Top Detectron issues

Model is being successfully trained on custom dataset, but when testing it outputs COCO's class names instead of custom class names

netw0rkf10w picture netw0rkf10w·4 Jun 2018·4 Comments

Why use sigmoid cross entropy instead of softmax in RPN?

twmht picture twmht·2 May 2018·3 Comments

convert_cityscapes_to_coco.py error:ValueError: too many values to unpack

wuqi930907 picture wuqi930907·12 Apr 2018·11 Comments

[enforce fail at conv_pool_op_base.h:253] training my own keypoint dataset

geyuying picture geyuying·6 Apr 2018·8 Comments

Inference with Pretrained Models OSError: /usr/local/lib/libcaffe2_detectron_ops_gpu.so: undefined symbol: _ZN6caffe211OperatorDef17default_instance_E

IgorKasianenko picture IgorKasianenko·5 Apr 2018·6 Comments

KeyError in json_dataset.py when training model on own dataset

gabriellapizzuto picture gabriellapizzuto·29 Mar 2018·2 Comments

Keypoints training using 1GPU error

7oud picture 7oud·10 Mar 2018·8 Comments

CUDA_ARCH_NAME error with conda version of caffe2

dhpollack picture dhpollack·8 Mar 2018·10 Comments

Output from training

fangpengcheng95 picture fangpengcheng95·8 Mar 2018·4 Comments

Anchor ratio generation

hgaiser picture hgaiser·4 Mar 2018·5 Comments
 Previous issues
Next issues