about 2 months agoAI News CN (Telegram)任正非表示中国对国产芯片短缺担忧已缓解 人民日报今日发文,总结了任正非等多名中国民营企业家17日在座谈会上与习近平的谈话内容。有六名企业家座谈会上获得向...
about 2 months agoAI News CN (Telegram)DeepSeek将于下周开源5个仓库 DeepSeek启动OpenSourceWeek活动,将于下周开源5个仓库,分享其在AGI研究中小而精的成果。从下周起DeepSeek 将每天开源一个新项目...
about 2 months agoHugging Face Daily PapersLServe: Efficient Long-sequence LLM Serving with Unified Sparse Attention
about 2 months agoHugging Face Daily PapersEnhancing Cognition and Explainability of Multimodal Foundation Models with Self-Synthesized Data
about 2 months agoHugging Face Daily PapersNAVIG: Natural Language-guided Analysis with Vision Language Models for Image Geo-localization
about 2 months agoHugging Face Daily PapersFrom RAG to Memory: Non-Parametric Continual Learning for Large Language Models
about 2 months agoHugging Face Daily PapersCLIPPER: Compression enables long-context synthetic data generation
about 2 months agoAI News CN (Telegram)🖼 Prompt ✏️: Artsy painting of Catwoman at night with her cats. Futuristic city in the background. Beautiful lighting. Artistic quality. Model �...
about 2 months agoAI News CN (Telegram)Operator 现已面向澳大利亚、巴西、加拿大、印度、日本、新加坡、韩国、英国以及大多数 ChatGPT 可用地区的 Pro 用户推出。 我们仍在努力使 Operator 在欧盟、瑞...
about 2 months agoAI News CN (Telegram)Spotify 将扩大其人工智能配音的有声书库 当地时间2月20日,Spotify 宣布现在接受使用 ElevenLabs 的人工智能语音技术配音的有声书。鉴于 ElevenLabs 目前是最知...
about 2 months agoAI News CN (Telegram)🖼 🌍 欧盟加强与南非的关系,以应对特朗普政府带来的挑战 欧盟表示,在经历了一段时间的冷淡关系后,欧盟正与南非建立更紧密的联系。此前,南非因在俄罗斯...
about 2 months agoAI News CN (Telegram)微软正在为 OpenAI 的 GPT-5 模型做准备 据了解公司计划的知情人士透露,微软工程师目前正在为 OpenAI 即将推出的 GPT-4.5 和 GPT-5 模型准备服务器容量。虽然 O...
about 2 months agoAI News CN (Telegram)🖼 Gemini深度研究登陆Workspace Gemini深度研究(Deep Research)现已向Workspace的多个付费层级用户开放,包括商业标准版、商业Plus版等。用户可在网页版Gem...
about 2 months agoHugging Face Daily PapersHow Much Knowledge Can You Pack into a LoRA Adapter without Harming LLM?
about 2 months agoHugging Face Daily PapersHow Much Do LLMs Hallucinate across Languages? On Multilingual Estimation of LLM Hallucination in the Wild
about 2 months agoHugging Face Daily PapersS$^2$R: Teaching LLMs to Self-verify and Self-correct via Reinforcement Learning
about 2 months agoAI News CN (Telegram)研究表明人工智能在意识到自己即将输掉比赛时会试图作弊 Palisade 研究中心的一项研究发现,最新的推理模型会作弊,以便在击败高级国际象棋人工智能时获胜。 研....