How to use the saved_model.pb to make prediction on python ?

On running the code through the method provided in Running Script, saved_model.pb & variables folder is obtained. I was able to make a prediction using

saved_model_cli run \
    --dir ./ \
    --tag_set serve \
    --signature_def predict \
    --input_examples 'examples=[{"2":[0.5348837209302325],"3":[0.75],"4":[33.68298368298368],"5":[14.4054],"6":[1707.0],"8":[1.0],"11":[1.0],"14":[1.0],"15":[1.0]},{"2":[0.3409090909090909],"4":[50.0],"5":[0.8047000000000001],"6":[833.0],"12":[1.0]}]'

The above CLI command contains 2 test cases provides me with 2 scores. I want to know a method wherein I could load the saved model on python or host it using docker on tf-serving & then make predictions on multiple inputs. My data is of the form of LibSVM with multiple features on each entry as shown below.

1 qid:0 2:0.6842105263157895 4:29.1951604418727 5:1.4356 6:1346 8:1 11:1 14:1 15:1
0 qid:0 2:0.8125 3:0.8571428571428571 4:32.55459624174708 5:35.69730000000001 6:1328 8:1 11:1 14:1 15:1

[hdatteln]

I am also a bit stuck on this. However, I am using TFRecords as training/test input data; It would be great to have a bit more information in this example notebook on how to save & load the trained model and make predictions: https://github.com/tensorflow/ranking/blob/master/tensorflow_ranking/examples/handling_sparse_features.ipynb

[ramakumar1729]

I am also a bit stuck on this. However, I am using TFRecords as training/test input data; It would be great to have a bit more information in this example notebook on how to save & load the trained model and make predictions: https://github.com/tensorflow/ranking/blob/master/tensorflow_ranking/examples/handling_sparse_features.ipynb

@hdatteln : Created a separate issue for this. #104