INSTALL.md

AdaMAE Installation

The codebase is mainly built with following libraries:

• Python 3.6 or higher

• PyTorch and torchvision.
  We can successfully reproduce the main results under two settings below:
  Tesla A100 (40G): CUDA 11.1 + PyTorch 1.8.0 + torchvision 0.9.0
  Tesla V100 (32G): CUDA 10.1 + PyTorch 1.6.0 + torchvision 0.7.0

• timm==0.4.8/0.4.12

• deepspeed==0.5.8

  DS_BUILD_OPS=1 pip install deepspeed

• TensorboardX

• decord

• einops

Note:

• We have observed accidental interrupt in the last epoch when conduct the pre-training experiments on V100 GPUs (PyTorch 1.6.0). This interrupt is caused by the scheduler of learning rate. We naively set --epochs 801 to walk away from issue :)

Acknowledgement

Instructions copied from the VideoMAE codebase. We thank the authors of the VideoMAE for making their code available online!