Hi,
I'm just started to use sagemaker.
Now, I'm testing encoder-decoder model for regression.
The model is coded in Sagemaker's script mode, and finished learning on a Jupyter notebook.
The learning code is written by keras + tensorflow, and the model is based on the encoder decoder model. It is similar to the Keras's seq2seq example linked below.
Below, I will take this as an example.
https://keras.io/examples/lstm_seq2seq/
In the above model, there are "encoder_model" and "decoder_model" apart from "model" to be trained, and in the inference, "encoder_model" and "decoder_model" are used to generate the prediction by the function "decode_sequence (input_seq)".
I would like to deploy this function "decode_sequence (input_seq)" as an endpoint, but it doesn't work as usual with estimator.deploy () and I don't know how to implement it.
Is there any sample code or resources to solve this?
Thanks in advance.