|
|
|||
|---|---|---|---|
| .riahub/workflows | |||
| README.md | |||
RIA Hub Technical Demo
This repository demonstrates a full ML pipeline via Gitea Actions:
-
Recordings
A collection of raw.npyradio recordings stored via Git LFS. -
Workflows
A CI pipeline that automatically:- Builds a labeled dataset from raw recordings
- Trains a model on that dataset
- Optimizes the model and packages an inference application
-
Scripts
scripts/build_dataset.sh
Reads throughrecordings/, applies preprocessing, and outputs training.npzor.csvfiles intodata/.scripts/train_model.sh
Consumesdata/, trains a PyTorch model, and writes checkpoints tocheckpoints/.scripts/build_app.sh
Takes the best checkpoint and builds a small inference CLI or server indist/.