Skip to content

Latest commit

 

History

History
32 lines (24 loc) · 1.02 KB

README.md

File metadata and controls

32 lines (24 loc) · 1.02 KB
Raw

A Dynamic Multi-Scale Voxel Flow Network for Video Prediction

Introduction

This project is the implement of A Dynamic Multi-Scale Voxel Flow Network for Video Prediction. Our paper is accepted by CVPR2023. 论文中文介绍

Usage

Installation

git clone https://github.com/huxiaotaostasy/DMVFN.git
cd 
pip3 install -r requirements.txt

  • Download the pretrained models from Google Drive. (百度网盘链接: 密码:，把压缩包解开后放在 pretrained_models/*)

  • Unzip and move the pretrained parameters to pretrained_models/*

Run

Train

python3 -m torch.distributed.launch --nproc_per_node=8 --master_port=4321 train.py --train_dataset CityTrainDataset --val_datasets CityValDataset --batch_size 8 --num_gpu 8

Test

python3 -m torch.distributed.launch --nproc_per_node=1 --master_port=4321 test.py --val_datasets CityValDataset --load_path /data/dmvfn/train_log_CityTrainDataset/DMVFN/dmvfn_299.pkl

Results