| Data | Version | Comment | Editor |
|---|---|---|---|
| 2024.05.15 | 0.0.1 | Draft | E.W |
| 2024.05.19 | 0.0.11 | Revise object and future | E.W |
| 2024.06.12 | 1.0 | MVP | Kisho |
| 2024.07.04 | 1.1 | Final | E.W & Kisho |
It is a product-type project for the deep learning class.
Music is a common way for most people to release stress, seek comfort, and pass the time. However, major music streaming platforms like Spotify and Apple Music lack mood-based recommendation features. Additionally, they do not seem to effectively integrate AI technology into their products, which can come across as insensitive to users' needs.
To fill the blank in the market, this project aims to develop a chatbot, leveraging AI to recommend the most suitable playlists based on user input. Users can write down a few sentences about their feelings or fulfill the questionsheet provided then receive their recommendations of music that is close to the mood.
git clone git@github.com:callmeeric5/MusicBot.git
- Build a Virtual Env if you need
pip install -r requirements.txt
-
Go to https://console.groq.com/playground create a project
-
Go to https://qdrant.tech/ create a cluster
-
Creat a
secrets.tomlunder the folder.streamlitEnter your API codes from the websites
GROQ_API_KEY =
QDRANT_API_KEY =
QDRANT_CLIENT_URL =
cd MusicBot
streamlit run 💬_Chat.py
docker build -t musicbot .
docker run -p 8501:8501 musicbot
Let's start this sepcial journery of music!
-
Index example
-
Chat example
-
Questionsheet example
-
Collaborative recommendation by user portrait.
-
Enhance model accuracy.
-
Expand dataset beyond Taylor Swift’s songs.
-
Integrate with other platforms for broader use.