Skip to content

IshitaG-2002IGK/lyrics-jina

Repository files navigation

JINICA


What is the Project about ?

  • Jinica -- Jina + Musica (En espagnol)
  • The project is based on matching a query word with the lyrics of a song.
  • Finding the best possible match for a word in the lyrics of a song, which can be helpful while playing "Song Association Game"

Why did we create it in the first place ?

  • We're all music lovers, and jam to any given song playing
  • We participated in the collaborated event with JinaAI X Featurepreneur, as we were genuinely interested in learning more about JinaAI, and wanted to explore different libraries in ML.
  • And lastly we're all music lovers, and would jam to any given song playing anywhere at anytime. So we wanted to focus the project primarily on "music" and finding/matching the words to the lyrics.

How did we make the project ?

  • We built our project using JinaAI (Document, Flow, Executor).
  • Made some graphs - Bargraphs, Density-Contour charts for visualizing the data that we manually built.
  • We've attached a video for a better explanation regarding our project.

Preview

https://www.youtube.com/watch?v=12GG5AhiPH4&t=7s

Instructions to run it locally:

WE'LL FIND THE SONG USING THE WORDS YOU GIVE US!!

To create and activate new JINA conda environment:

conda create --name JINA    
conda activate JINA   

Install requirements

pip install -r requirements.txt  
conda install -c conda-forge jupyterlab  

To run streamlit

streamlit run visual_stream.py

To view the graph

python density_contour.py
python bar_graph.py

Tech Stack used

Here are a few :

  • Jina AI
  • Python
  • Streamlit
  • Plotly
  • ML libraries : pandas, numpy, pytorch etc

Contributors




Ishita Gopalakrishnan

Ponshriharini V

Sanjjushri Varshini

Made with ☕ and ❤️

About

Jinica - The project is based on matching a query word with the lyrics of a song. This finds the best possible match for a word in the lyrics of a song.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published