m0sys / squalifier

Front-squat vs Back-squat classification web app.

Welcome to squalifier 👋

Front-squat vs Back-squat classification web app.

See notebooks for model demo.

This Deeplearning application was trained on a small curated dataset of images crawled off of google using a pretrained Resnet34 as the base. Click here or here to see an example implementation of Resnet models.

Install

See setup.md to get started!

Training Models

---

To train models run the following:

CNN Classification Model

python training/run_experiment.py --save '{"dataset": "FvbsDataset", "model": "CnnClassificationModel", "network": "resnet34"}'

or

python ./tasks/train_simple_cnn_classification_model_on_fvbs.sh

World Class CNN Classification Model (8-10% error rate)

python ./tasks/train_world_class_cnn_classification_model_on_fvbs.sh

Usage

see setup.md

Run tests

test/tests.py

Author

👤 m0sys

• Github: @m0sys

Show your support

Give a ⭐️ if this project helped you!

---