Submission specs

input representation: symbolic, monophonic

subtasks: 1

threads: see Parallelization

expected memory footprint:

expected runtime:

requirements: sbcl, quicklisp, sqlite3

Usage

Move the training data (monophonic midi files) into a new directory called training-data. Move the primes (monophonic midi files) into a new directory called primes. Create a directory results, where the results will be stored. Results are CSV files.

Alternatively, run

make all

To use the medium training set, and sample 100 primes at random from this training set.

To generate results, run

chmod +x run.sh
./run.sh

Parallelization

As many instances of this code as desired can be run in parallel on disjoint subsets of primes.