i am using keras-ocr in my project.
to do the detection i need to create a pipeline with 2 models recognizer and detector as folowing:
pipeline = keras_ocr.pipeline.Pipeline(detector=detector, recognizer=recognizer) then use the recognize methode: predictions = pipeline.recognize(images=[image])
how to serve my pipeline using TF serving ?
Tensorflow Serving provides out-of-the-box integration with Tensorflow models, but can be easily extended to serve other types of models and data.
I went through keras-ocr docs but couldn’t find methods to save the model or pipeline to serve using TF serving.
If you can save the model in TF saved model format as shown here, then you can serve it using Tensorflow serving.