Skip to content

Latest commit

 

History

History
71 lines (61 loc) · 1.78 KB

INSTALL.md

File metadata and controls

71 lines (61 loc) · 1.78 KB
Raw

Installation

We provide installation instructions for ImageNet classification experiments here.

Dependency Setup

Create an new conda virtual environment

conda create -n convnextv2 python=3.8 -y
conda activate convnextv2

Install Pytorch>=1.8.0, torchvision>=0.9.0 following official instructions. For example:

pip install torch==1.8.0+cu111 torchvision==0.9.0+cu111 -f https://download.pytorch.org/whl/torch_stable.html

Clone this repo and install required packages:

git clone https://github.com/facebookresearch/ConvNeXt-V2.git
pip install timm==0.3.2 tensorboardX six
pip install submitit
conda install openblas-devel -c anaconda -y

Install MinkowskiEngine:

(Note: we have implemented a customized CUDA kernel for depth-wise convolutions, which the original MinkowskiEngine does not support.)

git submodule update --init --recursive
git submodule update --recursive --remote
cd MinkowskiEngine
python setup.py install --blas_include_dirs=${CONDA_PREFIX}/include --blas=openblas

Install apex

git clone https://github.com/NVIDIA/apex
cd apex
pip install -v --disable-pip-version-check --no-cache-dir --global-option="--cpp_ext" --global-option="--cuda_ext" ./
cd ..

Dataset Preparation

Download the ImageNet-1K classification dataset and structure the data as follows:

/path/to/imagenet-1k/
  train/
    class1/
      img1.jpeg
    class2/
      img2.jpeg
  val/
    class1/
      img3.jpeg
    class2/
      img4.jpeg

For pre-training on ImageNet-22K, download the dataset and structure the data as follows:

/path/to/imagenet-22k/
  class1/
    img1.jpeg
  class2/
    img2.jpeg
  class3/
    img3.jpeg
  class4/
    img4.jpeg