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 13:08:02 -06:00
committed by GitHub
parent 9cead7813a
commit 23ca2da7f4
1 changed files with 20 additions and 10 deletions
Download Patch File Download Diff File Expand all files Collapse all files

30
advanced_training_notebook.ipynb
View File

@@ -173,9 +173,19 @@
"# -----------------------------\n",
"# Download and Process Audioset\n",
"# -----------------------------\n",
"from datasets import Dataset, Audio\n",
"import os\n",
"from pathlib import Path\n",
"import scipy.io.wavfile\n",
"from tqdm import tqdm\n",
"\n",
"# Directory setup\n",
"audioset_dir = \"./audioset\"\n",
"os.makedirs(audioset_dir, exist_ok=True)\n",
"\n",
"output_dir = \"./audioset_16k\"\n",
"os.makedirs(output_dir, exist_ok=True)\n",
"\n",
"# Full-scale dataset download links\n",
"dataset_links = [\n",
" f\"https://huggingface.co/datasets/agkphysics/AudioSet/resolve/main/data/bal_train0{i}.tar\"\n",
@@ -192,10 +202,6 @@
" print(f\"Extracting {file_name}...\")\n",
" os.system(f\"tar -xf {out_path} -C {audioset_dir}\")\n",
"\n",
"# Directory for 16kHz WAV files\n",
"output_dir = \"./audioset_16k\"\n",
"os.makedirs(output_dir, exist_ok=True)\n",
"\n",
"# Collect all FLAC files for processing\n",
"audioset_files = list(Path(audioset_dir).glob(\"**/*.flac\"))\n",
"print(f\"Number of FLAC files found: {len(audioset_files)}\")\n",
@@ -212,14 +218,18 @@
" # Define output file name and path\n",
" name = Path(row[\"audio\"][\"path\"]).stem + \".wav\"\n",
" output_path = Path(output_dir) / name\n",
" \n",
"\n",
" # Read and convert audio\n",
" data = row[\"audio\"][\"array\"]\n",
" if data is None or len(data) == 0:\n",
" raise ValueError(f\"Empty audio data in file: {row['audio']['path']}\")\n",
" scipy.io.wavfile.write(output_path, 16000, (data * 32767).astype(np.int16))\n",
" \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",
"\n",
" scipy.io.wavfile.write(\n",
" output_path,\n",
" 16000,\n",
" (row[\"audio\"][\"array\"] * 32767).astype(np.int16),\n",
" )\n",
" except Exception as e:\n",
" # Log and skip problematic files\n",
" print(f\"Error converting {row['audio']['path']}: {e}\")\n",
" corrupted_files.append(row[\"audio\"][\"path\"])\n",
"\n",