Dddred/qwen3-vl-2b-thinking-gguf-q4km-llamacpp-cpu

OpenAI-compatible endpoints:

• —POST /v1/chat/completions (supports stream: true via SSE)
• —GET /v1/models

Extra endpoints:

• —GET /health (fast, no model load)
• —GET /ready (forces model load and a tiny warmup)

Model (default):

• —Qwen/Qwen3-VL-2B-Thinking-GGUF / Qwen3VL-2B-Thinking-Q4_K_M.gguf

Notes:

• —VLM GGUF; this template supports optional mmproj/projector via env vars.