Requirements

python3

sudo apt-get install python3-dev

sudo apt-get install python3.8-dev

python3 -m venv venv
pip install pip --upgrade

pip install torch
pip install torchvision
pip install opencv-python
pip install tensorboardX
pip install tensorboard
pip install configargparse
pip install kornia
pip install matplotlib pip install pytorch3d

no longer required:

pip install pptk

dev-extension

(req python3.8-dev) pip install wandb

wandb login

cuda 10.1: pip install torch==1.6.0+cu101 torchvision==0.7.0+cu101 -f https://download.pytorch.org/whl/torch_stable.html

Data Structure

calibration to kitti-flow directory:

<datasets_dir>/KITTI_flow/data_scene_flow_calib

train-dataset:
<datasets_dir>/KITTI_flow_multiview/testing/image_2

eval-dataset:
<datasets_dir>/KITTI_flow/training/image_2'
<datasets_dir>/KITTI_flow/training/flow_occ'
<datasets_dir>/KITTI_flow/training/flow_noc'

models: .models

Run

note: in case you want to continue to train a model use the model_tag. e.g. '2020_08_12_v65'

python coach.py -s config/config_setup_0.yaml -c config/config_coach_def_usceneflow.yaml -e config/exps/r10/config_coach_exp_r10e0_disp_se3_mask3_val1.yaml

python presenter.py -s config/config_setup_0.yaml -c config/config_coach_def_usceneflow.yaml -e config/present/config_presenter_0.yaml