CODEWITHSUNDEEP
pythontensorflowkerastensorflow-serving
Ophir Yoktan
Ophir Yoktan

Reputation: 8449

Does tensorflow serving work with more than one input (or output)?

According to: https://blog.keras.io/keras-as-a-simplified-interface-to-tensorflow-tutorial.html#exporting-a-model-with-tensorflow-serving

Any Keras model can be exported with TensorFlow-serving (as long as it only has one input and one output, which is a limitation of TF-serving)

Is this (still) true? I couldn't find other references

Upvotes: 3

Views: 2641

Answers (1)

Burkhard Neidecker-Lutz
Burkhard Neidecker-Lutz

Reputation: 161

You can have multiple input and output tensors and expose any combination of them in different signatures for the SavedModel, as long as you use the PREDICT interface of TF-serving. For examples, look at

https://www.tensorflow.org/serving/signature_defs

So while there may be a limitation in the Keras-exporter code as written (I haven't looked), there is no such limitation in SavedModel on TensorFlow Serving (the product I am working on regularly uses the feature)

Upvotes: 2

Related Questions