To get the project running locally first install the requirements using
pip install -r requiremnets.txt
Once all the requirements are installed, run the data_formation.py to generate the data. If data is already available it will overwrite it with new random shapes.
.
├── data
│ ├── annulus
│ │ ├── test
│ │ ├── train
│ │ └── validation
│ ├── capsule
│ ├── cone
│ ├── cube
│ ├── cylinder
└── └── sphere
After data formation and run the classifier.py to train the on the labeled data. You can change the parameters in the __main__ module of the file.
The model will be saved after the training is completed. The default name for saved model is saved_model.pt.
To get inference on the a .obj file, run the infer.py by assigning the object file path and saved model path to the following variables obj_file_path and model_path respectively.
The infer.py will print out the predicted class label.