YOLOP

YOLOP copied to clipboard

In the \_\_getitem__() function in lib/dataset/AutoDriveDataset.py, we read an image and then convert from BGR to RGB. This image is then passed to augment_hsv() in lib/utils/augmentations.py file. Here, we convert...

RahulRewale

fixed plot dimension error

1

The demo.py file has an error plotting the results, which is fixed here.

arbit3rr

How do I need to change the code to train only lane detection?

1

Besides changing [ _C. train.lane_only = False ] to True, what else needs to be changed?

jasfa

单独训练车道线检测的话，需要如何修改代码呢

1

除了将_C.TRAIN.LANE_ONLY = False # Only train ll_segmentation task改为True还有哪里需要修改呢

jasfa

How to view image result in opencv

1

Hi!! I'm using Yolop with my images but I want to visualisate the result. How to could do it? My code is: `import torch import torchvision.transforms as transforms import cv2...

bb6698

```Dockerfile ARG PYTORCH="1.8.1" ARG CUDA="10.2" ARG CUDNN="7" FROM pytorch/pytorch:${PYTORCH}-cuda${CUDA}-cudnn${CUDNN}-devel ENV TORCH_CUDA_ARCH_LIST="6.0 6.1 7.0+PTX" ENV TORCH_NVCC_FLAGS="-Xfatbin -compress-all" ENV CMAKE_PREFIX_PATH="$(dirname $(which conda))/../" # To fix GPG key error when running apt-get update...

ozora-ogino

File "/home/dd/YOLOP/lib/core/function.py", line 301, in validate label_det_pred = f'{names[int(cls)]} {conf:.2f}' KeyError: 2

1

xun lian duo zhong mubiao jiance shi chu zhege cuowu,qingwen zengyang gaizheng ?

wanshuchun

Does anyone have a working source code of this yolop model for end to end training. There are so many bug I encountered with the code and I cannot fix...

HunterShinobiTitan

label_det_gt = f'{names[int(cls)]}' KeyError: 11

3

When I run train.py: Epoch: [1][3480/3500] Time 3268.846s (1639.564s) Speed 0.0 samples/s Data 3268.648s (1639.355s) Loss 0.32530 (0.28520) 0%| | 0/500 [00:00

HunterShinobiTitan

It is mentioned in the paper that Lcls and Lobj in the detection task use focal loss, but this is not the case in the code.

1

in the code loss.py, line 195: `gamma = cfg.LOSS.FL_GAMMA if gamma > 0: BCEcls, BCEobj = FocalLoss(BCEcls, gamma), FocalLoss(BCEobj, gamma)` but in the code default.py, line 39: `_C.LOSS.FL_GAMMA = 0.0`...

lss0510