# crossgcn

**Repository Path**: chenjoya/crossgcn

## Basic Information

- **Project Name**: crossgcn
- **Description**: No description available
- **Primary Language**: Unknown
- **License**: Not specified
- **Default Branch**: master
- **Homepage**: None
- **GVP Project**: No

## Statistics

- **Stars**: 6
- **Forks**: 0
- **Created**: 2020-10-30
- **Last Updated**: 2025-08-05

## Categories & Tags

**Categories**: multimedia

**Tags**: None

## README

# Cross Graph Align (CrossGCN)

## Acknowledgements
This repository is supported by **2018YFB1004300**.
```
Chen J, Du H, Wu Y F, et al. Cross-modal video moment retrieval based on visual-textual relationship alignment (in
Chinese). Sci Sin Inform, 2020, 50: 862–876, doi: 10.1360/SSI-2019-0292
```

### Requirements:
- CUDA >= 9.0

### Installation
```bash
# create and activate a clean conda env
conda create -n crossgcn
conda activate crossgcn 

# install the right pip and dependencies for the fresh python
conda install ipython pip

# install some dependencies
pip install yacs h5py terminaltables tqdm
# Note: you can use tsinghua mirror to speed up downloading if you are in China
# pip install yacs h5py terminaltables tqdm -i https://pypi.tuna.tsinghua.edu.cn/simple

# follow PyTorch installation in https://pytorch.org/get-started/locally/
# we give the instructions for CUDA 10.1, others are also okay
conda install pytorch torchvision cudatoolkit=10.1 torchtext -c pytorch
# Note: you can use tsinghua mirror to speed up downloading if you are in China
# conda install pytorch torchvision cudatoolkit=10.1 torchtext -c https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud/pytorch/

# clone crossgcn and enjoy it!
git clone https://github.com/ChenJoya/crossgcn
```

### Results

**1. TACoS Dataset**
| Repo | Rank1@0.1 | Rank1@0.3 | Rank1@0.5 | Rank5@0.1 | Rank5@0.3 | Rank5@0.5 |
| ---- |:-------------:| :-----:|:-----:|:-----:|:-----:|:-----:|
| CrossGCN | 51.9 | 39.8 | 26.4 | 74.5 | 60.0 | 47.2 |

**2. ActivityNet Dataset**
| Repo | Rank1@0.3 | Rank1@0.5 | Rank1@0.7 | Rank5@0.3 | Rank5@0.5 | Rank5@0.7 |
| ---- |:-------------:| :-----:|:-----:|:-----:|:-----:|:-----:|
| CrossGCN | 62.7 | 47.2 | 27.9 | 88.1 | 79.1 | 64.2 |

## Dataset
Please refer to [DATASET.md](DATASET.md) to prepare datasets.

## Quick Start
We provide scripts for simplifying training and inference. Please refer to [scripts/train.sh](scripts/train.sh), [scripts/eval.sh](scripts/eval.sh).

## Support
Please open a new issue. We would like to answer it. Please feel free to contact me: chenjoya@foxmail.com if you need my help.