{
"cells": [
{
"cell_type": "code",
"execution_count": 1,
"metadata": {},
"outputs": [],
"source": [
"import os\n",
"import glob\n",
"import numpy as np\n",
"from einops import rearrange\n",
"from suno_utils.audio import Audio"
]
},
{
"cell_type": "code",
"execution_count": 2,
"metadata": {},
"outputs": [],
"source": [
"# load files\n",
"music_emb = np.load(\"/app/suno/data/audio_mono_24khz/genius_hq/embeddings/music_emb.npy\")\n",
"text_emb = np.load(\"/app/suno/data/audio_mono_24khz/genius_hq/embeddings/text_emb.npy\")\n",
"ids = np.load(\"/app/suno/data/audio_mono_24khz/genius_hq/embeddings/ids.npy\")\n",
"start_times = np.load(\"/app/suno/data/audio_mono_24khz/genius_hq/embeddings/start_times.npy\")\n",
"lyrics = np.load(\"/app/suno/data/audio_mono_24khz/genius_hq/embeddings/lyrics.npy\")\n",
"\n",
"# reshape\n",
"music_emb = np.concatenate([line for line in music_emb])\n",
"text_emb = np.concatenate([line for line in text_emb])\n",
"start_times = np.concatenate([line for line in start_times])"
]
},
{
"cell_type": "code",
"execution_count": 16,
"metadata": {},
"outputs": [],
"source": [
"def lyric_to_audio(input_ix, is_display=False):\n",
" # print ID\n",
" if is_display:\n",
" print(\"input id: %s\" % ids[input_ix])\n",
"\n",
" # print lyrics\n",
" if is_display:\n",
" print(lyrics[input_ix][13:])\n",
"\n",
" # retrieval\n",
" input_emb = text_emb[input_ix][np.newaxis, :]\n",
" sim = np.dot(input_emb, music_emb.T)[0]\n",
" top5_indices = np.argsort(sim)[::-1][:5]\n",
"\n",
" # print retrieved ID\n",
" if is_display:\n",
" print(\"retrieved id: %s\" % ids[top5_indices[0]])\n",
"\n",
" # playback audio\n",
" audio_id = ids[top5_indices[0]]\n",
" if is_display:\n",
" audio_fn = os.path.join(\"/app/suno/data/audio_mono_24khz/genius_hq/audio/%s.wav\" % audio_id)\n",
" retrieved_audio = Audio.from_file(audio_fn, sample_rate=24000)\n",
" retrieved_audio.play()\n",
"\n",
" if not is_display:\n",
" return top5_indices[0]\n",
"\n",
"def audio_to_lyric(input_ix, is_display=False):\n",
" # print ID\n",
" if is_display:\n",
" print(\"input id: %s\" % ids[input_ix])\n",
"\n",
" # retrieval\n",
" input_emb = music_emb[input_ix][np.newaxis, :]\n",
" sim = np.dot(input_emb, text_emb.T)[0]\n",
" top5_indices = np.argsort(sim)[::-1][:5]\n",
"\n",
" # print retrieved ID\n",
" if is_display:\n",
" print(\"retrieved id: %s\" % ids[top5_indices[0]])\n",
"\n",
" # print lyrics\n",
" if is_display:\n",
" print(lyrics[top5_indices[0]][13:])\n",
"\n",
" # playback audio\n",
" audio_id = ids[input_ix]\n",
" if is_display:\n",
" audio_fn = os.path.join(\"/app/suno/data/audio_mono_24khz/genius_hq/audio/%s.wav\" % audio_id)\n",
" retrieved_audio = Audio.from_file(audio_fn, sample_rate=24000)\n",
" retrieved_audio.play()\n",
"\n",
" if not is_display:\n",
" return top5_indices[0]\n"
]
},
{
"cell_type": "code",
"execution_count": 17,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"lyric to audio accuracy: 0.9833\n"
]
}
],
"source": [
"correct = 0\n",
"for i in range(300):\n",
" out = lyric_to_audio(i, is_display=False)\n",
" if i == out:\n",
" correct += 1\n",
"print(\"lyric to audio accuracy: %.4f\" % (correct / 300))"
]
},
{
"cell_type": "code",
"execution_count": 19,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"audio to lyric accuracy: 0.9800\n"
]
}
],
"source": [
"correct = 0\n",
"for i in range(300):\n",
" out = audio_to_lyric(i, is_display=False)\n",
" if i == out:\n",
" correct += 1\n",
"print(\"audio to lyric accuracy: %.4f\" % (correct / 300))"
]
},
{
"cell_type": "code",
"execution_count": 10,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"input id: e4037136-f14b-4783-80cc-c3fa39e402ce\n",
"retrieved id: e4037136-f14b-4783-80cc-c3fa39e402ce\n",
"i open my eyes and suddenly feel your love wash over me let go of the past it's time to move on now that i'm finally free oh i wasted so much time\n"
]
},
{
"data": {
"text/html": [
"\n",
"\n"
],
"text/plain": [
""
]
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"audio_to_lyric(150)"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": []
}
],
"metadata": {
"kernelspec": {
"display_name": "suno_env",
"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": "3.10.13"
}
},
"nbformat": 4,
"nbformat_minor": 2
}