|| Web & ML 开发人员
|| 专注于 AI 研究、开发和学习
|| 分享 AI 趋势、论文和优秀产品
|| Web & ML Developer
|| Focus in AI research & learning
|| Share AI trends, papers, great products
AI 开源项目推荐
WeChatFerry - by Changhua
微信机器人底层框架,可以实现微信信息自动收发。结合大语言模型、多模态大模型和图像、音乐、视频生成模型,就可以实现对多模态信息的理解和输出,十分有想象力!
可接入 Gemini、ChatGPT、Claude、Groq、Llama-3、Yi-01
Only 3 days after LLAMA 3 is release, we've seen 3 Chinese fine-tune models available on
@huggingface
China Unicom:
ShareAI:
Blossom:
Great work everyone! Which model do you like most?
AI 开源项目推荐
Verba - 黄金 RAG 检索者 🪙 by
@weaviate_io
Verba 是一个完全可定制的个人助理,由 Weaviate 开发并开源,用于查询和与数据进行交互,无论是本地还是通过云端部署���解答有关文档的问题,交叉引用多个数据点,或从现有知识库中获得洞见。Verba 结合了最先进的 RAG 技术与 Weaviate
RAGCache: Efficient Knowledge Caching for Retrieval-Augmented Generation
Presents a novel multilevel dynamic caching system that efficiently caches and shares intermediate states of retrieved documents in RAG for LLMs.
📝
AI 开源项目推荐
🕷️ ScrapeGraphAI: You Only Scrape Once
by
@scrapegraphai
ScrapeGraphAI 是一个用于网络爬虫的开源 Python 库,使用大型语言模型(LLM)、Langchain 和 RAG 使网络爬虫变得更容易。
它使用 LLM 和直接图形逻辑来为网站、文档和 XML
Alongside our Mixtral 8x22B release, we are releasing our tokenizers, which go beyond the usual text <-> tokens, adding parsing of tools and structured conversation.
Repo:
Guide:
A Survey on Retrieval-Augmented Text Generation for LLMs
Presents a comprehensive overview of the RAG domain, its evolution, and challenges.
It includes a detailed discussion of four important aspects of RAG systems: pre-retrieval, retrieval, post-retrieval, and generation.
If
最近关注的一个 AI Agent 产品:MultiOn
官方介绍:可以使用自然语言自动执行网络任务的下一代 AI Agent
Agent API 地址:
Github 地址:
官方和创始团队信息:
@DivGarg9
@omarshaya
@MultiOn_AI
更多产品和 API 信息继续在下面更新 👇👇
📣📣 Super proud to present the most exciting project of my PhD so far: “HippoRAG: Neurobiologically Inspired Long-Term Memory for Large Language Models”.
HippoRAG, as the title suggests, is a brain-inspired RAG framework that enables LLMs to effectively and efficiently
ToDoList 产品推荐
Superlist - Home to all your ✅ lists
Superlist
@Superlist
可能是我用过颜值最高、交互最舒服的 ToDoList 产品了(可以开着声音用,不同的交互还有不同的音效)
创始团队是著名的 Wunderlist 核心团队,Wunderlist 被微软收购后,数据合并到 Microsoft To Do 中。
创始人
AI 开源项目推荐
fabric: 可以让你的生活自动化的开源 AI 框架
-
@DanielMiessler
fabric 是一个利用 AI 增强人类能力的开源框架。它提供了一个模块化框架,可以使用一组可在任何地方使用的众包AI Prompts 来解决特定问题。
fabric 出发点:
1. 将问题分解成单个部分,然后逐个应用 AI
2.
AI 文章推荐
Firecrawl Blog - by
@mendableai
Firecrawl 推出了博客,主要会分享:
- 从网络提取数据🌐
- Agent 实验🧪
- 智能抓取、爬取、提取🕸️
- AI 网络搜索🔎
- LLM 数据预处理📊
以最新一篇:
Build a 'Chat with website' using Groq Llama 3
Launching our Firecrawl🔥 blog today!
