easy-faster-rcnn.pytorch

easy-faster-rcnn.pytorch copied to clipboard

Both training and also inference needs [CUDA, the Nvidia framework for their GPU hardware](https://developer.nvidia.com/cuda-toolkit). This could be mentioned in the README documentation upfront.

xmacex

support problem

1

i run code 'python support/setup.py develop' and error. how i can solve it? 1 error detected in the compilation of "C:/Users/jizef/AppData/Local/Temp/tmpxft_0000a6d8_00000000-10_nms.cpp1.ii". error: command 'D:\\solfware\\CUDA-install\\CUDA Samples\\v8.0\\bin\\nvcc.exe' failed with exit status 1

jizefeng0810

add type-cast to fix the error as facebookresearch repo has done with this [issue](https://github.com/facebookresearch/maskrcnn-benchmark/pull/409)

linchenq

if i write python support/setup.py develop in ubunto app for window10 i have nvrtcGetProgramLogSize error how can i solve it?

sungbinson

Problems with infer

5

Thanks for your good implementation. I have run your code on my own dataset. And I want to know, do your code provide multiple images inferring?

Codermay

Hi, I noticed that the smooth L1 regression loss for the fast r cnn goes to 0 when there are no proposals that have an iou of > 0.5. Is...

peiherng97

About the NMS

5

Thanks for your clean and useful code implementation.I would like to consult you why 'class_bboxes' on the NMS are not arranged in descending order of 'class_probs' in 'model.py' when validating...

DrugRui

Other models from torchvision

2

Hi, I wanted to use light-weight base models from torchvision like mobilenet, squeezenet and densenet. Any tips on how to edit the code?

madiltalay

NMS without cuda

1

Is there any way to use NMS for inference in a system that has no GPU?

MurugeshMarvel