capstoneml

FPGA Bitstreams are required to run certain scripts on the Ultra96. The bitstreams were intentionally left out of the repo as since users can already be compiled from the code in this repo.

notebook

In notebooks, we have the notebooks used to train and evaluate models on both local computer and on the ultra96 board.

Training Model:

• conv_trainer.ipynb
• explore_conv.ipynb: Used during exploration of a CNN model
• main.ipynb: Training model on Unseen Data

Feature Extraction:

• feature_playground.ipynb: Playground to explore feature extraction
• feature_emd.ipynb: Exploration of more features

Evaluations:

• main_clean.ipynb: locally evaluate and test out algorithm
• eval_conv.ipynb: Evaluate and test algorithm with FPGA
• video.ipynb: Notebook used during Idividual Progress Check

Scripts

Main Scripts on Ultra96

Actual Implementation:

• ai.py
• start_detector.py
• modelling_utils.py
• filepaths_extComms.py
• dataloader.py

All Notebooks

• cpu_implement.py
• dataloader_notebook.py
• filepaths.py
• filereader.py
• model_utils_training.py
• visualisers.py

Ultra96 Notebook Exclusive Use:

• ai_notebook.py
• start_detector_notebook.py

Local Notebook Use:

• ai_tensor.py
• start_detector_tensor.py