Inspiration
Many people, especially elderly people, don't have great hiring and they have trouble listening. To help them out we made this tool, that can translate speech to text. This tool can save them a lot of energy and time at no extra cost. This hack is an accessibility hack.
What it does
There are 2 ways to use it:
- people can upload some text in a .wav file format.
- people can say something on the website using their microphone. Our ML algorithms then convert the speech to text using python's ML framework: Pytorch.
How we built it
First, we had to get the audio recordings into our server using a long script file. Then after we received the file in our folders, we used python, PyTorch, and Jupyter notebooks to train the AI model. We trained our model on Microsoft azure since our computer took a long time during the training process.
Challenges we ran into
One problem we ran into was saving the recording on our server. I couldn't find a way to do that, but eventually, I found a way to bypass this error.
Accomplishments that we're proud of
We are proud of finishing our project, we are also proud of all the people we could help out with our application
What we learned
We learned how to save audio recordings onto a server.
What's next for Speechify
We could improve our systems to make our application faster, and maybe we could also create a more accessible iPhone application for it.
Log in or sign up for Devpost to join the conversation.