bark-with-voice-clone/generate.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "from IPython.display import Audio\n",
    "from scipy.io.wavfile import write as write_wav\n",
    "\n",
    "from bark.api import generate_audio\n",
    "from bark.generation import SAMPLE_RATE, preload_models, codec_decode, generate_coarse, generate_fine, generate_text_semantic"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "semantic_path = \"semantic_output/pytorch_model.bin\" # set to None if you don't want to use finetuned semantic\n",
    "coarse_path = \"coarse_output/pytorch_model.bin\" # set to None if you don't want to use finetuned coarse\n",
    "fine_path = \"fine_output/pytorch_model.bin\" # set to None if you don't want to use finetuned fine\n",
    "use_rvc = True # Set to False to use bark without RVC\n",
    "rvc_name = 'mi-test'\n",
    "rvc_path = f\"Retrieval-based-Voice-Conversion-WebUI/weights/{rvc_name}.pth\"\n",
    "index_path = f\"Retrieval-based-Voice-Conversion-WebUI/logs/{rvc_name}/added_IVF256_Flat_nprobe_1_{rvc_name}_v2.index\"\n",
    "device=\"cuda:0\"\n",
    "is_half=True"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# download and load all models\n",
    "preload_models(\n",
    "    text_use_gpu=True,\n",
    "    text_use_small=False,\n",
    "    text_model_path=semantic_path,\n",
    "    coarse_use_gpu=True,\n",
    "    coarse_use_small=False,\n",
    "    coarse_model_path=coarse_path,\n",
    "    fine_use_gpu=True,\n",
    "    fine_use_small=False,\n",
    "    fine_model_path=fine_path,\n",
    "    codec_use_gpu=True,\n",
    "    force_reload=False,\n",
    "    path=\"models\"\n",
    ")\n",
    "\n",
    "if use_rvc:\n",
    "    from rvc_infer import get_vc, vc_single\n",
    "    get_vc(rvc_path, device, is_half)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# simple generation\n",
    "text_prompt = \"Hello, my name is Serpy. And, uh — and I like pizza. [laughs]\"\n",
    "voice_name = \"speaker_0\" # use your custom voice name here if you have on\n",
    "\n",
    "filepath = \"output/audio.wav\"\n",
    "audio_array = generate_audio(text_prompt, history_prompt=voice_name, text_temp=0.7, waveform_temp=0.7)\n",
    "write_wav(filepath, SAMPLE_RATE, audio_array)\n",
    "\n",
    "if use_rvc:\n",
    "    index_rate = 0.75\n",
    "    f0up_key = -6\n",
    "    filter_radius = 3\n",
    "    rms_mix_rate = 0.25\n",
    "    protect = 0.33\n",
    "    resample_sr = SAMPLE_RATE\n",
    "    f0method = \"harvest\" #harvest or pm\n",
    "    try:\n",
    "        audio_array = vc_single(0,filepath,f0up_key,None,f0method,index_path,index_rate, filter_radius=filter_radius, resample_sr=resample_sr, rms_mix_rate=rms_mix_rate, protect=protect)\n",
    "    except:\n",
    "        audio_array = vc_single(0,filepath,f0up_key,None,'pm',index_path,index_rate, filter_radius=filter_radius, resample_sr=resample_sr, rms_mix_rate=rms_mix_rate, protect=protect)\n",
    "    write_wav(filepath, SAMPLE_RATE, audio_array)\n",
    "\n",
    "Audio(audio_array, rate=SAMPLE_RATE)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# generation with more control\n",
    "text_prompt = \"Hello, my name is Serpy. And, uh — and I like pizza. [laughs]\"\n",
    "voice_name = \"speaker_0\" # use your custom voice name here if you have on\n",
    "\n",
    "filepath = \"output/audio.wav\"\n",
    "\n",
    "x_semantic = generate_text_semantic(\n",
    "    text_prompt,\n",
    "    history_prompt=voice_name,\n",
    "    temp=0.7,\n",
    "    top_k=50,\n",
    "    top_p=0.95,\n",
    ")\n",
    "\n",
    "x_coarse_gen = generate_coarse(\n",
    "    x_semantic,\n",
    "    history_prompt=voice_name,\n",
    "    temp=0.7,\n",
    "    top_k=50,\n",
    "    top_p=0.95,\n",
    ")\n",
    "x_fine_gen = generate_fine(\n",
    "    x_coarse_gen,\n",
    "    history_prompt=voice_name,\n",
    "    temp=0.5,\n",
    ")\n",
    "audio_array = codec_decode(x_fine_gen)\n",
    "write_wav(filepath, SAMPLE_RATE, audio_array)\n",
    "\n",
    "if use_rvc:\n",
    "    index_rate = 0.75\n",
    "    f0up_key = -6\n",
    "    filter_radius = 3\n",
    "    rms_mix_rate = 0.25\n",
    "    protect = 0.33\n",
    "    resample_sr = SAMPLE_RATE\n",
    "    f0method = \"harvest\" #harvest or pm\n",
    "    try:\n",
    "        audio_array = vc_single(0,filepath,f0up_key,None,f0method,index_path,index_rate, filter_radius=filter_radius, resample_sr=resample_sr, rms_mix_rate=rms_mix_rate, protect=protect)\n",
    "    except:\n",
    "        audio_array = vc_single(0,filepath,f0up_key,None,'pm',index_path,index_rate, filter_radius=filter_radius, resample_sr=resample_sr, rms_mix_rate=rms_mix_rate, protect=protect)\n",
    "    write_wav(filepath, SAMPLE_RATE, audio_array)\n",
    "\n",
    "Audio(audio_array, rate=SAMPLE_RATE)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "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": "3.10.8"
  },
  "orig_nbformat": 4
 },
 "nbformat": 4,
 "nbformat_minor": 2
}