Skip to content

README.md

Latest commit

 

History

History
34 lines (28 loc) · 1.35 KB

README.md

File metadata and controls

34 lines (28 loc) · 1.35 KB
Raw

Model Characteristics Extractor

The Model Characteristics Extractor is responsible for extracting model/submodel specifications.

Model Characteristics Extractor

Web Service

To use the Model Characteristics Extractor it is suggested to run its container version by doing the following steps:

Build the Docker image

The following command builds the image and stores it in Dockerhub

docker buildx build --platform linux/amd64 -t gkorod/model_characteristics_extractor:v1.0 --push --no-cache

Run Docker Container

The following command initiates the Model Characteristics Extractor inside a Docker container. Model Characteristics Extractor requires to know the Profiler endpoint in order provide the answer of the profiling upon request.

docker run -d -t -p 7002:7002 -e PROFILING_SERVICE_URL=147.102.19.159:7001/api/profiling gkorod/model_characteristics_extractor:v1.0

Send Request

Model Characteristics Extractor expects this kind of input:

curl -X POST http://147.102.19.159:7002/upload \
  -F "file=@crf12345AndLab123MI3_640.onnx" \
  -F "model_input_size=12345" \
  -F "model_output_size=67890"

In the above command the model_input_size parameter is the size of the image expected in bytes. The model_output_size parameter is the expected output of the model in bytes.