Back to Education
Deep Learning Specialization · Sequence Models

Jazz Solo Generation with LSTM

A creative sequence-generation project that trains an LSTM on jazz musical values and then uses an inference model to generate a new solo.

LSTM generationMany-to-many trainingInference-time samplingMusic sequencesSoftmax over notes/chords
GitHub

What this project is about

This project treats music as a sequence of symbolic events. The LSTM learns patterns over notes/chords during training and then generates a new sequence one step at a time during inference.

Implementation focus

The most valuable part was seeing the difference between training-time teacher forcing and inference-time generation, where the model feeds its own predictions forward.

Key results and artifacts

  • Training examples: 60
  • Sequence length: 30
  • Unique musical values: 90
  • Generated output length: 50 steps

What I learned

This milestone made sequence modeling tangible: the model learned a probability distribution over musical events and sampled from it.

Milestone deep dive

What this project proves in my learning path.

A creative sequence-generation project that trains an LSTM on jazz musical values and then uses an inference model to generate a new solo. The milestone is included because it shows a complete learning unit: a problem, a data representation, an implementation path, measurable outputs, and a lesson that later connects to larger AI systems.

Input layerordered sequences of text, audio, dates, or musical events
Model layerrecurrent or attention-based sequence model
Evaluation layertests, metrics, plots, or qualitative output
Artifact layerclassification, generation, alignment, or trigger probabilities
Problem framing

What I was trying to solve

This page is not just a lab description. It explains how I transformed the assignment idea into an engineering story: what the model receives, what it optimizes, what output it should produce, and how the result can be inspected.

Implementation

What I coded or assembled

The implementation focus was the mechanism behind the result: functions, model blocks, training logic, preprocessing, evaluation, and the small technical details that make the notebook work rather than only look correct.

Concepts

Core concepts

LSTM generationMany-to-many trainingInference-time samplingMusic sequencesSoftmax over notes/chords

These tags show the technical ideas this milestone added to my toolkit.

Learning result

What changed after it

This milestone made sequence modeling tangible: the model learned a probability distribution over musical events and sampled from it.

Important outputs and evidence

  • Training examples: 60
  • Sequence length: 30
  • Unique musical values: 90
  • Generated output length: 50 steps

How I would explain it in an interview

I would describe this milestone by starting from the data representation, then explain the model or algorithm, then show what was measured. The important part is not only the final number; it is the ability to reason about why the method works, what assumptions it makes, and what limitations should be improved next.