music_rnn.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "import note_seq\n",
    "from magenta.models.melody_rnn import melody_rnn_sequence_generator\n",
    "from magenta.models.performance_rnn import performance_sequence_generator\n",
    "from magenta.models.shared import sequence_generator_bundle\n",
    "from note_seq.protobuf import generator_pb2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "SF2_PATH = 'soundfonts/Yamaha-C5-Salamander-JNv5.1.sf2'\n",
    "SAMPLE_RATE = float(44100)"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# MelodyRNN"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "note_seq.notebook_utils.download_bundle(\"basic_rnn.mag\", \"models/music_rnn/\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "bundle = sequence_generator_bundle.read_bundle_file('models/music_rnn/basic_rnn.mag')\n",
    "generator_map = melody_rnn_sequence_generator.get_generator_map()\n",
    "melody_rnn = generator_map['basic_rnn'](checkpoint=None, bundle=bundle)\n",
    "melody_rnn.initialize()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "input_sequence = note_seq.midi_file_to_note_sequence('data/clair_de_lune.mid')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "num_steps = 128 # change this for shorter or longer sequences\n",
    "temperature = 1.0 # the higher the temperature the more random the sequence."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "last_end_time = (max(n.end_time for n in input_sequence.notes) if input_sequence.notes else 0)\n",
    "qpm = input_sequence.tempos[0].qpm \n",
    "seconds_per_step = 60.0 / qpm / melody_rnn.steps_per_quarter\n",
    "total_seconds = num_steps * seconds_per_step"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "generator_options = generator_pb2.GeneratorOptions()\n",
    "generator_options.args['temperature'].float_value = temperature\n",
    "generate_section = generator_options.generate_sections.add(\n",
    "  start_time=last_end_time + seconds_per_step,\n",
    "  end_time=total_seconds\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "sequence = melody_rnn.generate(input_sequence, generator_options)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "note_seq.play_sequence(sequence, synth=note_seq.fluidsynth, sample_rate=SAMPLE_RATE, sf2_path=SF2_PATH)\n",
    "note_seq.plot_sequence(sequence)"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Performance RNN"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "note_seq.notebook_utils.download_bundle(\"performance_with_dynamics.mag\", \"models/performance_rnn/\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "bundle = sequence_generator_bundle.read_bundle_file('models/performance_rnn/performance_with_dynamics.mag')\n",
    "generator_map = performance_sequence_generator.get_generator_map()\n",
    "generator = generator_map['performance_with_dynamics'](checkpoint=None, bundle=bundle)\n",
    "generator.initialize()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "input_sequence = note_seq.midi_file_to_note_sequence('data/clair_de_lune.mid')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "temperature = 1.0"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "last_end_time = (max(n.end_time for n in input_sequence.notes) if input_sequence.notes else 0)\n",
    "seconds_per_step = 0.1\n",
    "total_seconds = 30"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "generator_options = generator_pb2.GeneratorOptions()\n",
    "generator_options.args['temperature'].float_value = temperature # Higher is more random; 1.0 is default. \n",
    "generate_section = generator_options.generate_sections.add(start_time=last_end_time + seconds_per_step, end_time=total_seconds)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "sequence = generator.generate(input_sequence, generator_options)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "note_seq.play_sequence(sequence, synth=note_seq.fluidsynth, sample_rate=SAMPLE_RATE, sf2_path=SF2_PATH)\n",
    "note_seq.plot_sequence(sequence)"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "venv",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.10.9"
  },
  "orig_nbformat": 4
 },
 "nbformat": 4,
 "nbformat_minor": 2
}