taco/microWakeWord-Trainer-Nvidia-Docker
1
0
Fork 0
mirror of https://github.com/TaterTotterson/microWakeWord-Trainer-Nvidia-Docker.git synced 2026-06-13 04:20:19 -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 14:22:44 -06:00
committed by GitHub
parent 16f8e78846
commit 2137c50320
1 changed files with 5 additions and 10 deletions
Download Patch File Download Diff File Expand all files Collapse all files

15
advanced_training_notebook.ipynb
View File

@@ -177,9 +177,8 @@
"\n", "\n",
"# Directory setup\n", "# Directory setup\n",
"audioset_dir = \"./audioset\"\n", "audioset_dir = \"./audioset\"\n",
"os.makedirs(audioset_dir, exist_ok=True)\n",
"\n",
"output_dir = \"./audioset_16k\"\n", "output_dir = \"./audioset_16k\"\n",
"os.makedirs(audioset_dir, exist_ok=True)\n",
"os.makedirs(output_dir, exist_ok=True)\n", "os.makedirs(output_dir, exist_ok=True)\n",
"\n", "\n",
"# Full-scale dataset download links\n", "# Full-scale dataset download links\n",
@@ -215,22 +214,17 @@
" name = Path(row[\"audio\"][\"path\"]).stem + \".wav\"\n", " name = Path(row[\"audio\"][\"path\"]).stem + \".wav\"\n",
" output_path = Path(output_dir) / name\n", " output_path = Path(output_dir) / name\n",
"\n", "\n",
" # Read and convert audio\n", " # Check if audio data is valid before writing\n",
" if row[\"audio\"][\"array\"] is None or len(row[\"audio\"][\"array\"]) == 0:\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", " raise ValueError(f\"Empty or invalid audio data in file: {row['audio']['path']}\")\n",
"\n", "\n",
" # Ensure array data is valid before writing\n",
" array_data = row[\"audio\"][\"array\"]\n",
" if not isinstance(array_data, np.ndarray) or len(array_data.shape) == 0:\n",
" raise ValueError(f\"Invalid array data in file: {row['audio']['path']}\")\n",
"\n",
" scipy.io.wavfile.write(\n", " scipy.io.wavfile.write(\n",
" output_path,\n", " output_path,\n",
" 16000,\n", " 16000,\n",
" (array_data * 32767).astype(np.int16),\n", " (row[\"audio\"][\"array\"] * 32767).astype(np.int16),\n",
" )\n", " )\n",
" except (sf.LibsndfileError, ValueError, Exception) as e:\n", " except (sf.LibsndfileError, ValueError, Exception) as e:\n",
" # Log and skip problematic files\n", " # Log the error and skip the file\n",
" print(f\"Error converting {row['audio']['path']}: {e}\")\n", " print(f\"Error converting {row['audio']['path']}: {e}\")\n",
" corrupted_files.append(row[\"audio\"][\"path\"])\n", " corrupted_files.append(row[\"audio\"][\"path\"])\n",
"\n", "\n",
@@ -245,6 +239,7 @@
"\n", "\n",
"print(\"Audioset processing complete!\")\n", "print(\"Audioset processing complete!\")\n",
"\n", "\n",
"\n",
"# -----------------------------\n", "# -----------------------------\n",
"# Download and Process FMA\n", "# Download and Process FMA\n",
"# -----------------------------\n", "# -----------------------------\n",