CangHaiQingYue

这个仓库的代码应该没问题吧，大佬这边不是复现了么。我跑起来后发现一切都正常。 我自己的代码，训练10epoch后mAP就开始降了，完了才慢慢升，就很迷惑。

大佬你好，我是你的粉丝！

感觉像是数据集的问题，换了另一个数据集后，报错 “Error in './darknet': double free or corruption (!prev): 0x00007f8da0001110"

> @CangHaiQingYue，我也遇到了同样的问题，或许是代码里面data.c 里面的问题。我修改了data.c里面的第1421,1426,1515行的5xboxes改为 15xboxes就可以了。 原因是，在原yolo中，5 = （x,y,w,h,id）,而现在的里面 应该是15 = （x,y,w,h,id） + landmark(10). > @ouyanghuiyu,我理解的是这样，有不对的请指正，谢谢！ 谢谢大佬！

@wuzuowuyou 大佬搞明白这个问题了吗，我感觉二楼把我说糊涂了。。。

In Folder 'test_output'， for each image there are 7 outputs, 000 to 004 are side_outputs, 005 maybe the result you want. 005 is fuse output, 006 is average (of 000...

my GPU only has 6G memory which is fine to run the code. dont know your setting

Sorry, I have not had such error. This code will save a checkpoint each 100 steps, maybe 13000 is too big in your dataset( I trained at 30000 steps without...

I run it on Ubuntu 16.04.

@Jasontachiangwu Well, when I cancelled the 'summary_write' operator, this problem was gone. So I guess there maybe some bug in 'summary'. You may rebuild your own code.