How sirichandana-jpg/indian-folk-music-classification works
This is an Indian folk music genre-classification system: users provide audio and receive a predicted genre with confidence, while the project also supports feature analysis, model training, and evaluation. The graph shows the Streamlit inference route and the documented training/model capabilities. Preparation and evaluation details are limited to what the README and source index establish; unsampled relationships are omitted.
Main entry points: model_factory.py Model factory · app.py Streamlit interface · trainer.py Training engine · app_utils.py Inference helpers
Components 18
Audio inference
- Streamlit interface
streamlit_app/app.py - Inference helpers
streamlit_app/app_utils.py
Signal processing
- Audio preprocessing
preprocessing/audio_preprocessor.py - Feature extraction
feature_extraction/feature_extractor.py - Audio visualization
visualization/audio_visualizer.py
Model learning
- Audio augmentation
augmentation/augmentor.py - Training engine
training/trainer.py - Training metrics
training/metrics.py - Model factory
models/model_factory.py - Model base
models/base_model.py - 2D CNN
models/cnn_2d.py - CNN BiLSTM
models/cnn_bilstm.py - CRNN
models/crnn.py - Training entry
train.py - Model evaluation
evaluate_crnn.py
Shared support
- Configuration loader
utils/config_parser.py - Logging
utils/logger.py
Other components
- Audio user
Connections 22
- Audio user → Streamlit interface: provides audio
streamlit_app/app.py - Streamlit interface → Audio preprocessing: preprocesses audio
streamlit_app/app.py - Streamlit interface → Inference helpers: requests inference
streamlit_app/app.py - Streamlit interface → Feature extraction: extracts display features
streamlit_app/app.py - Streamlit interface → Audio visualization: renders plots
streamlit_app/app.py - Inference helpers → Feature extraction: extracts model features
streamlit_app/app_utils.py - Inference helpers → Model factory: builds model
streamlit_app/app_utils.py - Inference helpers → Audio user: returns genre scores
streamlit_app/app_utils.py - Training entry → Configuration loader: loads settings
train.py - Training entry → Model factory: builds model
train.py - Training entry → Training engine: starts training
train.py - Training engine → Audio augmentation: augments samples
training/trainer.py - Training engine → Feature extraction: extracts features
training/trainer.py - Training engine → Training metrics: computes metrics
training/trainer.py - Model factory → 2D CNN: constructs
models/model_factory.py - Model factory → CNN BiLSTM: constructs
models/model_factory.py - Model factory → CRNN: constructs
models/model_factory.py - 2D CNN → Model base: inherits from
models/cnn_2d.py - CNN BiLSTM → Model base: inherits from
models/cnn_bilstm.py - CRNN → Model base: inherits from
models/crnn.py - Model evaluation → Model factory: builds model
evaluate_crnn.py - Model evaluation → Training engine: evaluates model
evaluate_crnn.py