FlowFormer++: Masked Cost Volume Autoencoding for Pretraining Optical Flow Estimation
Xiaoyu Shi*, Zhaoyang Huang*, Dasong Li, Manyuan Zhang, Ka Chun Cheung, Simon See, Hongwei Qin, Jifeng Dai, Hongsheng Li
CVPR 2023
We first pretrain FlowFormer++ on the YoutubeVOS dataset. For finetuning, similar to RAFT, you will need to download the required datasets.
- YoutubeVOS
- FlyingChairs
- FlyingThings3D
- Sintel
- KITTI
- HD1K (optional)
By default datasets.py
will search for the datasets in these locations. You can create symbolic links to wherever the datasets were downloaded in the datasets
folder
├── datasets
├── Sintel
├── test
├── training
├── KITTI
├── testing
├── training
├── devkit
├── FlyingChairs_release
├── data
├── FlyingThings3D
├── frames_cleanpass
├── frames_finalpass
├── optical_flow
conda create --name flowformerpp
conda activate flowformerpp
conda install pytorch=1.6.0 torchvision=0.7.0 cudatoolkit=10.1 matplotlib tensorboard scipy opencv -c pytorch
pip install yacs loguru einops timm==0.4.12 imageio
The script will load the config according to the training stage. The trained model will be saved in a directory in logs
. For example, the following script will load the config configs/pretrain_config.py
. The trained model will be saved as logs/xxxx/final
.
python -u pretrain_FlowFormer_maemask.py --stage youtube
To finish the entire finetuning schedule, you can run:
./run_train.sh
We provide models trained in the four stages. The default path of the models for evaluation is:
├── checkpoints
├── chairs.pth
├── things.pth
├── sintel.pth
├── kitti.pth
├── things_288960.pth
The model to be evaluated is assigned by the _CN.model
in the config file.
Evaluating the model on the Sintel training set and the KITTI training set. The corresponding config file is configs/submissions.py
.
# with tiling technique
python evaluate_FlowFormer_tile.py --eval sintel_validation
python evaluate_FlowFormer_tile.py --eval kitti_validation --model checkpoints/things_kitti.pth
Generating the submission for the Sintel and KITTI benchmarks. The corresponding config file is configs/submissions.py
.
python evaluate_FlowFormer_tile.py --eval sintel_submission
python evaluate_FlowFormer_tile.py --eval kitti_submission
Visualizing the sintel dataset:
python visualize_flow.py --eval_type sintel --keep_size
Visualizing an image sequence extracted from a video:
python visualize_flow.py --eval_type seq
The default image sequence format is:
├── demo_data
├── mihoyo
├── 000001.png
├── 000002.png
├── 000003.png
.
.
.
├── 001000.png
FlowFormer++ is released under the Apache License
@inproceedings{shi2023flowformer++,
title={Flowformer++: Masked cost volume autoencoding for pretraining optical flow estimation},
author={Shi, Xiaoyu and Huang, Zhaoyang and Li, Dasong and Zhang, Manyuan and Cheung, Ka Chun and See, Simon and Qin, Hongwei and Dai, Jifeng and Li, Hongsheng},
booktitle={Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition},
pages={1599--1610},
year={2023}
}
In this project, we use parts of codes in: