Shikai Li comments

Results 11 comments of


                                            Shikai Li

I found that the image and label are not converted to be the type of tensor???

Yes. I think it may works in the old version like 0.40.

Train my own datasets

Same issue.

undefined symbol: _ZN10tensorflow16CorrelationGrad

@Iamanorange Hello, I still got some problems following the steps above. Here are some logs after compiling and running the test.py. ``` `python -m src.flownet2.test --input_a data/samples/0img0.ppm --input_b data/samples/0img1.ppm --out...

how to get DSB2018_application.py

Same issue.

Using pretrained models with data augmentation.

Same issue. Finally the accuracy of validation is only near 90%.

Using pretrained models with data augmentation.

Oh here is the solution. [#75](https://github.com/fchollet/deep-learning-with-python-notebooks/issues/75#issuecomment-442815465)

Using pretrained models with data augmentation.

@morenoh149 Sorry, I didn't check the issue number.

Using pretrained models with data augmentation.

@zgrkpnr I think it depends on the version of Keras. The code in that section is a bit old for newest version of Keras. Did you change `rescale=1./255` to `preprocessing_function=preprocess_input`...

I think the code is wrong

@nnUyi Well, have you trained it on other datasets?

Tool scraping data from model.com

same issue