search

daijifeng001 / R-FCN

R-FCN: Object Detection via Region-based Fully Convolutional Networks
Other
1.24k stars 542 forks source link

Is there any plan to realize these CPU parts? #44

Open zzningxp opened 7 years ago

zzningxp commented 7 years ago

SoftmaxWithLossLayer::Forward_cpu in src/caffe/layers/softmax_loss_layer.cpp is NOT_IMPLEMENTED, while ./src/caffe/layers/box_annotator_ohem_layer.cpp ./src/caffe/layers/psroi_pooling_layer.cpp ./src/caffe/layers/smooth_L1_loss_layer.cpp are the same.

Is there any plan to realize these CPU part?

zzningxp commented 7 years ago

Furthermore, I think that to test an image by using the pre-trained model on the CPU machine is really in demand. Also, I think that there is only one CPU forward (psroi_pooling_layer.cpp) should be implemented to fulfill the CPU test.

xiaoxiongli commented 7 years ago

I think you'd better to buy a GPU. such as M40. it is really fast if you use a GPU.

zzningxp commented 7 years ago

I have GPU machine, and I need to test the forward pregress on the a CPU machine.

chenxistephen commented 7 years ago

I have the same issue to test the CPU forward time (even I have a GPU machine).

Uchanka commented 7 years ago

I've implemented the cpu compatible version of PSROIPoolingLayer from this original psroi_pooling_layer.cu, although I'm not pretty sure if it's buggy or not. If you still want one of these, I'll push this to my github ASAP. @zzningxp

buaadf commented 7 years ago

@Grillnov I need the file now, have you push it to your gitHub, i didn't find it in your github. thank you

liyuanyaun commented 7 years ago

@Grillnov could you share the file? thank you very much