vllm-skills

Use OpenAI-compatible API, integrate with clients

Contribution guide, model integration, debugging

Distributed inference, tensor/pipeline parallelism

Multi-hardware backend configuration (CUDA, ROCm, NPU, XPU, CPU)

Text model inference (Llama, Qwen, Mistral, GPT, etc.)

LoRA adapter serving, multi-LoRA deployment

Vision, audio, and video model inference

Offline inference, batch processing, script integration

Performance optimization, profiling, benchmarking

Model quantization (AWQ, GPTQ, FP8, INT8)

Deploy API service, production environment configuration

Install vLLM, configure environment, resolve dependency issues

V1 engine architecture, migration guide