Add support for String input/output (triton-inference-server#60)

* x

* Fix string input tensor support

* Fix string list creation

* Add support for String output

* cleanup

* review edits

* Readme fixes

* review edits
- throw error during model loading

* follow up edits

* Fix typo in check

* Use verbose checking of output type to ensure appropriate error message is returned during model execution

* cleanup

* Wrap raw pointer to allow auto freeing of memory

* follow up review edits

Author: Hemant Jain

README.md

@@ -208,3 +208,11 @@ complex execution modes and dynamic shapes. If not specified, all are enabled by
   Setting the parameter `DISABLE_OPTIMIZED_EXECUTION` to "true" in the model
   configuration may help in some cases to avoid these negative interactions
   due to model specific caching and increase multiple instance performance.
+
+* PyTorch does not support Tensor of Strings but it does support models that accept
+a List of Strings as input(s) / produces a List of String as output(s). For these models
+Triton allows users to pass String input(s)/recieve String output(s) using the String
+datatype. As a limitation of using List instead of Tensor for String I/O, only for
+1-dimensional input(s)/output(s) are supported for I/O of String type.
+Batching is not allowed for PyTorch models with String I/O. For these models,
+the user must specify `max_batch_size: 0` in the configuration.