We read every piece of feedback, and take your input very seriously.
To see all available qualifiers, see our documentation.
Have a question about this project? # for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “#”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? # to your account
数据校验在平台上通不过,但是本地PaddleX可以通过
我的数据集如下: PS C:\Users\DMM\Desktop\tmp_data> tree /f 卷 Windows-SSD 的文件夹 PATH 列表 卷序列号为 D846-385E C:. ├─annotations │ instance_train.json │ instance_val.json │ └─images
但是PaddleX中校验通过 PS C:\Users\DMM\PycharmProjects\PaddleX> python.exe .\main.py -c paddlex/configs/instance_segmentation/Mask-RT-DETR-L.yaml -o Global.mode=check_dataset -o Global.dataset_dir=C:/Users/DMM/Desktop/tmp_data C:/Users/DMM/Desktop/tmp_data loading annotations into memory... Done (t=0.00s) creating index... index created! loading annotations into memory... Done (t=0.00s) creating index... index created! Only 5 images in instance_val.json loading annotations into memory... Done (t=0.00s) creating index... index created! loading annotations into memory... Done (t=0.02s) creating index... index created! Check dataset passed ! 3. 请提供您的星河uid和产线id Pipeline ID:p-6fdd7d78dc52
The text was updated successfully, but these errors were encountered:
数据校验失败。目标检测文档说使用coco数据集格式,Label Studio导出的coco格式YOLO正常使用,一比对paddlex官方给出的示例,不一样。
Sorry, something went wrong.
平台上这里确实存在问题(当数据量少于10张图片时),预计本周修复上线。
TingquanGao
No branches or pull requests
Checklist:
描述问题
数据校验在平台上通不过,但是本地PaddleX可以通过
复现
我的数据集如下:
PS C:\Users\DMM\Desktop\tmp_data> tree /f
卷 Windows-SSD 的文件夹 PATH 列表
卷序列号为 D846-385E
C:.
├─annotations
│ instance_train.json
│ instance_val.json
│
└─images
AI Studio上数据集校验失败,log为
dataset.log
但是PaddleX中校验通过
PS C:\Users\DMM\PycharmProjects\PaddleX> python.exe .\main.py -c paddlex/configs/instance_segmentation/Mask-RT-DETR-L.yaml -o Global.mode=check_dataset -o Global.dataset_dir=C:/Users/DMM/Desktop/tmp_data
C:/Users/DMM/Desktop/tmp_data
loading annotations into memory...
Done (t=0.00s)
creating index...
index created!
loading annotations into memory...
Done (t=0.00s)
creating index...
index created!
Only 5 images in instance_val.json
loading annotations into memory...
Done (t=0.00s)
creating index...
index created!
loading annotations into memory...
Done (t=0.02s)
creating index...
index created!
Check dataset passed !
3. 请提供您的星河uid和产线id
Pipeline ID:p-6fdd7d78dc52
The text was updated successfully, but these errors were encountered: