Sublime
An inspiration engine for ideas
TwelveLabs | Home
twelvelabs.io
Clay | Go to market with unique data—and the ability to act on it
clay.com
Generative AI Lab
ai-analytics.wharton.upenn.edu



转:关于 DeepSeek 的研究和思考 (Archerman Capital)
关于这几天很火的 DeepSeek, 我们 (Archerman Capital) 做了一些研究和思考, 和大家分享, enjoy! 灰色部分是技术细节, 不感兴趣的可略过。
几个事实
1) DeepSeek 不是套壳不是蒸馏美国的大模型。 虽然中国有些大模型是套壳和蒸馏的, 但 DeepSeek 不是。
2) 核心架构还是基于 Transformer, deepseek 在架构、工程设计上进行了创新和工艺提升, 实现效率优化。架构上, 采用了混合专家模型 (MoE)、多头潜注意力 (MLA)、多令牌预测 (MTP)、长链式推理 (CoT)、DualPipe 算法等设计, 并进行了依赖强化学习 (RL)... See more

In this studio we're building a corrective RAG (CRAG) agentic workflow. It's powered by a locally running DeepSeek-R1 and has ability to search through your docs, evaluate the context quality and fallback to web search if it needs more info.
Build a corrective RAG agentic workflow using DeepSeek-R1 - a Lightning Studio by akshay-ddods
🌳 Galileo LLM Studio
Algorithm-powered LLMOps Platform
Find the best prompt, inspect data errors while fine-tuning, monitor LLM outputs in real-time. All in one powerful, collaborative platform.
Testing framework for LLM Part
Collective Change Lab
collectivechangelab.org
LinearB | Software Engineering Intelligence - Unlock Insights and Automations
linearb.io