1. Prepare a pre-mturk method for downloading the training data dynamically (both annotated and not).
2. Improving upon the method mentioned in our IEEE paper.
Running for development
> running for development
-----------------------
If you'd like to reproduce our results, we encourage you to use the demo online or look at our `API spec`_ for raw ``json`` results. Alternatively, you can clone this repository and either call the scoring function or run the demo server yourself. We wrote `some instructions to help you`_.
...
...
@@ -29,7 +31,7 @@ If you'd like to reproduce our results, we encourage you to use the demo online