NIM Interactive Chat Launcher
Paste the following prompt into your AI chat to install this skill:
Please follow https://skillhub.cn/install/skillhub.md and install @user_93ce1c9f/nim-interactive-chat.
About this skill
Problem
When deploying NVIDIA NIM models locally, the workflow is often split into several manual steps: pulling the image, creating a container, mapping ports, confirming the service is healthy, and then calling an OpenAI-compatible API to chat. Stale containers, incomplete readiness checks, or premature requests can lead to port conflicts, connection failures, and interrupted sessions.
How It Works
- Generated scripts: creates a start script, a stop script, and a Python interactive chat client, reducing manual launch logic.
- Container management: removes an existing same-name container before starting, helping avoid container-state conflicts.
- Readiness handling: waits for health checks to pass before entering the chat flow.
- API integration: uses an OpenAI-compatible interface with streaming output and session context.
- Resource control: supports
gpu-memory-utilization, with a default value of0.9.
Boundaries
It is best suited for running NGC, TGC, or other NVIDIA NIM images in a local Docker environment for quick chat validation. Exiting the client does not necessarily stop the container; a stop script should be used when done. It assumes Docker, Python 3, and the openai package, and the image, model name, and port mapping must match the local setup.
Use Cases
- After running a local Qwen3.5 NIM Docker image, clean stale containers and wait for health checks before starting.
- When debugging an OpenAI-compatible endpoint, use the generated Python client to send consecutive messages and verify streaming and context.
- When validating GPU usage, set memory utilization, start the NIM container, and confirm it responds in chat.
- When taking over a local model demo environment, manage the container with start and stop scripts instead of missing manual cleanup.
Best For
- AI engineers deploying local LLMs who want to avoid hand-writing Docker launch scripts and enter NIM model chat directly.
- Model service testers who need to verify OpenAI-compatible streaming output, context retention, and health-check readiness.
- GPU operations engineers who want to control local NIM container memory usage with the utilization parameter.
- Model demo leads who need start and stop scripts to quickly demonstrate local Qwen or other NIM models.
Related Skills
An AI workflow that structures personal knowledge bases, IMA, Yuque, and Feishu content into approved personas for paid WeChat mini-agent deployment.
Local workflow memory with matching and SOP updates.
An OpenClaw live streaming executor that initializes TRTC streaming, starts a real-time dashboard, generates viewer URLs, and continuously reports live events.
Breaks down physical supply chains for super-trends to identify second- and third-layer bottlenecks, runs valuation and reverse checks, and maintains trackable reports.