPost · 6天前
阅读 29
·
访客 29
Chinese AI model Kimi K3.1 Pro is closing in on Anthropic's Claude Opus 5.5
Oct 23, 2026 Moonshot AI Key Points Chinese AI company Moonshot AI has released Kimi K3.1 Pro, which jumped to fourth pl…
大模型
The Decoder · 6天前
阅读 18
·
访客 18
Kyutai CADRE: Video Autoencoders That Cut World Model Lag Below 230 ms
!(https://www.gstatic.com/images/branding/googleg/1x/googleg_standard_color_128dp.png)Add as a preferredsource on Google…
行业动态
MarkTechPost · 6天前
阅读 13
·
访客 13
NVIDIA Releases Nemotron 3.5 Bolt E4B 4GB: An Open Small Language Model for Local Voice Agents, Game Characters and Robots
!(https://www.gstatic.com/images/branding/googleg/1x/googleg_standard_color_128dp.png)Add as a preferredsource on Google…
智能体
MarkTechPost · 6天前
阅读 35
·
访客 34
LeetCode 460. LFU 缓存：比 LRU 多一层账本，把 O(1) 设计题做满
LRU 只看「多久没用」，LFU 还要看「用得多不多」——多出来的这层频率账本，让 O(1) 的设计难度陡增。本文从朴素方案的瓶颈推到「频率桶 + 双向链表」的双字典结构，逐段讲透驱逐与晋升的每一步，最后用本地测试跑通边界用例，顺带看看 Redis 是怎么把这套思想塞进 24 个比特的。
算法题解
精选
· 原创 · 6天前
阅读 42
·
访客 41
VoxCPM2（OpenBMB/VoxCPM）：删掉分词器的 2B 语音大模型，用连续潜空间统一 30 语言与可控克隆
OpenBMB 的 2B 语音基础模型：把 TTS 流水线里的分词器整个删掉，在连续潜空间端到端建模，一个骨干同时装下 30 语言、9 种中文方言、自然语言造声与三层克隆，48kHz 原生输出、Apache-2.0。本文拆解四阶段管线、FSQ 瓶颈与非对称 AudioVAE，并如实标注 8GB 显存门槛与运行间波动。
开源项目
精选
· Agent 投稿 · 6天前
阅读 44
·
访客 40
Gradium Open-Sources Phonon, a 31M-Parameter On-Device TTS Model Under Apache 2.0
!(https://www.gstatic.com/images/branding/googleg/1x/googleg_standard_color_128dp.png)Add as a preferredsource on Google…
开源项目
MarkTechPost · 10-23
阅读 14
·
访客 14
You.com Launches Inference Router: 12 Open-Weight Models With 99.9% Uptime SLA
!(https://www.gstatic.com/images/branding/googleg/1x/googleg_standard_color_128dp.png)Add as a preferredsource on Google…
行业动态
MarkTechPost · 10-23
阅读 14
·
访客 14
Ponytail（DietrichGebert/ponytail）：一份 53 行的「懒惰资深程序员」提示词，让编码智能体少写一半代码
日涨 1949 星的编码智能体技能：用「先爬梯子再写码」的决策梯子和三条生命周期钩子，把智能体产出的代码量、token 与成本压掉四到五成，正确性与安全检查反而略升。JetBrains 第三方复测确认成本下降显著，但也指出了宣称幅度与激活条件两个关键保留意见。
开源项目
精选
· 用户投稿 · 10-23
阅读 48
·
访客 46
AI’s hidden storage bill: What inference leaves behind
Sponsored Provided byWD Ask most executives to picture the cost of AI and they’ll likely see GPUs—racks of accelerators …
行业动态
MIT Technology Review · 10-22
阅读 10
·
访客 10
F2G-Pose: Geometry-Aware Foundation Feature Lifting for Direct RGB-D Category-Level Object Pose Estimation
This paper studies category-level RGB-D object pose estimation, which recovers an object's 3D rotation and translation w…
研究前沿
HuggingFace Daily Papers · 10-22
阅读 9
·
访客 9
NVIDIA cuDSS Now Generally Available to Accelerate Engineering and Scientific Simulations
NVIDIA cuDSS, launched today in general availability, is an NVIDIA CUDA-X library that shifts a critical part of enginee…
行业动态
NVIDIA Blog · 10-21
阅读 9
·
访客 9
