feat: add dataset generator

Signed-off-by: Alexandre Gomez <gomez.a.corneille@gmail.com>

dataset-generator.ipynb

@@ -0,0 +1,138 @@
+{
+ "cells": [
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import os\n",
+    "import torch\n",
+    "\n",
+    "device = torch.device(\"cpu\")\n",
+    "torch.set_num_threads(4)\n",
+    "local_file = \"model.pt\"\n",
+    "\n",
+    "if not os.path.isfile(local_file):\n",
+    "    torch.hub.download_url_to_file(\"https://models.silero.ai/models/tts/fr/v3_fr.pt\",\n",
+    "                                   local_file)  \n",
+    "\n",
+    "model = torch.package.PackageImporter(local_file).load_pickle(\"tts_models\", \"model\")\n",
+    "model.to(device)\n",
+    "\n",
+    "sample_rate = 48000\n",
+    "num_files = 10\n",
+    "texts = [\n",
+    "    # Dis Cyril\n",
+    "    {\n",
+    "        \"text\": \"Dis Cyril\",\n",
+    "        \"yes\": True\n",
+    "    },\n",
+    "    # Dis Cyril\n",
+    "    {\n",
+    "        \"text\": \"Dis Cyril\",\n",
+    "        \"yes\": True\n",
+    "    },\n",
+    "    # Dis Cyril\n",
+    "    {\n",
+    "        \"text\": \"Dis Cyril\",\n",
+    "        \"yes\": True\n",
+    "    },\n",
+    "    # Dis Cyril\n",
+    "    {\n",
+    "        \"text\": \"Dis Cyril\",\n",
+    "        \"yes\": True\n",
+    "    },\n",
+    "    # Dis Cyril\n",
+    "    {\n",
+    "        \"text\": \"Dis Cyril\",\n",
+    "        \"yes\": True\n",
+    "    },\n",
+    "    # Comment tu t'appelles ?\n",
+    "    {\n",
+    "        \"text\": \"Comment tu t'appelles ?\",\n",
+    "        \"yes\": False\n",
+    "    },\n",
+    "    # Quel est ton nom ?\n",
+    "    {\n",
+    "        \"text\": \"Quel est ton nom ?\",\n",
+    "        \"yes\": False\n",
+    "    },\n",
+    "    # Tu t'appelles comment ?\n",
+    "    {\n",
+    "        \"text\": \"Tu t'appelles comment ?\",\n",
+    "        \"yes\": False\n",
+    "    },\n",
+    "    # Comment dois-je t'appeler ?\n",
+    "    {\n",
+    "        \"text\": \"Comment dois-je t'appeler ?\",\n",
+    "        \"yes\": False\n",
+    "    },\n",
+    "    # Peux-tu me dire ton nom ?\n",
+    "    {\n",
+    "        \"text\": \"Peux-tu me dire ton nom ?\",\n",
+    "        \"yes\": False\n",
+    "    },\n",
+    "    # Quelle heure est-il ?\n",
+    "    {\n",
+    "        \"text\": \"Quelle heure est-il ?\",\n",
+    "        \"yes\": False\n",
+    "    },\n",
+    "    # Peux-tu me donner la météo de demain ?\n",
+    "    {\n",
+    "        \"text\": \"Peux-tu me donner la météo de demain ?\",\n",
+    "        \"yes\": False\n",
+    "    },\n",
+    "    # Comment cuisiner une omelette ?\n",
+    "    {\n",
+    "        \"text\": \"Comment cuisiner une omelette ?\",\n",
+    "        \"yes\": False\n",
+    "    },\n",
+    "    # Raconte-moi une blague.\n",
+    "    {\n",
+    "        \"text\": \"Raconte-moi une blague.\",\n",
+    "        \"yes\": False\n",
+    "    },\n",
+    "    # Quel est le sens de la vie ?\n",
+    "    {\n",
+    "        \"text\": \"Quel est le sens de la vie ?\",\n",
+    "        \"yes\": False\n",
+    "    },\n",
+    "    # Comment se rendre à Paris depuis Lyon ?\n",
+    "    {\n",
+    "        \"text\": \"Comment se rendre à Paris depuis Lyon ?\",\n",
+    "        \"yes\": False\n",
+    "    },\n",
+    "    # Quels sont les meilleurs films de 2020 ?\n",
+    "    {\n",
+    "        \"text\": \"Quels sont les meilleurs films de 2020 ?\",\n",
+    "        \"yes\": False\n",
+    "    },\n",
+    "]\n",
+    "speaker = \"random\"\n",
+    "\n",
+    "for index, text in enumerate(texts):\n",
+    "    for i in range(num_files):\n",
+    "        audio_paths = model.save_wav(text=text,\n",
+    "                                    speaker=speaker,\n",
+    "                                    sample_rate=sample_rate)\n",
+    "    \n",
+    "        # Rename the file to test_(1...1000)\n",
+    "        os.rename('test.wav', f'../dataset/{\"yes\" if text == \"Dis Cyril\" else \"no\"}/{i + index * num_files}.wav')"
+   ]
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": ".venv",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "name": "python",
+   "version": "3.12.1"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 2
+}