https://github.com/zyc-cver/m2beats-2.0
This is the official code implementation for 《M2Beats 2.0: When Motion Meets Beats in Short-form Videos Twice》. More details will be released once the paper is published!
https://github.com/zyc-cver/m2beats-2.0
motion music rhythm
Last synced: 1 day ago
JSON representation
This is the official code implementation for 《M2Beats 2.0: When Motion Meets Beats in Short-form Videos Twice》. More details will be released once the paper is published!
- Host: GitHub
- URL: https://github.com/zyc-cver/m2beats-2.0
- Owner: zyc-cver
- Created: 2025-02-19T15:32:18.000Z (8 months ago)
- Default Branch: main
- Last Pushed: 2025-04-15T07:12:03.000Z (6 months ago)
- Last Synced: 2025-04-15T08:24:26.269Z (6 months ago)
- Topics: motion, music, rhythm
- Language: Python
- Homepage: https://zyc-cver.github.io/M2Beats-2.0/
- Size: 18.5 MB
- Stars: 1
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
M2Beats 2.0: When Motion Meets Beats in Short-form Videos Twice
## Installation
### 1. Create Conda Environment
```bash
conda create -n m2beats2 python=3.9 -y
conda activate m2beats2
```
### 2. Install Required Packages
```bash
pip install numpy==1.24.0 torch==1.12.1
```
## Model
Download our pretrained model from [Google Drive](https://drive.google.com/drive/folders/1DsTspcqAeyAo-SJl2rfyV4k_PWm3Haer?usp=sharing).
## Dataset
Download our dataset from [Google Drive](https://drive.google.com/file/d/1DkUsyZ58WW3TMauFbW7tggET46uFjP_g/view?usp=sharing).
The dataset follows the directory structure below:
```bash
AIST_M2B_2/
├── data_split/
├── test_data_split.csv
├── train_data_split.csv
```
Original video and motion data can be downloaded from [AIST++](https://google.github.io/aistplusplus_dataset/factsfigures.html)
## Evaluation
Run the following command to evaluate the model:
```bash
python eval.py --checkpoint checkpoints/params.pt --test AIST_M2B_2/train_data_split.csv --test_data AIST_M2B_2/data_split
```
## Test
### Testing on Sample Data from the Dataset
```bash
python test.py --checkpoint checkpoints/params.pt --test AIST_M2B_2/train_data_split.csv --test_data AIST_M2B_2/data_split
```
### Testing on Your Own Data
If you want to test on your own data, you need to extract 2D human keypoints from video data beforehand. We recommend using [mmpose](https://github.com/open-mmlab/mmpose).
---
More details will be available once our paper is published!