search

sabeenlohawala / tissue_labeling

0 stars 1 forks source link

Simple_Unet is missing MaxPool layers #42

Closed sabeenlohawala closed 7 months ago

sabeenlohawala commented 7 months ago

https://github.com/sabeenlohawala/tissue_labeling/blob/18de49f5b831edc428b3c42fc32dab8504376b69/TissueLabeling/models/simple_unet.py#L19

The downsampling blocks have an extra convolutional layer here. Remove this and restructure the code to have MaxPool after the encoder blocks, like in the paper.

hvgazula commented 7 months ago

not resolved in simple_unet.py but resolved in nobrainer_unet.py (which is the same as https://github.com/MGH-LEMoN/ddpm-labels/blob/main/ddpm_labels/models/model3.py)