This is where I’ve gathered all my implementations for machine learning in Python.
LlamaIndex-implementations:
- RAG (Retrieval-Augmented Generation): Enhanced LLM responses with real-time, accurate external information.
- AI Agents: Automated data tasks and develop advanced apps using LlamaIndex agents.
- Query and Chat Engines: Created stateful systems that maintain context for more engaging conversations.
- Streamlit Interfaces: Transformed Python scripts into interactive web apps with minimal code.
Langchain-implementations:
- Prompting Techniques: Utilized few-shot prompting, Chain of Thought, and ReAct prompting to guide model responses.
- Chat & Open Source Models: Explored conversational agents and community-driven model alternatives.
- Prompt Engineering: Involves crafting effective Prompts, using PromptTemplates (e.g., langchainub), and output parsers like Pydantic.
- Chaining Operations: Leveraged chains (e.g., create_retrieval_chain, create_stuff_documents_chain) to sequentially process tasks.
- Agents & Customization: Developed various agents (custom, Python, CSV, Agent Routers) for specialized data tasks.
- Tool Integration: Employed OpenAI Functions and tool calling within broader toolkits.
- Memory & Vectorstores: Integrated memory modules and vector stores (Pinecone, FAISS) for context management and data retrieval.
- RAG (Retrieval Augmentation Generation): Merged real-time external knowledge with LLM responses.
- Document Handling: Used DocumentLoaders and TextSplitters to process and manage text data.
- UI Development: Applied Streamlit for quick creation of interactive web interfaces.