Please go to demo website for more details.

Sample applications that can be served and profiled by ML Model CI.

The web application contains two parts (tabs):

• BERT Application on Descriptive Text Sentiment Analysis
• Mask R-CNN Application on Image Object Detection

Click the following links to download the models.

• BERT, IMDB Dataset
• Mask R-CNN.

You can build a script to start the server or using command line tool.

Please refer to the ModelCI doc.

By using the serve.py in ModelCI, we can start the inference server easily, but you need to register the model first.

python serving.py name --m MRCNN -f tensorflow -e tfs --device cuda:1

The same as BERT model.

You need to modify the API address in the application source code, to start the services.

Address location:

• BERT
• Mask R-CNN

After all of these, you can start the web application to see the serving and inference results by:

npm install
npm start

BERT Application on Descriptive Text Sentiment Analysis	Mask R-CNN Application on Image Object Detection

For more details about the screenshots.