# MultiVisionModels Platform

## 1.Dependences

| Name  | Version | Build|  Channel |
| -------- |---------| -------- |------|
| python    | 3.12.8  | h3f84c4b_1_cpython|conda-forge|
|pytorch| 2.3.1   |py3.12_cuda12.1_cudnn8_0|pytorch
pytorch-cuda| 12.1    | hde6ce7c_6| pytorch
torchvision| 0.17.2  |pypi_0|pypi
numpy| 1.26.3  |py312h8753938_0|conda-forge
matplotlib| 3.10.0  | pypi_0  |  pypi

A100(40G) train edition

## 2.Overview

Include objection dectection ,keypoint detection,instance segment detection and line dectection.





第3版的修改



D:\python\PycharmProjects\lcnn-master\lcnn_\20250212\MultiVisionModels\models\line_detect\predict.py

加载权重进行训练。设置box和line的阈值，将box和line画在原图上

D:\python\PycharmProjects\lcnn-master\lcnn_\20250212\MultiVisionModels\models\line_detect\predict2.py

加载权重进行训练。设置box和line的阈值，将box和line画在原图上，且一一对应。但没有并行训练

此处将所有一一对应的box与line都画出来，且进行阈值限制，box为0.7，line为0.9