Testing result issue

[candy0530]

I tried the project you provided, but I got some errors. I used model trainedmodel/fphb_crack.caffemodel which downloaded from Google Drive, dataset CRACK500 test.txt also downloaded from Google Drive and test.py with test_fphb_crack.prototxt, but I got error messages. F0613 18:57:03.865557 146139 net.cpp:861] Cannot copy param 0 weights from layer 'conv4_fuse'; shape mismatch. Source param shape is 512 1024 1 1 (524288); target param shape is 512 2048 3 3 (9437184). To learn this layer's parameters from scratch rather than copying from a saved net, rename the layer.

I replaced layers name conv1_fuse, conv2_fuse, conv3_fuse, conv4_fuse, conv5_fuse, score-dsn1, score-dsn2, score-dsn3, score-dsn4, score-dsn5 to conv1_fuse_1, conv2_fuse_1, conv3_fuse_1, conv4_fuse_1, conv5_fuse_1, score-dsn1_1, score-dsn2_1, score-dsn3_1, score-dsn4_1, score-dsn5_1. It works, but I got all black result images.

Did I do wrong steps or miss some steps? How can I get normal results? Thanks!

[fyangneil]

@candy0530 i will double check the model and prototxt file this weekend

[fyangneil]

@candy0530 i have tested it and it works. Have you tried the FPHB-tutorial in example folder? please try it first, let me know if it does not work.

[candy0530]

@fyangneil I re-downloaded model trainedmodel/fphb_crack.caffemodel again, finally it works with test.py and with test_fphb_crack.prototxt. But I got another issue.

I trained a model by solve_fphb_crack.py with train_fphb_crack.prototxt and get a model named fphb_iiter_20000.caffemodel. When I used this model with test_fphb_crack.prototxt, I got the same error messages. Aren't test_fphb_crack.prototxt and train_fphb_crack.prototxt same structures? How can I alter train_fphb_crack.prototxt to a correct test_fphb_crack.prototxt?

Thanks a lot!

[fyangneil]

yes，the training and test files are a little different, I will modify it tomorrow.

On Tue, Jun 18, 2019 at 1:43 AM candy0530 notifications@github.com wrote:

@fyangneil https://github.com/fyangneil I re-downloaded model trainedmodel/fphb_crack.caffemodel again, finally it works with test.py and with test_fphb_crack.prototxt. But I got another issue.

I trained a model by solve_fphb_crack.py with train_fphb_crack.prototxt and get a model named fphb_iiter_20000.caffemodel. When I used this model with test_fphb_crack.prototxt, I got the same error messages. Aren't test_fphb_crack.prototxt and train_fphb_crack.prototxt same structures? How can I alter train_fphb_crack.prototxt to a correct test_fphb_crack.prototxt?

Thanks a lot!

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/fyangneil/pavement-crack-detection/issues/7?email_source=notifications&email_token=ALAUWTWR5JJDXVZBIZ3NPFDP3CNZRA5CNFSM4HYDF4Z2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGODX5UXUA#issuecomment-503008208, or mute the thread https://github.com/notifications/unsubscribe-auth/ALAUWTTT4NFM7D4AMZVAFDTP3CNZRANCNFSM4HYDF4ZQ .

[fyangneil]

@candy0530 you could try the test_fphb_crack_v1 file. if it does not work, let me know

[candy0530]

@fyangneil It does not work. But I found typo in solver_fphb_crack.prototxt.

net: "train_fpn_crack.prototxt"

Due to this line, I used FPN model structure to train the FPHB model, so it works with test_fpn_crack.prototxt. So it should be "net: train_fphb_crack.prototxt" at first line in solver_fphb_crack.prototxt. Do I misunderstand the code?

[fyangneil]

yes，you are right. does it work？

On Wed, Jun 19, 2019 at 1:33 AM candy0530 notifications@github.com wrote:

@fyangneil https://github.com/fyangneil It does not work. But I found typo in solver_fphb_crack.prototxt.

net: "train_fpn_crack.prototxt"

Due to this line, I used FPN model structure to train the FPHB model, so it works with test_fpn_crack.prototxt. So it should be "net: train_fphb_crack.prototxt" at first line in solver_fphb_crack.prototxt. Do I misunderstand the code?

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/fyangneil/pavement-crack-detection/issues/7?email_source=notifications&email_token=ALAUWTTU5FOK3BYVBNO3UGLP3HVLZA5CNFSM4HYDF4Z2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGODYBD2TA#issuecomment-503463244, or mute the thread https://github.com/notifications/unsubscribe-auth/ALAUWTRPTYRUQ2U4W5GT6ALP3HVLZANCNFSM4HYDF4ZQ .

[candy0530]

It works! Thanks a lot!