mirror of
https://github.com/neuphonic/neutts-air.git
synced 2025-10-10 02:44:44 +03:00
115 lines
2.5 KiB
Plaintext
115 lines
2.5 KiB
Plaintext
{
|
|
"cells": [
|
|
{
|
|
"cell_type": "markdown",
|
|
"id": "a0fa9718",
|
|
"metadata": {},
|
|
"source": [
|
|
"Import required libraries"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": null,
|
|
"id": "939f4fdc",
|
|
"metadata": {},
|
|
"outputs": [],
|
|
"source": [
|
|
"import sys\n",
|
|
"#Append the parent directory to the sys.path to allow imports from neuttsair package\n",
|
|
"sys.path.append('..')\n",
|
|
"from neuttsair.neutts import NeuTTSAir\n",
|
|
"from IPython.display import Audio"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "markdown",
|
|
"id": "e4e61937",
|
|
"metadata": {},
|
|
"source": [
|
|
"Downloads files and loads the model into memory"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": null,
|
|
"id": "cfabf9dd",
|
|
"metadata": {},
|
|
"outputs": [],
|
|
"source": [
|
|
"%%capture\n",
|
|
"tts = NeuTTSAir(\n",
|
|
" backbone_repo=\"neuphonic/neutts-air-q8-gguf\",\n",
|
|
" backbone_device=\"cpu\",\n",
|
|
" codec_repo=\"neuphonic/neucodec\",\n",
|
|
" codec_device=\"cpu\"\n",
|
|
")"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "markdown",
|
|
"id": "75c87818",
|
|
"metadata": {},
|
|
"source": [
|
|
"Pick your speaker and type up your input text - and generate!"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": null,
|
|
"id": "914e5e00",
|
|
"metadata": {},
|
|
"outputs": [],
|
|
"source": [
|
|
"speaker = \"dave\" # default speakers are 'dave' and 'jo'\n",
|
|
"input_text = \"Hey there, I'm Andy. I'm 25 and I just moved to London. The underground is pretty confusing, but it gets me around in no time at all.\"\n",
|
|
"\n",
|
|
"ref_text = f\"../samples/{speaker}.txt\"\n",
|
|
"ref_audio_path = f\"../samples/{speaker}.wav\"\n",
|
|
"\n",
|
|
"ref_text = open(ref_text, \"r\").read().strip()\n",
|
|
"ref_codes = tts.encode_reference(ref_audio_path)\n",
|
|
"wav = tts.infer(input_text, ref_codes, ref_text)"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "markdown",
|
|
"id": "e8da4a9d",
|
|
"metadata": {},
|
|
"source": [
|
|
"Listen to your generation!"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": null,
|
|
"id": "21774af1",
|
|
"metadata": {},
|
|
"outputs": [],
|
|
"source": [
|
|
"Audio(wav, rate=24000)"
|
|
]
|
|
}
|
|
],
|
|
"metadata": {
|
|
"kernelspec": {
|
|
"display_name": "",
|
|
"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"
|
|
}
|
|
},
|
|
"nbformat": 4,
|
|
"nbformat_minor": 5
|
|
}
|