py-R-FCN

py-R-FCN copied to clipboard

When the train is unexpected interrupt, what should I do to continue it?

1

ZoneFv

About position-sensitive score maps

3

How the position-sensitive score maps generate? I don not find anything in detail in paper and author just say that use a bank of specialized convolutional layers as the FCN...

qianyeqiang

is it possible to extend the method to train/detection for object with arbitrary number of channel (not only the 3 RGB image standard)?

laupl0082

In the source code of tools/rfcn_alt_opt_5stage_ohem.py, stage 2 rfcn training uses initial net weight from initial imagenet model. Why didn't it use net weight trained from stage 2 rpn but...

xxxxxxxx-dl

nms_wrapper.py uses CPU mode when cfg.USE_GPU_NMS isn't set or it's passed force_cpu as True. When --cpu was provided, demo.py and demo_rfcn.py did not set cfg.USE_GPU_NMS, which defaulted to True. So,...

samlich

Ignore Difficult_Case set to True

1

Hello @YuwenXiong I have a doubt regarding ignoring difficult_case while training/testing on Pascal0712. As per my understanding, for both training and testing phase, difficult annotation from pascal data is ignored...

ananddb90

peformance on smaller network

3

Hi, I try to train on a smaller network such as CaffeNet and GoogleNet. But the result is map= 0.002, something like that. Do you ever get a chance to...

newwhitecheng

Testing in my model failed.

1

When i test my model, Check failed: top_shape[j] == bottom[i]->shape(j) (38 vs. 37) All inputs must have the same shape, except at concat_axis. *** Check failure stack trace: *** How...

ashichao

can't reproduce the mAP reported in readme, only 28.0 mAP in coco minival

1

hi: I have run the traning code without change anyting for three times, but get lower mAP than you report in the readme, here's the details: VOC: 07+12 trainval ,07test,...

zengarden

IOError: [Errno 2] No such file or directory

2

4.Since py-faster-rcnn does not support multiple training datasets, we need to merge VOC 2007 data and VOC 2012 data manually. Just make a new directory named VOC0712, put all subfolders...

cervantes-loves-ai