Skip to main content

图像场景图生成使用说明

1. 算法描述

该算法用于图像场景图生成(SGG)。该算法基于PyTorch框架开发,输入一张图片,算法会检测图片中人和物的交互关系,输出图像场景图。

2. 依赖及安装

CUDA版本: 11.3 其他依赖库的安装命令如下:

conda install pytorch==1.10.1 torchvision==0.11.2 torchaudio==0.10.1 cudatoolkit=11.3 -c pytorch -c conda-forge

可使用如下命令下载安装算法包:

pip install -U mmkg-sgg

3. 使用示例及运行参数说明

需要准备 pretrained_ckpt, labelmap_file, freq_prior 三个文件,可以通过 mmkg-sgg 上所提供的链接进行下载

from mmkg_vrd import SGG

sgg = SGG(
    pretrained_ckpt='custom_io/ckpt/RX152FPN_reldn_oi_best.pth',
    labelmap_file='custom_io/ji_vrd_labelmap.json', 
    freq_prior='custom_io/vrd_frequency_prior_include_background.npy', 
)

sgg.det_and_vis(
    img_file='custom_io/imgs/1024px-Gen_Robert_E_Lee_on_Traveler_at_Gettysburg_Pa.jpg',
    save_file='custom_io/out/1024px-Gen_Robert_E_Lee_on_Traveler_at_Gettysburg_Pa.reldn_relation.jpg')

print('all done')

Metadata

Release files for mmkg-sgg 0.2.0

For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.

Built distribution (wheel)

Table of built distributions (wheels) for mmkg-sgg 0.2.0
File Interpreter ABI Platform
mmkg_sgg-0.2.0-cp37-cp37m-manylinux1_x86_64.whl CPython 3.7 CPython 3.7 pymalloc Linux glibc 2.5+ x86-64 Details

Release files / mmkg_sgg-0.2.0-cp37-cp37m-manylinux1_x86_64.whl

Download URL mmkg_sgg-0.2.0-cp37-cp37m-manylinux1_x86_64.whl
Size 5.5 MB
Tags CPython 3.7 CPython 3.7 pymalloc Linux glibc 2.5+ x86-64
SHA-256 checksum
How to use checksums
16f0d8661e30902bbd52857ecb6890a9fe87831ca7e8fd295c6ba1bcbf81bc0c
BLAKE2b-256 checksum
How to use checksums
749fa3d0b16760455542684865c54c2b771dd29ebe3620c171285071d5b97082
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
No
Uploaded via twine/4.0.2 CPython/3.7.16

Release history Release notifications | RSS feed

This release

0.2.0 This release

1 release file

0.1.5

1 release file

0.1.4

1 release file

0.1.3

1 release file

0.1.1

1 release file

0.1.0

1 release file

Anthropic, PBC Visionary sponsor Bloomberg Visionary sponsor Hudson River Trading Visionary sponsor Meta Visionary sponsor NVIDIA Visionary sponsor Microsoft Sustainability sponsor Depot Continuous Integration AWS Cloud computing and Security Sponsor Datadog Monitoring Fastly CDN Google Download Analytics Sentry Error logging StatusPage Status page