Connotate.AI

• 

  This is the UI.

• 

  This pretty much sums up the entire process.

Inspiration

English is our second language and we've been struggling with analyzing quotes ever since the first grade. ICE SPICE

What it does

It takes in a phrase as input, puts it through the AI and it outputs the predicted sentiment of the phrase.

How we built it

We used HTML and CSS to build the UI. We used datasets from kaggle, trained them on Colab with models from Distillberg.

Challenges we ran into

One of our teammates went on a "boba run" and didn't come back for the entire first day. We also spent a long time obtaining good datasets.

Accomplishments that we're proud of

The program can output phrase sentiment with 70% real accuracy, which is good considering we created this AI. Plus, the UI looks good for first-time HTML and CSS coders.

What we learned

We learned that we need large datasets to create accurate AI and that Macs train AI very slowly. I BE GOIN HARD

What's next for Connotate.AI

In the future, we want the model to also be able to give connotative meaning to the phrases. GRAH

Built With

• colab
• css
• html
• huggingface
• javascript
• python