+130Mb

Not allowed.

Large files (models, >1Mb) for DNN should be stored on external resources (Google Drive, DropBox, OneDrive, etc).

Test images should be kept small (no need to store FullHD images in PNG format - 320x240 may be enough for regression test)

+130Mb

Not allowed.

Large files (models, >1Mb) for DNN should be stored on external resources (Google Drive, DropBox, OneDrive, etc).

Test images should be kept small (no need to store FullHD images in PNG format - 320x240 may be enough for regression test)

@alalek I have already uploaded the model weights to Google Drive:

1. crnn.onnx: https://drive.google.com/file/d/1ooaLR-rkTl8jdpGy1DoQs0-X0lQsB6Fj/view?usp=sharing
2. DB_IC15_resnet50.onnx: https://drive.google.com/file/d/1cmAs91yFRyC2LC3nTTBNNYooyEmb0i5t/view?usp=sharing
3. DB_TD500_resnet50.onnx: https://drive.google.com/file/d/1_9KB6QkIghp2VNZti_dJ3Araz-nDZBsL/view?usp=sharing

Then, I should resize the test images, and only put the images to the new branch? How about the model weights?
Thanks in advance.

Put links on models weights into sample/documentation. So Users can download these files by using provided links.
In samples assume that these external files are located in the current folder.

For tests, see https://github.com/opencv/opencv_extra/blob/master/testdata/dnn/download_models.py (including this file history and related PRs). Use cvtest::findDataFile() with required=False (share on CI is updated manually, so files may missing)

Please squash commits to remove huge remove files from Git history.

Please squash commits to remove huge remove files from Git history.

@alalek
sorry, this is my first time to pull a request, and I am not sure whether I successfully squash my commits.
If I make a mistake again, can I just close this pr, and open a new one?

just close this

Do NOT do that. Just push your changes with "--force" option.

git reset --hard b9d7d2b83efe552b3eb43ea9311e1e53f01456dd
git push --force ...

@alalek Thank you for your help.