Post-Processing and Markov Chains Program-
Project Overview -
Goals:
- Create a text generator that uses a Markov chain
- Practice starting from scratch
- Using data to train a generative model
- Post-filtering the output of a generative model
For this assignment, you will need a technical collaborator/buddy to work with. Also, you will share two screenshots of your project. One a few days early, and one on the due date.
Post-Processing -
Your generator should use post-processing on the generated text:
- Use generate-and-test to constrain the output of your generator to between 100 and 140 characters.
Feel free to do additional post-processing if it helps your framing or theme!
Tips & Advanced Techniques -
- Get the code from these slides working early
- Experiment with different settings for the Markov chain
- Try using a short section of text, not a full novel: Tracery grammars are an inefficient representation for giant Markov chains.
- Try other forms of post-processing! - What about making sure sentences are properly capitalized?
- If you use code that you didn't write, cite the source it came from.
Attachment:- Assignment File.rar