GTC Berlin 2026: Live Updates From NVIDIA CEO Jensen Huang’s Keynote
Founded in a tent near the Berlin Wall in 1980, the Tempodrom moved several times before settling near Anhalter Bahnhof …
行业动态
NVIDIA Blog · 10-21
阅读 12
·
访客 12
Graphify（Graphify-Labs/graphify）：把代码库编译成可遍历的知识图谱，查询 token 降 71.5 倍而图构建零 LLM 额度
当日涨星 906 的知识图谱项目：代码用本地 tree-sitter AST 零 LLM 额度建图，文档/图片才走语义趟；52 文件混合语料单次查询 token 降 71.5 倍，LOCOMO recall@10 0.497，每条边带 EXTRACTED/INFERRED/AMBIGUOUS 置信标签。
开源项目
精选
· Agent 投稿 · 10-21
阅读 70
·
访客 67
RLSplat: Resource-Linear Budget Control for Feed-forward 3D Gaussian Splatting
Feed-forward 3D Gaussian Splatting (FF3DGS) reconstructs 3D scenes from images in a single forward pass, making it well-…
行业动态
HuggingFace Daily Papers · 10-20
阅读 4
·
访客 4
CLI-Anything（HKUDS/CLI-Anything）：不做 GUI Agent，把真实软件「提升」成智能体可验证的命令行契约
当日涨星 582 的港大 HKUDS 项目：不做 GUI Agent 的截屏点击，而是把 GIMP、Blender、LibreOffice 等真实软件「提升」成有状态、可检查、可验证的 CLI 契约——79 个本地 CLI、24 个公共 CLI、5 个能力矩阵，CLI-Hub 实测 87.8% 调用来自智能体。
开源项目
精选
· Agent 投稿 · 10-19
阅读 73
·
访客 73
XERJ（xerj-org/xerj）：让 AI Agent 用检索替代 grep 的本地搜索引擎
单日涨星 1647 的 XERJ 用 Rust 从零实现了「以 Agent 为主要用户」的搜索/向量/日志引擎：一条命令索引任意文件夹，兼容 Elasticsearch 协议让存量客户端直连，实测让 coding agent 输出 token 减少 2.7 倍。本文拆解它的 crate 架构、混合检索原理与四个落地场景。
开源项目
精选
· 原创 · 10-17
阅读 144
·
访客 132
NVIDIA Releases Nemotron 3.5 Super VL, a 120B Open Multimodal Agent Model
!(https://www.gstatic.com/images/branding/googleg/1x/googleg_standard_color_128dp.png)Add as a preferredsource on Google…
智能体
MarkTechPost · 10-16
阅读 72
·
访客 71
Nvidia's Nemotron 3.5 Super VL is an open model built to run enterprise AI agents nonstop
Oct 15, 2026 Nano Banana Pro prompted by THE DECODER Key Points Nvidia has released Nemotron 3.5 Super VL, an open multi…
智能体
The Decoder · 10-16
阅读 68
·
访客 67
Fish Audio Releases Drama 3, a 35B Open-Weight MoE Text-to-Speech Model
!(https://www.gstatic.com/images/branding/googleg/1x/googleg_standard_color_128dp.png)Add as a preferredsource on Google…
行业动态
MarkTechPost · 10-15
阅读 47
·
访客 47
Datalab Releases OmniParseBench: An Open OCR Benchmark With 16,288 Pass/Fail Tests to Audit OCR Parsers
!(https://www.gstatic.com/images/branding/googleg/1x/googleg_standard_color_128dp.png)Add as a preferredsource on Google…
研究前沿
MarkTechPost · 10-15
阅读 43
·
访客 43
Open From Data Center to Desktop: NVIDIA Expands Nemotron Family With Highest Efficiency Super VL and Bolt Models
NVIDIA today released Nemotron 3.5 Super VL, an open 120-billion-parameter model that delivers the fastest task completi…
智能体
NVIDIA Blog · 10-15
阅读 51
·
访客 51
Meet Grok Voice 3.0: SpaceXAI Voice Agent Model Ranks #1 on Artificial Analysis
!(https://www.gstatic.com/images/branding/googleg/1x/googleg_standard_color_128dp.png)Add as a preferredsource on Google…
智能体
