SegNet model build fail

[fatrybl]

Dear author, Thank you very much for your research. Ive read it with great interest! I tried to implement classification with Bayessian SegNet model, but faced this problem while launching SegNet-CamVid11-Bayesian.ipynb file. The packages are pretty fresh and the only difference from your requirements is Pillow package. My Pillow version is 7.0 But it seems like it isnt gonna influence on that

I also tried to change parameter name to dropout_location = dropout_location. It didnt help.

[janisgp]

Hi @fatrybl

Thanks for pointing this problem out. The parameter dropout_loc was a development artifact. I removed it in the latest commit. I further added a yaml file with the newer conda environment in which I just ran the notebook. Feel free to create a conda environment from it.

Best, Jansi

[janisgp]

Hi @fatrybl

does this solve your problem?

Best, Janis

[fatrybl]

Thanks for fast solution and reply! yeah, that problem has been solved, but another occurred.

[janisgp]

Hey,

is the a folder called "y3595640ea0b061d5422951ed3e1a7bf0" (or equivalently y"some other hash") in the folder ".../Sampling-free-Epistemic-Uncertainty/Bayesian_SegNet/src/datasets/camvid" ? And if so, does it contain a file metadata.csv?

Best, Janis

[janisgp]

Hey @fatrybl

did you resolve the problem?

Best, Janis

[janisgp]

@fatrybl

can the issue be closed?

Best, Janis

[fatrybl]

Frankly speaking, I did not chech yet, so you might close it, if you are sure that evrrything wotkd well 11.04.2020, 11:22, "janisgp" notifications@github.com: @fatrybl can the issue be closed? Best, Janis

—You are receiving this because you were mentioned.Reply to this email directly, view it on GitHub, or unsubscribe.