search

ohhhyeahhh / SiamCAR

SiamCAR: Siamese Fully Convolutional Classification and Regression for Visual Tracking (CVPR 2020, Oral)
https://openaccess.thecvf.com/content_CVPR_2020/html/Guo_SiamCAR_Siamese_Fully_Convolutional_Classification_and_Regression_for_Visual_Tracking_CVPR_2020_paper.html
322 stars 63 forks source link

关于测试 #78

Closed job333 closed 2 years ago

job333 commented 2 years ago

在测试UAV123的时候，通过可视化发现，测试的速度特别满，这是正常的吗？希望得到您的回复十分感谢！

WangJun-ZJUT commented 2 years ago

您好！这是由于UAV的图像分辨率较大导致的，将可视化功能关闭就可以了。

job333 commented 2 years ago

感谢您的回复，关闭了可视化还是不行，我想把骨干网换成alexnet,修改了config.yaml和config.py中的骨干网名字为alexnet，报错Traceback (most recent call last): File "test.py", line 126, in main() File "test.py", line 57, in main model = ModelBuilder() File "E:\siamese-rpn-pytorch\za\SiamCAR-master\pysot\models\model_builder.py", line 26, in init cfg.BACKBONE.KWARGS) File "E:\siamese-rpn-pytorch\za\SiamCAR-master\pysot\models\backbone__init__.py", line 23, in get_backbone return BACKBONES[name](kwargs) File "E:\siamese-rpn-pytorch\za\SiamCAR-master\pysot\models\backbone\alexnet.py", line 91, in alexnet return AlexNet(**kwargs) TypeError: init() got an unexpected keyword argument 'used_layers'您再给看看怎么解决，谢谢！

Capricorn231 commented 2 years ago

您好，这是因为alexnet与resnet中结构不同导致的配置文件输入有所不同，您可以参考pysot中对应alexnet的配置文件对SiamCAR中的配置文件进行修改。