Fix Tapas bug due to exceeding the maximum rank value

[vankov]

A bug fix for the Tapas model.

Description

The problem is that the maximum value rank in Tapas is limited to 255 and if the text is long enough you may end up with higher ranks. The solution is to replace ranks higher than 255 with 255.

Motivation and Context

The issue was reported by a user in the slack channel. If you feed a long text (i.e table) you may end up with ranks higher than 255 and this is the max value allowed by the model (perhaps they used a byte variable to represent it or a static array). Such ranks case a Tensorflow exception.

How Has This Been Tested?

I first reproduced the problem by running Tapas on long enough texts and made sure that the fixes solves the problem

Screenshots (if appropriate):

Types of changes

• Bug fix (non-breaking change which fixes an issue)
• Code improvements with no or little impact
• New feature (non-breaking change which adds functionality)
• Breaking change (fix or feature that would cause existing functionality to change)

Checklist:

• My code follows the code style of this project.
• My change requires a change to the documentation.
• I have updated the documentation accordingly.
• I have read the CONTRIBUTING page.
• I have added tests to cover my changes.
• All new and existing tests passed.