{"id":51402595,"url":"https://github.com/adi21-dev/stuttering","last_synced_at":"2026-07-04T08:03:05.026Z","repository":{"id":325932357,"uuid":"1102899037","full_name":"adi21-dev/stuttering","owner":"adi21-dev","description":null,"archived":false,"fork":false,"pushed_at":"2025-11-25T14:29:58.000Z","size":178,"stargazers_count":0,"open_issues_count":0,"forks_count":1,"subscribers_count":0,"default_branch":"master","last_synced_at":"2026-06-18T10:08:11.165Z","etag":null,"topics":[],"latest_commit_sha":null,"homepage":null,"language":"Python","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":null,"status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/adi21-dev.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":null,"code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null,"zenodo":null,"notice":null,"maintainers":null,"copyright":null,"agents":null,"dco":null,"cla":null}},"created_at":"2025-11-24T07:23:21.000Z","updated_at":"2025-11-25T14:30:10.000Z","dependencies_parsed_at":null,"dependency_job_id":null,"html_url":"https://github.com/adi21-dev/stuttering","commit_stats":null,"previous_names":["thecoder1232/stuttering","adi21-dev/stuttering"],"tags_count":null,"template":false,"template_full_name":null,"purl":"pkg:github/adi21-dev/stuttering","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/adi21-dev%2Fstuttering","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/adi21-dev%2Fstuttering/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/adi21-dev%2Fstuttering/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/adi21-dev%2Fstuttering/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/adi21-dev","download_url":"https://codeload.github.com/adi21-dev/stuttering/tar.gz/refs/heads/master","sbom_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/adi21-dev%2Fstuttering/sbom","scorecard":null,"host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":286080680,"owners_count":35114174,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2026-05-26T15:22:16.424Z","status":"online","status_checked_at":"2026-07-04T02:00:05.987Z","response_time":113,"last_error":null,"robots_txt_status":"success","robots_txt_updated_at":"2025-07-24T06:49:26.215Z","robots_txt_url":"https://github.com/robots.txt","online":true,"can_crawl_api":true,"host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":[],"created_at":"2026-07-04T08:03:04.491Z","updated_at":"2026-07-04T08:03:05.012Z","avatar_url":"https://github.com/adi21-dev.png","language":"Python","funding_links":[],"categories":[],"sub_categories":[],"readme":"# Stuttering Detection Platform\n\nThis repository contains code for detecting and classifying stuttering events in speech audio using deep learning (PyTorch). It is designed for research and experimentation as part of a final year project.\n\n## Features\n- **Frame-level stutter detection** using Wav2Vec2-based models\n- **Custom preprocessing pipeline** for converting raw Parquet datasets to training-ready format\n- **Training scripts** for both CPU and GPU environments\n- **Inference scripts** for analyzing new audio files\n- **Class imbalance handling** via computed class weights\n- **Evaluation and reporting** of model performance\n\n## Directory Structure\n```\n├── dataset/\n│   ├── raw/           # Raw Parquet files (input)\n│   ├── processed/     # Processed samples (.pt files, class weights)\n├── models/            # Saved model checkpoints (.pth)\n├── docs/              # Documentation and plans\n├── extras/            # Utility scripts and notebooks\n├── preprocessing.py   # Main preprocessing script\n├── train_gpu.py       # Training script (GPU)\n├── train_cpu.py       # Training script (CPU)\n├── inference.py       # Inference script\n├── inference2.py      # Alternative inference script\n├── test_dataset.py    # Dataset validation script\n```\n\n## Setup Instructions\n1. **Clone the repository**\n2. **Install Python 3.8+** and recommended packages:\n   ```bash\n   pip install torch torchaudio transformers pandas pyarrow tqdm librosa\n   ```\n3. (Optional) For Jupyter notebooks:\n   ```bash\n   pip install notebook\n   ```\n4. (Optional) For GPU support, ensure CUDA is installed and available.\n\n## Data Preparation\n- Place your raw Parquet files in `dataset/raw/`.\n- Run the preprocessing script to generate processed samples:\n  ```bash\n  python preprocessing.py\n  ```\n- Processed files will be saved in `dataset/processed/`.\n\n## Training\n- **GPU:**\n  ```bash\n  python train_gpu.py\n  ```\n- **CPU:**\n  ```bash\n  python train_cpu.py\n  ```\n- Model checkpoints are saved in `models/`.\n\n## Inference\n- Use `inference.py` or `inference2.py` to analyze new audio files:\n  ```bash\n  python inference.py --audio_path path/to/audio.wav --model_path models/stutter_model_epoch_6.pth\n  ```\n  (Edit the script to set paths as needed)\n\n## Classes\n- `fluent` (0)\n- `block` (1)\n- `word_rep` (2)\n- `syllab_rep` (3)\n- `prolongation` (4)\n\n## Dataset Reference\n- Synthetic stuttering dataset: [miosipov/Stutter_EN_Synthetic](https://huggingface.co/datasets/miosipov/Stutter_EN_Synthetic)\n- Parquet schema: audio samples, metadata (start_time, end_time, stutter_type), transcription\n\n## Example Workflow\n1. Download dataset and place in `dataset/raw/`\n2. Run `preprocessing.py` to create processed samples\n3. Train model with `train_gpu.py` or `train_cpu.py`\n4. Run inference on new audio with `inference.py`\n\n## Contributors\n- Aditya (Final Year Project)\n\n## License\nThis project is for academic/research use. Please cite appropriately if used.\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fadi21-dev%2Fstuttering","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fadi21-dev%2Fstuttering","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fadi21-dev%2Fstuttering/lists"}