bimahasto/antispoofing-ockd

Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference

Local Gradio UI

This folder contains a Gradio demo (app_gradio.py) and a small entrypoint app.py that launches it. The demo scans available trained models (folders under implementasi_new/code/models/ and single .pt files under deploy/antispoofing-ockd/models/).

Quick steps to run locally (Windows PowerShell):

1. 1.Create and activate a Python 3.8+ environment and install requirements:

python -m venv .venv; .\.venv\Scripts\Activate.ps1; pip install -r requirements.txt

1. 1.Start the Gradio UI:

python app.py

1. 1.The UI will open in your browser (default port 7861). Select a model from the dropdown, upload one .wav file, pick a scenario (1/2/3), and click Run. The UI returns a single numeric score and a label ("bonafide"/"spoof").

Notes on Hugging Face Spaces:

• —This project is Gradio-based and can be deployed to Hugging Face Spaces. The repo already contains an app.py entrypoint and requirements.txt — Spaces will run app.py by default. You may need to pin exact versions for reproducibility.
• —Keep in mind: Spaces has resource limits (CPU/RAM) and may not support a large PyTorch GPU workload. For CPU-based demo it should work; for GPU you'd need a paid GPU-enabled space.