This repository offers a comprehensive collection of end-to-end Generative AI projects, focusing on Large Language Models (LLMs), Retrieval-Augmented Generation (RAG), AI Agents, and multimodal applications. It targets developers and researchers seeking practical, industry-relevant examples with deployment considerations, providing a valuable resource for learning and implementing cutting-edge AI solutions.

How It Works

The project showcases a wide array of LLM applications, leveraging popular frameworks like Langchain, LlamaIndex, Haystack, and CrewAI. It demonstrates various techniques including RAG for knowledge retrieval, AI agents for task automation, fine-tuning LLMs with custom datasets (e.g., LoRA, QLoRA), and multimodal capabilities. The implementations often utilize cloud-based LLMs (Google Gemini, OpenAI GPT, Llama 3) and open-source models (Mistral, Llama 2, Gemma), with deployment examples using Streamlit and Hugging Spaces.

Quick Start & Requirements

• Installation: Typically involves cloning the repository and installing Python dependencies via pip. Specific project requirements vary.
• Prerequisites: Python 3.x, potentially specific versions of libraries like Langchain, LlamaIndex, Hugging Face Transformers, PyTorch. Some projects may require API keys for services like OpenAI or Google Gemini. GPU acceleration is often beneficial or required for fine-tuning and efficient inference.
• Resources: Setup time and resource requirements depend heavily on the specific project. Running larger models locally may necessitate significant RAM and GPU VRAM.

Highlighted Details

• Extensive coverage of RAG implementations with diverse vector databases (FAISS, ChromaDB, Qdrant, Objectbox).
• Numerous examples of AI agent frameworks (Autogen, Langraph, CrewAI, Phidata) and their applications.
• Demonstrations of LLM fine-tuning techniques (LoRA, QLoRA, ORPO, DPO) and quantization methods (GGUF, AWQ).
• Projects span various domains including finance, healthcare, and content generation, with multimodal capabilities.

Maintenance & Community

The repository is maintained by GURPREETKAURJETHRA. Community contributions are welcomed, with a call for stars and ideas. Links to community platforms are not explicitly provided in the README.

Licensing & Compatibility

The repository is distributed under the MIT License, permitting commercial use and closed-source linking.

Limitations & Caveats

The README presents a broad list of projects without detailed setup instructions or individual project documentation for each. Users may need to consult external resources for specific project dependencies, configurations, and troubleshooting. The sheer volume of projects might imply varying levels of maturity and maintenance for individual implementations.