Please follow MonST3R and Spann3R to download ScanNet, TUM-dynamics,Sintel, Bonn, KITTI and 7scenes datasets.
To prepare the ScanNet dataset, execute:
python datasets_preprocess/long_prepare_scannet.py # You may need to change the path of the datasetTo prepare the TUM-dynamics dataset, execute:
python datasets_preprocess/long_prepare_tum.py # You may need to change the path of the datasetTo prepare the Bonn dataset, execute:
python datasets_preprocess/long_prepare_bonn.py # You may need to change the path of the datasetTo prepare the KITTI dataset, execute:
python datasets_preprocess/long_prepare_kitti.py # You may need to change the path of the datasetResults will be saved in eval_results/*.
CUDA_VISIBLE_DEVICES=6,7 bash eval/relpose/run_scannet.sh # You may need to change [--num_processes] to the number of your gpus and choose sequence length in datasets=('scannet_s3_1000')
CUDA_VISIBLE_DEVICES=6,7 bash eval/relpose/run_tum.sh # You may need to change [--num_processes] to the number of your gpus and choose sequence length in datasets=('tum_s1_1000')
CUDA_VISIBLE_DEVICES=6,7 bash eval/relpose/run_sintel.sh # You may need to change [--num_processes] to the number of your gpusCUDA_VISIBLE_DEVICES=5 bash eval/video_depth/run_kitti.sh # You may need to change [--num_processes] to the number of your gpus and choose sequence length in datasets=('kitti_s1_500')
CUDA_VISIBLE_DEVICES=5 bash eval/video_depth/run_bonn.sh # You may need to change [--num_processes] to the number of your gpus and choose sequence length in datasets=('bonn_s1_500')
CUDA_VISIBLE_DEVICES=5 bash eval/video_depth/run_sintel.sh # You may need to change [--num_processes] to the number of your gpusCUDA_VISIBLE_DEVICES=5 bash eval/mv_recon/run.sh # You may need to change [--num_processes] to the number of your gpus and hoose sequence length in max_frames