Please read the topic category description to understand what this is all about
Description
Many email and word processing applications can now automatically detect and correct common grammatical errors as you write. For example, the sentence “I am doing fine. How is you?” might be corrected to “I am doing fine. How are you?”. The goal of this project is to fine-tune a Transformer model to be able to automatically provide these corrections, similar to how Grammarly does
Model(s)
This task can be viewed as a sequence-to-sequence task, so models like T5 would be a great starting point
Datasets
Challenges
If you use T5, you’ll need to define a suitable prefix for the text-to-text task. You’ll also need to think about suitable metrics for the evaluation.
Desired project outcomes
- Create a Streamlit or Gradio app on
Spaces that can automatically provide suggestions to improve the grammar of some input text. Check out Grammarly for some inspiration on the visualization side.
- Don’t forget to push all your models and datasets to the Hub so others can build on them!
Additional resources
- https://towardsdatascience.com/fine-tune-a-transformer-model-for-grammar-correction-b5c8ca49cc26
- https://github.com/PrithivirajDamodaran/Gramformer
Discord channel
To chat and organise with other people interested in this project, head over to our Discord and:
- Follow the instructions on the
#join-course
channel - Join the
writing-assistant
channel
Just make sure you comment here to indicate that you’ll be contributing to this project
Team organization on the Hub
To join this team, make sure you join the following organisation on the Hub: