Skip to content

Latest commit

 

History

History
26 lines (20 loc) · 663 Bytes

File metadata and controls

26 lines (20 loc) · 663 Bytes

Testing Environments

  • Linux
  • Python 3.7.9
  • PyTorch 1.9.0
  • CUDA 10.2
  • GCC 7.5.0
  • mmcv-full 1.3.9

Note that the mmcv and dgl version must be compatable with your torch and cuda version!

Package Requirements

a. Prepare the testing environments.

b. Install additional packages.

pip install -r requirements.txt

Hardware Requirements

  • 60GB RAM for loading the whole dataset.

    • 60GB for Apparel dataset
    • 40GB for Food dataset
    • 20GB for Cosmetic dataset
  • At least 10GB GPU memory to reproduce RADAR model. Some baselines and ablation studies requires more.