A Song is more than just the Lyrics
A Hugging Face & AWS Case Study
Summary
Musixmatch is a leading music data company providing data, tools, and services that enrich the way we experience music such as searching for songs and sharing song lyrics.
Musixmatch is the largest service of this kind in the world with over 80 million users and over 8 million distinct lyrics.
Every month, over 20 million people from around the world contribute to create the largest Music Catalog data ever.
Challenge & Solution
Musixmatch wanted to enhance the experience of music, and offer more out of songs than just spoken lyrics. They needed to support their unique music human curation process, which includes transcription, proofreading and syncing of song lyrics, as well as further analyze those lyrics with Natural Language Processing (NLP) in multiple languages.
Musixmatch's challenge was finding a way to perform classical NLP seamlessly on their Music Catalog, using cutting-edge, state-of-the-art technology. To accomplish this, Musixmatch turned to Hugging Face and AWS SageMaker. Using Hugging Face models allowed Musixmatch to perform a variety of classical NLP tasks on their Music Catalog seamlessly and with ease. This included Named Entity Recognition, Part-of-speech tagging as well as further custom classification tasks like lyrics Mood and Emotion Detection.
By utilizing the power of Hugging Face models, Transformers matched and eventually out-performed traditional algorithms that Musixmatch had tested in the past. This combined with SageMaker as an end-to-end Machine Learning platform, allowed them to train Hugging Face models and then deploy these models to production with ease, to support the human curated review of the Music Catalog and ensure lyrics perfection.
This solution saved the Musixmatch team time, resources and cost, all while creating business value by performing cutting edge, state-of-the-art Machine Learning processing to their Music Catalog.
About Musixmatch
Musixmatch is a leading music data company providing data, tools, and services that enrich the way we experience music such as searching for songs and sharing song lyrics, and it licences content to top music services such as Apple Music, Spotify, Amazon, Microsoft Bing, Tidal, Facebook, Google and others. With over 8 million distinct lyrics, Musixmatch is the world largest music database.
"One of our winners of the FbStart Program, Musixmatch, they’re doing great work." — Mark Zuckerberg, Co-founder and CEO of Meta Platforms (formerly Facebook, Inc.)
Benefits
The Benefits of a Hugging Face & AWS Backed Solution
Resource Efficiency
- Freeing up valuable time for Musixmatch's Data Scientist to focus on their core business tasks.
Agility
- The agility of this solution when trying new models and moving onto the next if it does not meet standards is key to Musixmatch's success.
Availability
- The availability of state of the art models from top research teams. (Google, Facebook, Microsoft, etc.)
Hugging Face & AWS, a Better Together Strategy
Why Hugging Face?
Musixmatch decided to use Hugging Face for their Natural Language Processing needs due to the ease of use of Transformer models and wide ranging availability of models on the Hugging Face hub.
With over 26k models in the Hugging Face model hub, Musixmatch was able to try a wide variety of models for their use case and easily fine-tune them to their business use case using their own data. If a model didn't perform at a high level they simply moved onto a new model.
Having access to all those models in one central location - the Hugging Face model hub - and the native integration with SageMaker, made loading models into their AWS environment simple.
Why AWS?
With high-performance compute options powered by machine learning, Amazon Web Services (AWS) enables organizations to undergo broad digital transformations with modern, cloud-native solutions. Improved processes, increased efficiency, and accelerated innovation are just some of the benefits realized from the inclusion of machine learning in business operations.
Offering a broad set of machine learning services and supporting cloud infrastructure, AWS enables organizations to tailor their machine learning solution to meet the unique needs of their business. Organizations are already realizing great value from AWS, enabling them to provide new experiences for their customers and drive business growth.
Improved processes, increased efficiency, and accelerated innovation are just some of the benefits realized from the inclusion of machine learning in business operations.
