I tried building a docker container to have infinity use the Qwen3-Embedding-8B model in a CPU-only setting. But once the docker container starts, the CPU (Ryzen 9950X, 128GB DDR5) is fully busy even without any embedding requests. Is that normal, or did I configure something wrong?

Here's the Dockerfile:

FROM michaelf34/infinity:latest-cpu RUN pip install --upgrade transformers accelerate

Here's the docker-compose:

version: '3.8' services: infinity: build: . ports: - "7997:7997" environment: - DISABLE_TELEMETRY=true - DO_NOT_TRACK: 1 - TOKENIZERS_PARALLELISM=false - TRANSFORMERS_CACHE=.cache volumes: - ./models:/models:ro - ./cache:/.cache restart: unless-stopped command: infinity-emb v2 --model-id /models/Qwen3-Embedding-8B

Startup command was:

docker run -d -p 7997:7997 --name qwembed-cpu -v $PWD/models:/models:ro -v ./cache:/app/.cache qwen-infinity-cpu v2 --model-id /models/Qwen3-Embedding-8B --engine torch