gengyanlei / reflective-clothes-detect-yolov5

reflective-clothes-detect-dataset、helemet detection yolov5、工作服(反光衣)检测数据集、安全帽检测、施工人员穿戴检测

author is leilei 👋

• 🎓 CSU(master's degree)
• 📖 leilei's Bio
• 📖 leilei's blog
• 🔭 Computer vision(目标检测-语义分割-分类)
• ❤ Program language: Python C++ (Go Rust)-toLearn

reflective-clothes-detect-yolov5's Issues

Hello! Can you provide a link to your dataset?

Hello! Can you provide a link to your dataset?

Cant download dataset and weight files?

Probably because of Biadu's permission issue. Kindly provide OneDrive or Google Drive link.
Thank you

yolov5版本为v3

标注有问题

开源的这一千多张数据标注有问题。例如33650146-4792-4b8e-9adf-44eeade0db86.jpg图片对应的xml只标注了反光衣，没有标注头部，很多这样的图片xml，，，存在的问题也就是没有戴安全帽的头部有的标注有的没有标注。

ValueError: invalid literal for int() with base 10: ''

My OS: Ubuntu
Python: 3.7
Issue:
I came across "ValueError: invalid literal for int() with base 10: '' when I ran "train.py".
Above error happened in folder "utils" > general.py around line 956 -ish...
Solution:

update_NUM = 0
for x in d:
    if '_' in x:
        tmp_NUM = int(x[len(dir):])
        if tmp_NUM > update_NUM:
            update_NUM = tmp_NUM
n = update_NUM + 1
# n = max([int(x[len(dir):x.find('_') if '_' in x else None]) for x in d]) + 1  # increment

The actual error is at "int(x[len(dir):x.find('') if '' in x else None])", I solved it via changing "in-line" code to several lines.
Please correct me if I got any inappropriate

请问有在数据集上的测试指标吗

您好，完整的标注数据可以共享吗？

您好！您共享的数据貌似是只标了反光衣和其他衣物，但是根据您给出的demo那张图，上面是有标整体人的，我想问一下标有整体人和衣服这一部分数据可以共享吗？

请问，没穿反光衣是否需要标注？

没穿反光衣是否需要标注？

where is the dataset?

Hello! Can you provide a link to your dataset?

程序运行环境？

请问您的代码具体运行环境？Python版本、Pytorch版本这样信息方便告知吗？非常感谢！

反光衣的训练集格式是coco128格式的吗

版本问题

pytorch版本必须1.6.0以上么，我用1.6.0以下版本，torch.load(w, map_location=map_location)这个加载权重文件报错，_pickle.UnpicklingError: A load persistent id instruction was encountered

dataset train problem

您所说的训练方式，是首先采用反光衣数据集 ，分为两类(反光衣和别的衣服)得到的模型。随后将训练SHWD安全帽数据集，以得到的模型作为网络的预训练模型，分为两类(帽子和头)进行训练得到最后的模型，是这么理解么，请批评指正