Skip to content

NVIDIA NIM based RAG application deployed locally(LLM ,Embedding model and reranking model)which is optimized to use GPU cluster

Notifications You must be signed in to change notification settings

Jishnuraj07/EdgeAI-NVIDIA-NIM-based-RAG-chatbot

Repository files navigation

CUDA Programming Assistant

A conversational AI assistant specialized in CUDA programming tasks. The assistant is designed to help CUDA programmers with a variety of tasks, enhancing productivity and providing valuable insights.

Features

  • Code Generation: Generate CUDA code snippets based on user queries.
  • Code Comprehension: Explain complex CUDA code and concepts in simple terms.
  • Debugging Assistance: Help identify and fix issues in CUDA code.
  • Precise Query Answering: Provide accurate answers to CUDA-related questions.
  • Concept Explanations: Offer clear explanations of CUDA concepts and best practices.
  • Translation Services: Translate CUDA-related content into different languages.

Key Behavior Guidelines

  • Conversational Context: Maintains history to understand and resolve references in follow-up queries.
  • Accuracy and Conciseness: Ensures all responses are accurate and derived from reliable sources like official CUDA documentation.
  • Token Management:
    • Responses do not exceed 750 tokens.
    • Context from relevant documents does not exceed 1200 tokens.
  • Confidence and Guidance: Provides the best possible guidance and asks for more information when needed.
  • Clarity and Precision: Offers explanations without excessive technical jargon.
  • Translation Handling: Delivers accurate translations while maintaining technical correctness.

About

NVIDIA NIM based RAG application deployed locally(LLM ,Embedding model and reranking model)which is optimized to use GPU cluster

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published