https://github.com/hesamsheikh/ai-researcher-agent

https://github.com/hesamsheikh/ai-researcher-agent

Last synced: 8 days ago
JSON representation

• Host: GitHub
• URL: https://github.com/hesamsheikh/ai-researcher-agent
• Owner: hesamsheikh
• Created: 2024-05-24T22:50:55.000Z (8 months ago)
• Default Branch: main
• Last Pushed: 2024-06-01T22:33:27.000Z (7 months ago)
• Last Synced: 2024-06-01T23:49:45.597Z (7 months ago)
• Language: Jupyter Notebook
• Size: 2.2 MB
• Stars: 0
• Watchers: 1
• Forks: 1
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
## 🚀Let's build 𝐀𝐈 𝐀𝐠𝐞𝐧𝐭𝐬 to read a research paper and write a blog post for it: (100% 𝐥𝐨𝐜𝐚𝐥𝐥𝐲) 

We will create AI Agents that:

📖 Read a scientific paper

✍ Write a blog post about it

🏷️ Choose a great title for the post

I will use Lightning AI for fast and easy development and Ollama for running AI 100% locally, no API is needed! 

We will use OpenHermes, a state-of-the-art model fine-tuned on Mistral,

🪶 It is lightweight, (7B)

👍 It works great with agentic instructions

To setup our LLM:

- Setup Ollama,

- Pull the model,

- Initiate it.

Let's create a tool for reading the paper in PDF format and to return the contents of it. Our Agent will use this tool later... 📄

We create three agents:

🤖 Reads paper

🤖 Write a post about it

🤖 Come up with a title

We clearly explain the role of our agent, and its goal, and adding a backstory also helps with it playing as we would like. 

We define three tasks and assign each of them to an agent. Each task needs:

📝 description: A concise description of what needs to be done

👤 agent: Which agent is responsible for it

🏁 expected_output: The output we would like to receive

Putting all the Agents and Tasks together,

Next, let's kick off! 🚀

The Agents 🕵️‍♂️ read through the research paper "Attention is All You Need", and came up with a title 📝 and a post article for the blog! 📰

I have published the code as a Lightning AI studio,

Everything is ready for you to see and run,

and you can use free resources.