Describe the bug
The Unity usage example assumes that user has Rantionary.rpkg file in assets/resources folder but there is no step by step information on how to import those and from what source.
The Rant3.zip from github doesn't contain said file.

To Reproduce
Steps to reproduce the behavior:

1. Follow the Running in Unity tutorial
2. Run the example in Unity
3. See the erro

1. Add CI test for building documentations (Do not ignore warnings and add spellcheck).
2. Fix docstrings with incorrect/inconsistent Sphinx format. Currently, such issues are treated as warnings in the docs building.

https://github.com/minimaxir/gpt-2-simple/blob/ca6bc61d958fd4c474af9a412ace27279b88dd90/gpt_2_simple/src/encoder.py#L8

According to the docs the lru_cache decorator uses memozation to speed up a function call when it's called with the same arguments.

But this function doesn't need any arguments...

The generated article already uses a headline as a prompt, but it ignores anything already written in the article section when generating. I think it would be a good addition to allow users to write the first paragraph or so of the article in addition to the headline -- perhaps as a separate field? Is this possible?

• TransformerDecoder.forward: where does self.training come from?
  https://github.com/asyml/texar-pytorch/blob/d17d502b50da1d95cb70435ed21c6603370ce76d/texar/torch/modules/decoders/transformer_decoders.py#L448-L449

• All arguments should say their types explicitly in the docstring. E.g., what is the type of infer_mode? The [method signature](https://texar-pytorch.readthedocs.

谢谢分享
@imcaspar 能写一下怎样使用吗?

1. 使用该模型继续进行预训练的步骤
2. 使用该模型进行inference的步骤

谢谢