GregorySchwing / L2M

Solving Maximum Weighted Matching on Large Graphs with Deep Reinforcement Learning

Solving Maximum Weighted Matching on Large Graphs with Deep Reinforcement Learning

---

Datasets

• Synthetic Datasets, to get implementation from Networkx.

  1. erdos_renyi_graph
  2. barabasi_albert_graph
  3. holme_kim_graph
  4. watts_strogatz_graph

• Real-world Datasets, to get from SNAP.

Compare with

• LwD@Learning What to Defer for Maximum Independent Sets
• DGL-TreeSearch@WHAT’S WRONG WITH DEEP LEARNING IN TREE SEARCH FOR COMBINATORIAL OPTIMIZATION

INSTALL

sudo docker run --gpus all --ipc=host --ulimit memlock=-1 --ulimit stack=67108864 --rm -v ./:/L2M/ -it -p 8888:8888  nvcr.io/nvidia/pytorch:22.07-py3

pip install  dgl -f https://data.dgl.ai/wheels/cu118/repo.html

pip install  dglgo -f https://data.dgl.ai/wheels-test/repo.html

Running

sudo docker run --gpus all --ipc=host --ulimit memlock=-1 --ulimit stack=67108864 --rm -v ./:/L2M/ -it -p 8888:8888  nvcr.io/nvidia/pytorch:22.07-py3

cd /L2M

For training,

python run_l2m.py --mode train

For testing,

python run_l2m.py --mode test