Skip to content

Latest commit

 

History

History
44 lines (35 loc) · 1.67 KB

INSTALL.md

File metadata and controls

44 lines (35 loc) · 1.67 KB

Preparation for pre-training & ImageNet fine-tuning

Pip dependencies

  1. Prepare a python environment, e.g.:
$ conda create -n spark python=3.8 -y
$ conda activate spark
  1. Install PyTorch and timm (better to use torch~=1.10, torchvision~=0.11, and timm==0.5.4) then other python packages:
$ pip install torch==1.10.0+cu113 torchvision==0.11.1+cu113 -f https://download.pytorch.org/whl/torch_stable.html
$ pip install timm==0.5.4
$ pip install -r requirements.txt

It is highly recommended to install these versions to ensure a consistent environment for re-implementation.

ImageNet preparation

Prepare the ImageNet-1k dataset

  • assume the dataset is in /path/to/imagenet
  • it should look like this:
/path/to/imagenet/:
    train/:
        class1: 
            a_lot_images.jpeg
        class2:
            a_lot_images.jpeg
    val/:
        class1:
            a_lot_images.jpeg
        class2:
            a_lot_images.jpeg
  • that argument of --data_path=/path/to/imagenet should be passed to the training script introduced later

PS: In our implementation, we use pytorch built-in operators to simulate the submanifold sparse convolution in encoder.py for generality, due to the fact that many convolution operators (e.g., grouped conv and dilated conv) do not yet have efficient sparse implementations on today's hardware. If you want to try those sparse convolution, you may refer to this sparse convolution library or MinkowskiEngine.