SJTU CS3308 Machine Learning final project, created by [Junxuan Liang], [Jiang Nan], [Yue Yuan].

git clone https://github.com/hitefork/ML-proj-BPP.git
conda create -n binpacking python=3.8
conda activate binpacking
cd ML-proj-BPP
pip install -r requirements.txt

Our trained model checkpoints can be found in Link

python task1.py --load_path="your path to task1.pt"

After running the instruction, you could check /Box_data/task1.jpg to see the visualization

python task2.py --load_path="your path to task2.pt"

After running the instruction, you could check /Box_data/task2.jpg to see the visualization

python task3.py --load_path="your path to task3.pt"

python task3.py --load_path="your path to task3_online.pt"

After running the instruction, you could check /Box_data/task3_xx*xx*xx.jpg to see the visualization

See demo_sb3_train_ipynb and demo_packing_env.ipynb. The name of model weights is "ppo_mask_interm_10000.zip",you could check /gifs/strain_5_boxes.gif to see the visualization