search

zhou13 / lcnn

LCNN: End-to-End Wireframe Parsing
MIT License
494 stars 94 forks source link

Rectangle training and training your own dataset #51

Open carlsummer opened 3 years ago

carlsummer commented 3 years ago

Thank you very much for your sharing. I modified some of your code according to my own requirements, mainly to meet the following functions After my own modification, I can train my own data set, and I can train rectangle, so I don't have to be square

zhou13 commented 3 years ago

Thanks for the PR! Have you tested your model and notice any performance improvement over image resizing?

carlsummer commented 3 years ago

I tried 512 512 and 64 512, because the graph for my requirements is 400 3000, so it is better to test 64 512. The line appeared at the first verification, but it didn't appear until 512 * 512 for the third time

carlsummer commented 3 years ago

I didn't compare the effects of different sizes on wireframe datasets. But you can try, training should not be too long, I because the company has the need to detect the line, so I changed the code to run rectangular; And I also hope that you can merge or open a new branch to help the readers

carlsummer commented 3 years ago

image image

carlsummer commented 3 years ago

I saw several issues in the issue about how to train my own data set. The task my company now gives me is also to train my own data set, and then it is not a square picture. So I changed the code myself. Share with you, hope to help readers behind

sachinkaundal commented 3 years ago

Hello @carlsummer, i modified train_test_split.py as per my requirements and trying to run but i am getting some error : ModuleNotFoundError: No module named 'python_developer_tools' can you please help me to solve the issue. Thanks in advance for response.

carlsummer commented 3 years ago

@sachinkaundal https://github.com/carlsummer/python_developer_tools That is a development tool I wrote. There are many tool functions in it. You can copy it,

carlsummer commented 3 years ago

@sachinkaundal I annotate it with label

sachinkaundal commented 3 years ago

Thank you so much for your response.

On Tue, 13 Jul 2021 at 13:27, carlsummer @.***> wrote:

@sachinkaundal https://github.com/sachinkaundal I annotate it with label

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/zhou13/lcnn/pull/51#issuecomment-878865423, or unsubscribe https://github.com/notifications/unsubscribe-auth/ASXDPSE6PSTINQPEGHINA63TXPWXHANCNFSM5AABRE6Q .

carlsummer commented 3 years ago

deploy onnx，and you can use onnxruntime run