Website: https://ivory-parallax-spot.wayscript.cloud/
Inspiration
What made me want to create Co:Dense was when I started to get frustrated over reading complex paragraphs. I wanted to create an application where it could summarize the paragraph for me and others. This motivated me to use the co:here technology since it utilizes NLP.
What it does
Co:Dense prompts the user a text area in which users can input texts. After clicking a button to submit the input, Co:Dense runs the user's input into an example model provided by co:here which returns a generated sentence that summarizes the user's input.
How I built it
Co:Dense wouldn't be what it is if it wasn't for co:here's example model. Co:Dense is mainly a not so beautiful website form in which user's will type in their input into a text area. It takes the user's input and runs it into co:hers's example model; which returns a generalize sentence that summarize the user's input.
Challenges I ran into
A challenge I ran into and was stuck on for about 3 hours was learning how to deploy my website. The website was built with the flask framework. At first, I wanted to deploy my website in Heroku; however I came across multiple errors such as 403, 500, and 503. This errors continue to haunt me as I tried different approaches in Google the Cloud. Despite, trying app engine, cloud run, and cloud functions I couldn't seem to solve this challenge. So, my last option was to go watch a youtube video to find where I can deploy my app. This was when I came across Wayscript and finally I got to deploy my website with ease and no errors.
Accomplishments that I'm proud of
A great accomplishment I am proud of was my ability to find different solutions to solve a problem. Like my the problem with deployment, I motivated myself to find a way to deploy my app. I tried different products from the Google Cloud, a popular PAAS platform like Heroku, and watch youtube videos. This accomplishment/ability is what I'm most proud of during this hackathon.
What I learned
I learned how co:here has open NLP to the public in a very easy and accessible way. As I was browsing through the resources provided to the hackatees, I stumbled upon co:here which consist of NLP technology. After giving some time exploring the website and the product, I came to love co:here for it's simple design and upfront explanation to it's apis. Overall, co:here was a fun and exciting technology I learned throughout this hackathon.
What's next for Co:Dense
There is a lot of potential for Co:Denser to improve. For instance a great improvement would be to create a finetune model from scratch instead of using an example model from co:here. Another great implementation to improve Co:Denser is to make it so that it is a chrome extension. This makes it more accessible for users to access since it'll be at the top right of their browser instead of having to go to a website.
Log in or sign up for Devpost to join the conversation.