yolov5_obb

yolov5_obb copied to clipboard

OpenCV DNN C++ inference

3

The prediction of an Image using OpenCV DNN with python detect.py works very well. SInce I have little to no experience with C++ and I intent to use it in...

blackHorz

How can I improve the orientation regression performance on a custom dataset? I get high training and validation scores (600 epochs) **HBBmAP.5** > 90% **precision&recall** ~95% **val/theta_loss** ~0.18 But sometimes...

guglie

请问大家在dota上跑到作者跑到的精度了吗？为什么我跑了很多次都到不了73.3的精度。除了batchsize设置的是8其他参数看着也没啥不一样的啊？求各位老哥指教指教下

1

用了作者给的超参数和yolov5官方的预训练权重，yolov5n最多只能到72.2。想问下大家跑到73.3的精度了吗？看了作者的结果文件，hpy参数除了batchsize我的是8其他的没发现不一样的，求大佬指教下。到底是哪里有问题。对了我没有test直接看的val结果最多72.2。是不是要去把结果融合下？还是其他的？ 下载玩作者的这个代码只把数据及位置改了，还有预训练权重位置。batchsize的原因吗？？？

tom123478

模型无法收敛识别效果差

1

您好，我用这个代码训练自己的数据集，总样本数量为116张，100张训练，16张测试。但是模型一直震荡无法收敛，识别效果也不好，会有多个框出现。能否指点一下？下面是我的数据集，标签以及测试结果。    

RemeberMeX

Help Wanted 🙏 to add Oriented Bounding Boxes (OBB) to YOLOv8

3

Hello @hukaixuan19970627! I hope you're doing well! I'm the author of [YOLOv8](https://github.com/ultralytics/ultralytics) and founder of Ultralytics. I've been following your work on the YOLOv5 OBB implementation and am genuinely impressed...

glenn-jocher

demo已经跑通了，但是图像不需要分割，想问一下在哪个文件哪里删merge部分

9

只在NMS那里找了merge，不清楚是不是需要删这里。而且在第一步载入图像时，显示train: WARNING: data1/train/images/img_99.jpg: ignoring corrupt image/label: list index out of range，是因为文件名不对吗，还是因为没有切割图像呢。 

tianmingguxiang

CSL

2

博主你好，请问怎么关掉CSL呢，想做一个用CSL和不用的对比

XIAO0WANG999

RuntimeError: CUDA out of memory

1

Is anyone encoutering this problem here ? it was working fine yesterday, but i tried to use --cache today and it did not work and i tried to run without...

liamh1999

Windows10进行python setup.py develop 报错

7

error: command 'cl.exe' failed: None

aniseanise

遇到了RuntimeError: Not compiled with GPU support

4

Traceback (most recent call last): File "E:/WYH/yolov5_obb-master/train.py", line 635, in main(opt) File "E:/WYH/yolov5_obb-master/train.py", line 530, in main train(opt.hyp, opt, device, callbacks) File "E:/WYH/yolov5_obb-master/train.py", line 373, in train compute_loss=compute_loss) File "D:\Anaconda\envs\dx\lib\site-packages\torch\autograd\grad_mode.py",...

qqdalao