A recipe directory and improved documentation with input/output examples for all tasks

**Is your feature request related to a problem? Please describe.**
Yes, this feature request is related to a lack of clarity in the documentation for the data labeling tool. The tool supports different types of labeling tasks like bounding box annotation, semantic segmentation, image classification, intent classification and slot filling, response selection, named entity recognition, etc. However, there are no examples provided of the expected input/output data formats for each of these task types and where examples do exist they are buried in the website.

As a user, I have to guess or dig through the code to understand the format. For example, for intent classification and slot filling, what is the expected JSON input and output? These are very basic pieces of information that would help me get started with the tool much more quickly.

Providing an /examples directory in the repo with actual example input and output files and or a recipes similar to a "model zoo" would be helpful

**Describe the solution you'd like**
Having a robust "examples" or "recipes" section on your documentation as well as a directory in the repo with these types of examples for each of your labeling tasks would greatly improve the usability and accessibility of this tool for new users.

**Describe alternatives you've considered**
A clear and concise description of any alternative solutions or features you've considered.

**Additional context**
Add any other context or screenshots about the feature request here.