##参数设置MODEL:BACKBONE:NAME:"build_resnet_fpn_backbone"RESNETS:OUT_FEATURES:["res2","res3","res4","res5"]FPN:IN_FEATURES:["res2","res3","res4","res5"] (相同道理,全局搜索build_resnet_fpn_backbone,如果有能运行起来demo的话,那就单步跟进最方便) 改环节主要包含两个部分: 创建基网络...
```# 等价于如下foo=registry.BACKBONES[cfg.MODEL.BACKBONE.CONV_BODY]backbone=foo(cfg)returnbackbone```returnregistry.BACKBONES[cfg.MODEL.BACKBONE.CONV_BODY](cfg) 至此backbone的获取过程就介绍完了,本篇内容并没有涉及到backbone中ResNet网络,FPN网络是如何构造的,如果需要了解相关细节,可以查看其源代码。
下面分别是key为model、train_cfg、test_cfg的源代码(该配置文件为faster_rcnn_r50_fpn_1x.py) model: AI检测代码解析 # model settings model = dict( type='FasterRCNN', pretrained='torchvision://resnet50', backbone=dict( type='ResNet', depth=50, num_stages=4, out_indices=(0, 1, 2, 3...
SpineNet is a backbone model specific for detection, it's a backbone but can do FPN's thing!! More info pls reference google's paperlink. fromnb.torch.bakbones.spinenetimportSpineNetmodel=SpineNet() 2020.09.11: New added blocks: Support Matrix ...
_, C2, C3, C4, C5 = resnet_graph(input_image, "resnet101", stage5=True) #构建共享卷积层。 #自下而上的图层 #返回每个阶段的最后一个图层列表,共5个。 #不要创建thead(阶段5),所以我们选择列表中的第4个项目。 # Top-down Layers