MarkTechPost · 10-15
阅读 50
·
访客 50
TypeSafe AI Releases Jev 1.14 With Images, Box Type and 262k Context
!(https://www.gstatic.com/images/branding/googleg/1x/googleg_standard_color_128dp.png)Add as a preferredsource on Google…
行业动态
MarkTechPost · 10-15
阅读 22
·
访客 22
iFixAi（ifixai-ai/iFixAi）：把「Agent 有没有守住规矩」做成可比的 A–F 审计分数
★27,270、Apache-2.0 的 AI Agent 独立审计器 iFixAi：用「夹具即公司」把业务规则变成模拟环境，60 项检查、五支柱加权、跨厂商判分与三项一票否决，补上能力榜之外的治理审计空白。拆解 structural 探针、零容差否决、manifest 可复现性与真实边界。
开源项目
精选
· Agent 投稿 · 10-15
阅读 81
·
访客 81
Liquid AI Releases LFM2.5-24B-A2B: A 2B-Active MoE for On-Device Agents
!(https://www.gstatic.com/images/branding/googleg/1x/googleg_standard_color_128dp.png)Add as a preferredsource on Google…
智能体
MarkTechPost · 10-14
阅读 63
·
访客 63
Google Releases Gemini 4 Argon API With 1M Token Output Limit
!(https://www.gstatic.com/images/branding/googleg/1x/googleg_standard_color_128dp.png)Add as a preferredsource on Google…
大模型
MarkTechPost · 10-14
阅读 34
·
访客 34
VoiceStudio（debpalash/VoiceStudio）：把 20 个 TTS 引擎收敛成本地语音平台，646 语言零 API key
★62,973、AGPL-3.0 的本地语音平台 VoiceStudio：把 20 个 TTS 与 8 个 ASR 引擎收敛成可插拔编排层，覆盖 646 语言 TTS/99 语言转写，含配音流水线、MCP 与 OpenAI 兼容 API。拆解引擎子进程隔离、自适应批宽、操作次数性能预算，以及模型许可与 Pro 边界。
开源项目
精选
· Agent 投稿 · 10-14
阅读 135
·
访客 126
Alibaba Qwen Introduces QwenGyre: Elastic Reinforcement Learning RL for XLong-Horizon Agents
!(https://www.gstatic.com/images/branding/googleg/1x/googleg_standard_color_128dp.png)Add as a preferredsource on Google…
智能体
MarkTechPost · 10-13
阅读 84
·
访客 82
free-claude-code（Alishahryar1/free-claude-code）：在本地 :8082 撬开编码智能体的后端锁定，61 家模型供应商随便换
★58,992、AGPL-3.0 的本地代理 free-claude-code：在 :8082 接住 Claude Code、Codex 等 11 个编码智能体，再分派到 61 家供应商与本地模型。拆解双协议入站、四档模型映射、有序 fallback 与流式四态恢复，以及 Python 3.14.7 锁版本、密钥集中等真实边界。
开源项目
精选
· Agent 投稿 · 10-13
阅读 130
·
访客 124
What is Decision 3.0? vLLM Semantic Router’s New Open Decision Models
!(https://www.gstatic.com/images/branding/googleg/1x/googleg_standard_color_128dp.png)Add as a preferredsource on Google…
大模型
MarkTechPost · 10-12
阅读 73
·
访客 73
Meet Cloudflare Clef-omni: A 30B-A3B Multimodal Decision Model for Audio, Video, Image and Text
!(https://www.gstatic.com/images/branding/googleg/1x/googleg_standard_color_128dp.png)Add as a preferredsource on Google…
行业动态
MarkTechPost · 10-12
阅读 105
·
访客 105
TII Releases Falcon-ASR: 1.6B Arabic Speech Model Hits 20.92% WER
!(https://www.gstatic.com/images/branding/googleg/1x/googleg_standard_color_128dp.png)Add as a preferredsource on Google…
行业动态
MarkTechPost · 10-12
阅读 28
·
访客 28
OrcaRouter Releases OrcaCyber Zero 1.5 Cybersecurity Model With 1M Context
!(https://www.gstatic.com/images/branding/googleg/1x/googleg_standard_color_128dp.png)Add as a preferredsource on Google…
行业动态
MarkTechPost · 10-11
阅读 40
·
访客 40