How rifat-binreza/mowaveqformer works
MoWaveQFormer is a research experiment snapshot, not an installable inference package. Its principal workflow is to load BUT PPG data, train and evaluate the notebook’s motion-conditioned, quality-gated heart-rate model, and retain recorded results. The README documents the experiment stages and outputs; notebook internals were not sampled, so the graph does not assert unverified calls or a fully automatic motion-conditioning pipeline.
Main entry points: brnodatajournal.ipynb Experiment notebook · kaggle-run.txt Recorded run · metrics.csv Comparison metrics
Components 11
Notebook experiment
- Experiment notebook
notebooks/brnodatajournal.ipynb - Data preparation
notebooks/brnodatajournal.ipynb - Motion and quality
notebooks/brnodatajournal.ipynb - Model training
notebooks/brnodatajournal.ipynb - Evaluation and baselines
notebooks/brnodatajournal.ipynb - Statistical analysis
notebooks/brnodatajournal.ipynb
Recorded results
- Training outputs
notebooks/brnodatajournal.ipynb - Recorded run
results/logs/kaggle-run.txt - Comparison metrics
results/metrics.csv
Other components
- Researcher
- BUT PPG data
Connections 10
- Researcher → Experiment notebook: runs
README.md - BUT PPG data → Experiment notebook: supplies data
README.md - Experiment notebook → Data preparation: executes cells
README.md - Experiment notebook → Motion and quality: executes stages
README.md - Experiment notebook → Model training: trains model
README.md - Experiment notebook → Evaluation and baselines: runs evaluation
README.md - Experiment notebook → Statistical analysis: runs analysis
README.md - Model training → Training outputs: generates
README.md - Evaluation and baselines → Recorded run: records run
README.md - Recorded run → Comparison metrics: transcribed into
README.md