Object_Detection_Python Model Server
The object_detection_python model server is a simple FastAPI application written specifically for use in the object_detection recipe with "DEtection TRansformer" (DETR) models. It relies on huggingface's transformer package for AutoImageProcessor and AutoModelforObjectDetection to process image data and to make inferences respectively.
Currently, the server only implements a single endpoint, /detection, that expects an image in bytes and returns an image with labeled bounding boxes and the probability scores of each bounding box.
Build Model Server
To build the object_detection_python model server image from this directory:
podman build -t object_detection_python . base/Containerfile
or
make build
Download Model(s)
You can download models from huggingface.co for this model server. This model server is intended to be used with "DEtection TRansformer" (DETR) models. The default model we've used and validated is facebook/detr-resnet-101.
You can download a copy of this model into your models/ with the make command below.
make download-model-facebook-detr-resnet-101
or any model with
cd ../../models/ && \
python download_hf_models.py -m <MODEL>
Deploy Model Server
The model server relies on a volume mount to the localhost to access the model files. It also employs environment variables to dictate the model used and where its served. You can start your model server using the following make command from the model_servers/object_detection_python directory, which will be set with reasonable defaults:
make run
or
podman run -it -d -p 8000:8000 \
-v <LOCAL/PATH/TO/MODLES>/facebook/detr-resnet-101:/models/facebook/detr-resnet-101:ro \
-e MODEL_PATH=/models/facebook/detr-resnet-101 \
-e HOST=0.0.0.0 \
-e PORT=8000 \
ghcr.io/ai-lab/model_servers/object_detection_python:latest