TensorFlow Serving 2 preinstalled for AWS. The official tensorflow/serving Docker image with the canonical half_plus_two sample model, an nginx basic-auth gateway on port 80, and per-instance credentials on first boot. Backed by 24/7 cloudimg support.
## TensorFlow Serving 2 by cloudimg
TensorFlow Serving 2 is Google's high-performance, flexible model server for TensorFlow SavedModels. This Amazon Machine Image delivers TF Serving 2 running under Docker behind an nginx basic-auth gateway, so a production-grade model inference endpoint is running within minutes of launch.
## Application Stack
The image runs the official tensorflow/serving 2 CPU Docker container managed by docker compose v2 and supervised by systemd. REST predictions are served on port 8501, gRPC predictions on port 8500. An nginx reverse proxy on port 80 fronts the REST API and enforces HTTP Basic authentication.
## Sample Model Included
Google's canonical half_plus_two SavedModel is bundled on a dedicated 20 GiB model volume, so the server has a working model on first boot. Replace it with any TensorFlow SavedModel by dropping a new versioned directory under /var/lib/tfserving/models/.
## Secure First Boot
On the first boot of every instance a one-shot service generates a fresh nginx password unique to that instance and writes credentials to a root-only file.
## cloudimg Support
cloudimg provides 24/7 technical support for TF Serving deployment, model upgrades, gRPC and REST integration, nginx hardening and TLS termination.