Installation
ONNX
1. Create a conda environment with python 3.8 (3.9 for pytorch110):
conda create -n <env_name> python=3.8
conda activate <env_name>
#### 2. We use pytorch 1.9.0, torchvision 0.10.0, and cudatoolkit 11.1:
2. We use pytorch:
Please refer to https://pytorch.org/get-started/locally/ for installing.
conda install pytorch torchvision torchaudio cudatoolkit=11.1 -c pytorch-lts -c nvidia
or
conda install pytorch torchvision torchaudio cudatoolkit=11.1 -c pytorch-nightly -c nvidia
3. Use pip to install some requirements:
pip install matplotlib easydict tqdm lmdb tensorboard
pip install transformers # for BERT https://huggingface.co/transformers/
pip install opencv_python
Albumentations for data augmentation:
4. We usepip install -U albumentations
5. Install TensorRT-8.0.1.6
Evaluation
Training
Acknowledgments
Thanks for the great DETR and pytracking.
- We borrow
transformer
from DETR. - We refer to pytracking for data augmentation.
License
Our work is released under the GPL 3.0 license. Please see the LICENSE file for more information.