TODOs

This is an umbrella issue where we can collectively tackled some problems and improve general open source reading comprehension quality.

**Goal**
The network is already there. We just need to add more features on top of the current model.
- [x] Implement full features stated in the original paper
- [ ] Achieve EM/F1 performance stated in the original paper with a single model settings

**Model**
- [x] Increase the hidden units to 128. #15 reported performance increase when the hidden units increased from 96 to 128
- [ ] Increase the number of heads to 8
- [ ] Add dropouts in better locations to maximize regularization
- [ ] Train "unknown" word embedding

**Data**
- [ ] Implement paraphrasing by back-translation to increase the data size

Contribution to any of these issues is welcome and please comment on this issue and let us know if you want to work on these problems.

Start by reviewing the unchecked Model and Data items against the original paper and the repository's current QANet implementation. Choose one specific item, such as additional attention heads, dropout placement, unknown-word embeddings, or back-translation, and define completion through implementation and comparison with the paper's EM/F1 results.

Written by the indexing model from the issue text.