[Enhancement] Consolidate the training and inference launch scripts into separate folders.

[shijieliu]

Is your feature request related to a problem? Please describe.
Currently in hstu example, the training and inference launch scripts are placed under the same directory and it is hard for new users to find where to start.

Describe the solution you'd like
It is expected that benchmark, training/inference launch scripts, related configs and readme under examples/hstu will be consolidated and placed under examples/hstu/training and examples/hstu/inference seperately. Other codes should be kept unchanged to avoid unnecessay file change.

The final structure will be like:

- examples/hstu
   - ...
   - utils
     - gin_config_args.py
     - ...
   - tmp_data # folder to place processed data, this should be aligned with preprocessor.py
   - training
      - benchmark
         - readme.md
         - training benchmark scripts
      - configs
      - readme.md
      - pretrain_gr_ranking.py
      - ...
   - inference
      - benchmark
         - readme.md
         - inference benchmark scripts
      - configs
      - readme.md
   - readme.md
   - preprocessor.py
- examples/commons

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, code examples, or references to existing implementations about the feature request here.

---

By submitting this issue, you agree to follow our code of conduct and our contributing guidelines.

[geoffreyQiu]

Create a base PR #190 for further README reorganization.
WIP in PR #202 for further work.