taco/microWakeWord-Trainer-Nvidia-Docker
1
0
Fork 0
mirror of https://github.com/TaterTotterson/microWakeWord-Trainer-Nvidia-Docker.git synced 2026-06-12 20:10:19 -06:00
Code Issues Packages Projects Releases Wiki Activity

Update advanced_training_notebook.ipynb

Browse Source
This commit is contained in:
MasterPhooey
2025-01-05 15:46:31 -06:00
committed by GitHub
parent 4af19466fc
commit cc962f2069
1 changed files with 11 additions and 14 deletions
Download Patch File Download Diff File Expand all files Collapse all files

23
advanced_training_notebook.ipynb
View File

@@ -202,31 +202,28 @@
"print(f\"Number of FLAC files found: {len(audioset_files)}\")\n",
"\n",
"if audioset_files:\n",
" audioset_dataset = Dataset.from_dict({\"audio\": [str(file) for file in audioset_files]})\n",
" audioset_dataset = audioset_dataset.cast_column(\"audio\", Audio(sampling_rate=16000))\n",
"\n",
" corrupted_files = []\n",
"\n",
" print(\"Converting Audioset files to 16kHz WAV...\")\n",
" for row in tqdm(audioset_dataset, desc=\"Processing Audioset files\"):\n",
" for file_path in tqdm(audioset_files, desc=\"Processing Audioset files\"):\n",
" try:\n",
" # Define output file name and path\n",
" name = Path(row[\"audio\"][\"path\"]).stem + \".wav\"\n",
" output_path = Path(output_dir) / name\n",
" # Attempt to load the file and handle any errors\n",
" audio, sampling_rate = sf.read(file_path)\n",
" \n",
" # Check if audio data is valid before writing\n",
" if row[\"audio\"][\"array\"] is None or len(row[\"audio\"][\"array\"]) == 0:\n",
" raise ValueError(f\"Empty or invalid audio data in file: {row['audio']['path']}\")\n",
" if audio is None or len(audio) == 0:\n",
" raise ValueError(f\"Empty or invalid audio data in file: {file_path}\")\n",
"\n",
" # Resample audio to 16kHz\n",
" output_path = Path(output_dir) / (file_path.stem + \".wav\")\n",
" scipy.io.wavfile.write(\n",
" output_path,\n",
" 16000,\n",
" (row[\"audio\"][\"array\"] * 32767).astype(np.int16),\n",
" (audio * 32767).astype(np.int16),\n",
" )\n",
" except (sf.LibsndfileError, ValueError, Exception) as e:\n",
" # Log the error and skip the file\n",
" print(f\"Error converting {row['audio']['path']}: {e}\")\n",
" corrupted_files.append(row[\"audio\"][\"path\"])\n",
" print(f\"Error converting {file_path}: {e}\")\n",
" corrupted_files.append(str(file_path))\n",
"\n",
" # Log corrupted files\n",
" if corrupted_files:\n",