Tag: #OpenAI 호환 API
Kubernetes, DevOps and MLOps, written from things we ran and measured ourselves · 1 posts
Can a custom-built embedding server be swapped out for a Dynamo embedding worker?
We deployed the same model used by our operational embedding server (a custom FastAPI server) onto a Dynamo embedding worker and compared them by sending identical inputs. The vectors matched (cosine similarity of at lea
2026-11-15 · 11 min read #NVIDIA Dynamo#vLLM#임베딩#Qwen3-Embedding#쿠버네티스