Faster RCNN训练出现问题：generated_database_->Add(encoded_file_descriptor, size)

最新推荐文章于 2023-02-04 15:41:19 发布

cv.exp

最新推荐文章于 2023-02-04 15:41:19 发布

阅读量2.3k

点赞数

CC 4.0 BY-SA版权

分类专栏： Deep Learning

本文链接：https://blog.youkuaiyun.com/forest_world/article/details/78515438

Deep Learning 专栏收录该内容

172 篇文章

订阅专栏

本文记录了使用Caffe框架实现Faster R-CNN时遇到的错误信息及解决方案，包括忽略某些源层及复制特定源层的过程，并针对特定错误提供了重启MATLAB的解决办法。

摘要生成于 C知道，由 DeepSeek-R1 满血版支持，前往体验 >

I1112 18:49:04.704902 18730 net.cpp:743] Ignoring source layer relu3
I1112 18:49:04.704903 18730 net.cpp:743] Ignoring source layer conv4
I1112 18:49:04.704905 18730 net.cpp:743] Ignoring source layer relu4
I1112 18:49:04.704907 18730 net.cpp:743] Ignoring source layer conv5
I1112 18:49:04.704908 18730 net.cpp:743] Ignoring source layer relu5
I1112 18:49:04.704910 18730 net.cpp:746] Copying source layer roi_pool5
I1112 18:49:04.704917 18730 net.cpp:746] Copying source layer fc6
I1112 18:49:04.726521 18730 net.cpp:746] Copying source layer relu6
I1112 18:49:04.726529 18730 net.cpp:746] Copying source layer drop6
I1112 18:49:04.726531 18730 net.cpp:746] Copying source layer fc7
I1112 18:49:04.736151 18730 net.cpp:746] Copying source layer relu7
I1112 18:49:04.736160 18730 net.cpp:746] Copying source layer drop7
I1112 18:49:04.736176 18730 net.cpp:746] Copying source layer fc7_drop7_0_split
I1112 18:49:04.736178 18730 net.cpp:746] Copying source layer cls_score
I1112 18:49:04.736222 18730 net.cpp:743] Ignoring source layer cls_score_cls_score_0_split
I1112 18:49:04.736225 18730 net.cpp:746] Copying source layer bbox_pred
I1112 18:49:04.736385 18730 net.cpp:743] Ignoring source layer loss
I1112 18:49:04.736389 18730 net.cpp:743] Ignoring source layer accuarcy
I1112 18:49:04.736392 18730 net.cpp:743] Ignoring source layer loss_bbox
[libprotobuf ERROR google/protobuf/descriptor_database.cc:57] File already exists in database: caffe.proto
[libprotobuf FATAL google/protobuf/descriptor.cc:1018] CHECK failed: generated_database_->Add(encoded_file_descriptor, size):

%% -------------------- CONFIG --------------------
opts.caffe_version          = 'caffe_faster_rcnn';
opts.gpu_id                 = auto_select_gpu;
active_caffe_mex(opts.gpu_id, opts.caffe_version);

解决方法：