Skip to content

Conversation

sashulyak
Copy link

  1. Fresh version of unpooling added. Now it can deal with variable batch size. (took it from here https://github.com/tensorflow/tensorflow/issues/2169)
  2. 0-dimension of placeholders were set to None. This allows setting a different batch size for train and test runs.
  3. Trimap argument added to the test script, so now you can test pretrained model on your own images, even if you don't have alpha-mask.
  4. Redme edited.

@ImMrMa
Copy link

ImMrMa commented Mar 22, 2018

Thank you very much for your help with this image size, but I see that after you change the code, the input image will still be resized to that 320 pixel , so how can you make it has the size of according to the input size?

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

2 participants