Shuai Kyle Zheng comments

Results 51 comments of


                                            Shuai Kyle Zheng

You initialised your network on CPU, please initialize it on GPU.

I think you should check if you run the train / test on GPU by specifying '-GPU #gpu_id'. #gpu_id is your GPU device id.

eltwise_layer.cpp:34] Check failed: bottom[0]->shape() == bottom[i]->shape() bottom[0]: 1 21 333 500 (3496500), bottom[1]: 1 21 375 500 (3937500)

You need to pad the image as we did in the demo inference code. Otherwise, you would see this error.

worse and worse performance during finetune

Hi, Have you checked your data? CRFasRNN requires input image aligning with ground truth mask. Otherwise, the performance will undoubtedly get worse. From what I understand what you said, it...

Linking Problems During Building

Thanks for reporting this error. From your error message, It seems to be the error of opencv 3.0. Could you provide more detail information, e.g. ubuntu linux? gcc version? opencv...

Linking Problems During Building

It seems to be something to do with the image encoding libraries. Any error messages have you got?

Installing on Mac

Hi, Could you use the previous version of gcc? e.g. 4.6.x. I am not sure if this is the problem of gcc. However, the code has been tested on ubuntu...

Installing on Mac

@phil0stine It should work with Mac, what's the error message have you got? Which mac version and which gcc or clang version have you tried?

crf-rn demo vs trained algorithm

We used the same model running on the background for the demo. This could be caused by the scaling issue. If you have an input image with significant large size,...

crf-rn demo vs trained algorithm

If it is a gray input image, the demo script will convert them into 3-channels like image, each channel will be filled with the gray image. For an image with...

Reconcile caffe with upstream caffe

@mtourne Thanks for this. Did you test python script? It seems to be broken.