📸 Disaster reporting w/Open Source LLaVA/AI for photo analysis

❔ Context

My country recently faced severe events:

🤔 Inception

These events generated a lot of visual contents, especially photos at an amazing pace and from a very wide variety of media sources.

These contents generated a lot of emotions while we were lacking of sleep. Still we need to be able to perform analysis.

The purpose of this post is to show a way to analyze these contents, with Open Source software... and what we can learn from this technology in a context of social crisis and emotion overwhelm.

🎯 What we'll cover

Below some keypoints I'll cover:

🎬 Demo

📑 Demo keypoints

• 💡 Intro
• 🪝 Pitch
• 📜 Project workflow
• 🤖 First image analysis
• 📕 Extracting keywords with open source AI : llama3
• 🔭 Teasing on text2image perspectives

🔖 Resources

• 📜 Kaggle Notebook
• 🆓 Local & Open Source AI: a kind ollama & LlamaIndex intro
• 💰 Unleash ollama potential w. data and LLM capabilities
• 📑 Ollama Series' Articles