Go to file

Liyu Xiao 19d151872d removed warnings during conversion part of workflow demo		2025-05-27 14:35:55 -04:00
.riahub/workflows	removed warnings during conversion part of workflow demo	2025-05-27 14:35:55 -04:00
checkpoint_files	formatted all files	2025-05-22 14:12:36 -04:00
conf	added in final changes to now do 5 epochs, and added in conversion to ORT at the end of workflow	2025-05-26 09:44:53 -04:00
data	increased the batch size for every update	2025-05-26 15:02:57 -04:00
helpers	added in final changes to now do 5 epochs, and added in conversion to ORT at the end of workflow	2025-05-26 09:44:53 -04:00
onnx_files	completed converting the pytorch model -> ONNX	2025-05-22 14:12:10 -04:00
onnx_scripts	added profiling for the onnx model	2025-05-26 10:28:11 -04:00
wheel	reversed the changes, for now will install from the wheel folder	2025-05-16 14:17:08 -04:00
.gitattributes	added LFS tracking to the .whl files	2025-05-16 13:23:13 -04:00
.gitignore	optimized script	2025-05-26 12:04:20 -04:00
README.md	Add ReadME	2025-05-14 15:36:31 -04:00
requirements.txt	added profiling for the onnx model	2025-05-26 10:28:11 -04:00

RIA Hub Technical Demo

This repository demonstrates a full ML pipeline via Gitea Actions:

Recordings
A collection of raw .npy radio recordings stored via Git LFS.
Workflows
A CI pipeline that automatically:
1. Builds a labeled dataset from raw recordings
2. Trains a model on that dataset
3. Optimizes the model and packages an inference application
Scripts
- scripts/build_dataset.sh
  Reads through recordings/, applies preprocessing, and outputs training .npz or .csv files into data/.
- scripts/train_model.sh
  Consumes data/, trains a PyTorch model, and writes checkpoints to checkpoints/.
- scripts/build_app.sh
  Takes the best checkpoint and builds a small inference CLI or server in dist/.