N-gram Models

N-gram Models

14 devlogs
35h 43m
•  Ship certified
Created by GB114

This project is basically my approach to N-grams through the use of dictionaries. Please vote :)

Timeline

Ship 2

1 payout of shell 30.0 shells

GB114

4 days ago

GB114 Covers 5 devlogs

Hid the scroll bar on scrollable elements.

Update attachment
GB114
GB114
1h 10m 4 days ago

Added light mode and dark mode switch.

Fixed issues with the download button.

If you don't have Ideas for the corpus use this.

Update attachment

Finally deployed my website on Vercel. Apparently, the method I used before only worked when I opened the terminal, so here is the updated link:

<a href = https://ngram.vercel.app>https://ngram.vercel.app</a>

(I don't think HTML works here, but here you go.)

Update attachment

Ship 1

1 payout of shell 252.0 shells

GB114

about 1 month ago

GB114 Covers 9 devlogs

Finally deployed this project on a server now I am going to submit it.

Update attachment
  • Added more content to the website.
  • Added navigation bar.
  • Stylized the navigation bar.
  • Tried to make the website more friendly to smaller screens.

Here are the things I added:
- Customized the layout a bit.
- styled the popup notifications and added an interaction to hide them faster when clicked.
- Styled the output field.
- Added a button to generate more words (with a field to specify the word count.)

Worked some more on my project:
- Changed the style a bit,
- Reworked the information presentation by making it give you pop up like notifications,
- Added the output to be at the bottom of the program,
- Figured out why javascript's syntax highlighting didn't work
- Also learned a few things about flask
- How it handles css and javascript files (as static) and how it requires them to be at a specific place to access them.
- You can transfer information from python to javascript using Jsonify and creating json files.
Here is a demo for the current version of my project:

Haven't done much today but I tried to style my website.

Update attachment

Made my library be in an all accessible form, a website. I have used flask to connect the library with HTML and I think it works fine. It needs a bit of polishing and fixing of some bugs, but it is a great start.

Added a method to write a text (readable) file that shows the relationship of the words in another way. I also changed the N-gram from count based to weight (probability) based to help with the next process (generating text). Additionally, I got git working. (so no more added via upload).

Decided to change the N-gram generation logic to make it a bit more modular other than only being able to make bigrams. (And making the model more accessible for later.)

I have started working on my Ngram and the Ngram generator successfully generates Ngrams.