154 lines
3.6 KiB
Plaintext
154 lines
3.6 KiB
Plaintext
{
|
|
"cells": [
|
|
{
|
|
"cell_type": "markdown",
|
|
"id": "b56703b7",
|
|
"metadata": {},
|
|
"source": [
|
|
"*this notebook requires a working PyTorch GPU environment* "
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "markdown",
|
|
"id": "3967f4b4",
|
|
"metadata": {},
|
|
"source": [
|
|
"# OpenAI's Whisper multimodel\n",
|
|
"\n",
|
|
"Speech to text...\n",
|
|
"\n",
|
|
"more information at\n",
|
|
"- https://openai.com/blog/whisper\n",
|
|
"- https://github.com/openai/whisper\n",
|
|
"\n",
|
|
"\n"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": null,
|
|
"id": "dac0ca1b-b098-426a-982a-777049f40581",
|
|
"metadata": {},
|
|
"outputs": [],
|
|
"source": [
|
|
"!pip install git+https://github.com/openai/whisper.git "
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": null,
|
|
"id": "2d10ec51-1c42-40de-a6ac-f2b70566b9a3",
|
|
"metadata": {},
|
|
"outputs": [],
|
|
"source": [
|
|
"import time\n",
|
|
"import torch\n",
|
|
"import whisper\n",
|
|
"from datetime import datetime"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": null,
|
|
"id": "7ffe9e04-cc7f-4c65-b9a3-10bb1a5eac41",
|
|
"metadata": {},
|
|
"outputs": [],
|
|
"source": [
|
|
"Models = [\n",
|
|
" 'base',\n",
|
|
" 'medium',\n",
|
|
" ]\n",
|
|
"\n",
|
|
"Files = [\n",
|
|
" \"./TestdateiAudiotranskription.mp3\"\n",
|
|
"]"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": null,
|
|
"id": "ca9f2fae-f7f9-4846-afd5-7c27253e4061",
|
|
"metadata": {},
|
|
"outputs": [],
|
|
"source": [
|
|
"LogFile = \"./log.md\""
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": null,
|
|
"id": "4f7fbfb3-a116-4dcb-a154-188a26291f47",
|
|
"metadata": {},
|
|
"outputs": [],
|
|
"source": [
|
|
"def transcribe(model, file):\n",
|
|
" s_transcribe = time.time()\n",
|
|
" result = model.transcribe(file, verbose=True)\n",
|
|
" e_transcribe = time.time()\n",
|
|
" writeLog(f'- **Transcribe ({file}) : {e_transcribe-s_transcribe}**\\n')\n",
|
|
" writeLog(f' - ({file}) : {result[\"text\"]}\\n')\n",
|
|
"\n",
|
|
"\n",
|
|
"def writeLog(mes):\n",
|
|
" with open(LogFile, mode=\"a\") as f:\n",
|
|
" f.write(mes)"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": null,
|
|
"id": "46042b2a-507b-48c8-b0f2-dd9e8260c52a",
|
|
"metadata": {},
|
|
"outputs": [],
|
|
"source": [
|
|
"writeLog(\"# Whisper Research\\n\")\n",
|
|
"writeLog(f\"## Research Start {datetime.now().strftime('%Y/%m/%d %H:%M:%S')} \\n\")\n",
|
|
"writeLog(f' ** cuda available: {torch.cuda.is_available()} **\\n')\n",
|
|
"for m in Models:\n",
|
|
" writeLog(f' ### Model : ({m}) ---\\n')\n",
|
|
" s_loadmodel = time.time()\n",
|
|
" model = whisper.load_model(m)\n",
|
|
" e_loadmodel = time.time()\n",
|
|
" writeLog(f' ** Load Model ({m}) : {e_loadmodel-s_loadmodel}**\\n')\n",
|
|
" for file in Files:\n",
|
|
" transcribe(model, file)\n",
|
|
" writeLog(' -------- \\n')"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": null,
|
|
"id": "6d6bb338-cbe4-4bc1-9ce4-204771f9721b",
|
|
"metadata": {},
|
|
"outputs": [],
|
|
"source": [
|
|
"# clear resources\n",
|
|
"from numba import cuda\n",
|
|
"device = cuda.get_current_device()\n",
|
|
"device.reset()"
|
|
]
|
|
}
|
|
],
|
|
"metadata": {
|
|
"kernelspec": {
|
|
"display_name": "Python 3 (ipykernel)",
|
|
"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": "3.10.12"
|
|
}
|
|
},
|
|
"nbformat": 4,
|
|
"nbformat_minor": 5
|
|
}
|