Join us as we share tutorials and announcements on:
- Extracting data from the web 🌐
- Agent experiments 🧪
- Smart scraping, crawling, extraction 🕸️
- RAG optimization 🦜
- Web search for AI 🔎
- Data pre-processing for LLMs 📊
Happy to share Browserbase with the world today.
We help AI applications browse the web.
And we just raised $6.5 million to do it.
Now, we're opening signups to developers everywhere.
I can't wait to see what you 🅱️uild.
The Prompt Report
A Systematic Survey of Prompting Techniques
Generative Artificial Intelligence (GenAI) systems are being increasingly deployed across all parts of industry and research settings. Developers and end users interact with these systems through the use of
AI 开源项目推荐
Chat with your data - Solution accelerator by
@Azure
本开源项目是在 Azure 中运行的 RAG 模式的解决方案加速器,使用 Azure AI 搜索进行检索,并使用 Azure OpenAI 大语言模型来支持 ChatGPT 风格和问答体验。这包括最常见的要求和最佳实践。
项目主要亮点:
· 私有 LLM
Github - TurboSeek
TurboSeek 是由 Together AI
@togethercompute
开发的免费开源的 AI 搜索引擎,主推更快速更智能的搜索。
主要技术栈:
· Next.js app router with Tailwind
· Together AI for LLM inference
· Mixtral 8x7B & Llama-3 for the LLMs
· Bing for the search API
· Helicone for
树莓派 AI Kit
@RaspberryPi_org
@Hailo_ai
树莓派 AI Kit 与 Hailo 合作开发,提供了一种将本地、高性能、节能推理集成到各种应用程序中的便捷方式,价格为 70 美元。
AI Kit 包括 M.2 HAT+,它预装有一个 Hailo-8L AI 加速模块。安装在树莓派5 上,AI Kit 允许你快速构建复杂的 AI
excited to announce the release of the new version of . Key updates include:
- Introduction of the RAG feature, employing advanced retrieval algorithms to improve response accuracy (you can add pdf file up to 10 MB, website content and texts)
- PRO
AI 开源项目推荐
Jina - 使用云原生技术构建多模式 AI 应用程序
20.2K 🌟 Apache-2.0 license -
@JinaAI_
Jina 可让用户构建通过 gRPC、HTTP 和 WebSockets 进行通信的多模式 AI 服务和管道,然后对其进行扩展并部署到生产环境中。用户可以专注于逻辑和算法,而不必担心基础设施的复杂性。
📽️ New 4 hour (lol) video lecture on YouTube:
"Let’s reproduce GPT-2 (124M)"
The video ended up so long because it is... comprehensive: we start with empty file and end up with a GPT-2 (124M) model:
- first we build the GPT-2 network
- then we optimize
Introducing Gemma with a 10M context window
We feature:
• 1250x context length of base Gemma
• Requires less than 32GB of memory
• Infini-attention + activation compression
Check us out on:
• 🤗:
• GitHub:
• Technical
I think AI agentic machine translation has huge potential for improving over traditional neural machine translation, and am releasing as open-source a demonstration I'd been playing with as a fun weekend project.
Using an agentic workflow, this demonstration (i) Prompts an LLM
Announcing v1 of the Together Python SDK!
◆ More intuitive OpenAPI compatible API
◆ Async support for batching requests
◆ More robust with better error handling
AI 产品&团队推荐
Unsloth - Easily finetune & train LLMs Get faster with unsloth
@UnslothAI
LLM 微调和训练平台,更快地训练和推理速度、更少的显存占用、算力托管更新、开源。
平台 UI & UX 极其可爱,两兄弟 Daniel Han & Michael Han 创建!
@danielhanchen
My Sunday project: "training" Python code with an LLM:
The problem:
I thought it would be neat if an LLM could go through all my emails and tell me all the places I've traveled to in the world by extracting the destinations from the flight itineraries.
We've been in the kitchen cooking 🔥 Excited to release the first
@AIatMeta
LLama-3 8B with a context length of over 1M on
@huggingface
- coming off of the 160K context length model we released on Friday!
A huge thank you to
@CrusoeEnergy
for sponsoring the compute. Let us know