July 1, 2016

Report Abuse
This is a example of not working webservice although the experiment was running smoothly with no error at all
This service takes input as string, which represents for a feature vector using svmlight format as follows: # [class] [feature_id:feature_value] [feature_id:feature_value] ... such as 1 2609:0.061183 3822:0.005865 4876:0.007535 5612:0.058160 5652:0.005863 The issue is that with a given input, the service is running without any issue from the Azure ML Studio However, when publishing as a web service, it is failed from the test function as well as from the response/request call. I have tested with an input like this (included the double quote) from the test function. "1 2609:0.061183 3822:0.005865 4876:0.007535 5612:0.058160 5652:0.005863" The error shows that the input data has been taken and send to the Python module correctly, but still don't know why it still fails.