- Python>=3.7.0
- tensorflow==2.8.0
- pip install -r requirements.txt
The dataset is available on download here
python train.py --path path/to/dataset --epochs 12 --model_dir path/for/model
- path: path to dataset
- epochs: number of epochs
- model_dir: path where weights will be saved
python test.py --path path/to/dataset --model path/for/trained/model
The result on test data will be printed.
- path: path to dataset
- model: path to model
python inference.py --image path/to/image --model path/for/trained/model
The result of inference will be printed.
- image: path, image from dataset
- model: path to model
python image_projection.py --path path/to/data --model path/for/trained/model
The resulted image will be shown on a new window.
- path: path to test data, or data which will be projected
- model: path to model
cd flask
python run.py --path path/for/trained/model
- path: a required path to weights
Refer to youtube.
| Name | Test accuracy | Model |
|---|---|---|
| ResNet50 | 0.9821 | model |
