AWS DeepComposer adds new Transformers algorithm that allows developers to extend an input melody

Posted on: Nov 9, 2020

Today, we are excited to announce a Transformers-based algorithm in AWS DeepComposer. AWS DeepComposer gives developers a creative way to get hands-on using a musical keyboard and the latest generative AI techniques to expand their skills. Transformers are state-of-the-art models that work on sequential data such as text, stock values, music, and genome sequences. With the newly launched Transformers feature, developers can iteratively extend a melody in AWS DeepComposer to create new and longer compositions.  

To get started, developers can choose the Transformers generative AI technique and the TransformerXLClassical model. Then, use one of the seven parameters to adjust the musical composition: sampling technique, sampling threshold, input duration, track extension duration, maximum rest time, maximum note length, and creative risk. After creating or selecting the input melody and inference parameters, developers can create their first composition by choosing extend input melody. The model will extend the input melody by up to 20 seconds. Developers can learn more about the Transformers algorithm in the learning capsules provided in the AWS DeepComposer console.  

Get hands-on with the AWS DeepComposer keyboard to start composing today. AWS DeepComposer keyboard is now available on at a special price of $79.20 (20% off) for a limited time during November and December. The pricing includes the keyboard and a 3-month free trial of AWS DeepComposer services.

Learn more about AWS DeepComposer.