Segmentation fault (core dumped)

[ChrisHJC]

I try to train a model without pretrained resnet50, and I get this problem. how can I solve this problem？

[DongGuangchang]

Loaded dataset voc_2007_trainval for training Set proposal method: gt Appending horizontally-flipped training examples... voc_2007_trainval gt roidb loaded from /home//faster-rcnn.pytorch-pytorch-1.0/data/cache/voc_2007_trainval_gt_roidb.pkl done Preparing training data... done before filtering, there are 10022 images... after filtering, there are 10022 images... 10022 roidb entries Loading pretrained weights from data/pretrained_model/resnet101_caffe.pth 段错误 (核心已转储) I trained on 1080ti and encountered the same problem. Has your issue been resolved?

[Haienzi]

I have the same problem. Have you solved it？

[Summer-fox9]

I encountered the same problem, want a solution.

[qinzzz]

I have the same problem. Did you find the solution?

[Summer-fox9]

Sorry, I didn't. Finnaly I use the type of torchvision to replicate faster r-cnn.

------------------ 原始邮件 ------------------ 发件人: "qinzzz"<notifications@github.com>; 发送时间: 2020年5月9日(星期六) 下午5:02 收件人: "jwyang/faster-rcnn.pytorch"<faster-rcnn.pytorch@noreply.github.com>; 抄送: "李嘉丽"<495596191@qq.com>;"Comment"<comment@noreply.github.com>; 主题: Re: [jwyang/faster-rcnn.pytorch] Segmentation fault (core dumped) (#749)

I have the same problem. Did you find the solution?

— You are receiving this because you commented. Reply to this email directly, view it on GitHub, or unsubscribe.

[loolzaaa]

Try to write separately from pycocotools import coco in python console. Is there any error? If yes, you can try install pycocotools separately:

pip uninstall -y pycocotools
pip install --no-binary :all: pycocotools

[qinzzz]

I reinstalled using python2.7 and compiled pycocotool from cocoapi. The problem is gone.

409 this might be helpful!