Env-Doctor

소개

Env-Doctor addresses the primary pain point in GPU computing: persistent CUDA version mismatches between NVIDIA drivers, system toolkits, cuDNN, and popular Python AI libraries like PyTorch and TensorFlow. It provides a swift, one-command diagnosis to pinpoint exact conflicts and offers actionable solutions, including platform-specific CUDA installation guides and safe `pip install` commands tailored to your system's configuration. Beyond basic checks, Env-Doctor validates Dockerfiles, assesses AI model VRAM compatibility, and seamlessly integrates into CI/CD workflows and AI assistants via its Model Context Protocol (MCP) server, ensuring a robust and functional GPU environment for AI development.

주요 기능

• One-Command GPU Driver, CUDA, cuDNN, and AI Library Compatibility Check
• Provides Safe `pip install` Commands for AI Libraries (PyTorch, TensorFlow, JAX)
• Generates Platform-Specific CUDA Toolkit Installation Guides
• Checks AI Model VRAM Requirements Against Your GPU
• 104 GitHub stars
• Validates Dockerfiles for GPU Configuration and AI Library Issues

사용 사례

• Empower AI assistants (e.g., Claude Desktop) to diagnose and resolve GPU-related problems via MCP server integration.
• Rapidly debug and fix GPU and CUDA environment issues in local AI development.
• Automate GPU environment validation and compatibility checks in CI/CD pipelines.