We customize pointnet.pytorch for the binary tree classification.
We add tree dataloader for the classification and pointnet_demo.ipynb for colab users.
If you want to train this model, you should install below library. If you don't want to install libraries, you can use our demo file too.
- python 3.9.12
- torch 1.0.2
- sklearn 1.10.1+cu102
pip install torch
pip install sklearn
You can use .off or .ply but My code will convert your .off files to .ply files. If you don't want to change it, give convert_off_to_ply parameter to False
If you want to make your own dataset, change paths of dataset (larch, pine) and run train_valid_maker.py
Also, add id.txt to misc directory for your tast. Here is example
python train_valid_maker.py
cd ./utils
python train_classification.py --batchSize [batch size] --num_points [number of points for each data] --nepoch [training epoch] --dataset [dataset path] --dataset_type [dataset type]
example
cd ./utils
python train_classification.py --dataset ./dataset --dataset_type tree
cd ./utils
python train_classification.py --batchSize [batch size] --num_points [number of points for each data] --nepoch [training epoch] --dataset [dataset path] --dataset_type [dataset type] -train False
example
cd ./utils
python train_classification.py --dataset ./dataset --dataset_type tree -train False
If you want to run demo file, you can use pointnet_demo.ipynb.
Oct 24, 2022 - Nov 30, 2022