pytorch-yolo-v3 issues

Darknet issue trying to run slimyolov3

8

I dowload this weights and cfg: https://github.com/PengyiZhang/SlimYOLOv3 original tensor size: torch.Size([1, 3, 608, 608]) I got this error: with map1 and map2 with sizes in darknet.py file: torch.Size([1, 12, 2,...

cartovarc

use resnet-101 as backbone for yolo

1

Hi. I want to implement yolov3 with resnet-101 as its backbone. I find resnet-101.cfg file on dartnet official website and place it as backbone on yolov3-spp.cfg instead of darknet53. But...

thancaocuong

Wrong withdifferent width and height

1

tomjerrygithub

Picture saving bug fixed.

1

Picture output fails in Windows system, since the path is not correctly obtained. `det_names = pd.Series(imlist).apply(lambda x: "{}/det_{}".format(args.det, x.split("/")[-1]))` This trouble result from that Windows use '\' as seperator, which...

yang-jin-hai

adding try exception

I tried to use it and I had some error without try exception handling, so I pull request this matter.

yujiny97

remove reference to num_classes in get_test_input

1

This might be the most trivial pull request ever as it not impacting the execution of the code. However, when deconstructing the code to understand it this line caused a...

rlshuhart

The output of the model has no gradients

Hi! Thank you for building up this pytorch project. I am modifying the project to my own needs, which requires the output of the model has gradients. However, the darknet...

wzjin2017

Extracting the name of the classes detected.

1

I have 3 class(person,sofa and cats). I would like to print output when cats are being detected and no output when person and sofa being detected. Been going through the...

iqrammm