Skip to content

Bug in seg_output #102

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
FabianIsensee wants to merge 23 commits into
base: master
Choose a base branch
from
Open

Bug in seg_output #102

Changes from all commits
Commits
Show all changes
23 commits
Select commit Hold shift + click to select a range
eb3f145
added UNet + example (road segmentation)
FabianIsensee Aug 24, 2016
4cd59a4
Reduced number of plotted test images, Added a plot of the road proba…
FabianIsensee Aug 24, 2016
ea84aef
renamed some layers of UNet to make their names more clear, added con…
FabianIsensee Aug 26, 2016
fcff2da
code restructuring (now it's not cramped into a single .py file anymo…
FabianIsensee Aug 26, 2016
9cbb33f
now using urlretrieve for downloading the data. Also uses multiproces…
FabianIsensee Aug 26, 2016
f10200d
updated batch_iterator so that it now supports deterministic batch ge…
FabianIsensee Aug 26, 2016
db99d82
added center crop generator
FabianIsensee Aug 26, 2016
90aa43d
now uses compressed numpy arrays to store the dataset
FabianIsensee Aug 26, 2016
c657d51
some minor fixes
FabianIsensee Aug 26, 2016
49df413
some more minor fixes. It now runs out of the box
FabianIsensee Aug 26, 2016
f9c7fca
switched back to uncompressed numpy arrays for saving the dataset (it…
FabianIsensee Aug 29, 2016
b89f4d6
minor bugfixes, new pretrained weights
FabianIsensee Aug 29, 2016
654c2d7
minor bug fixes and cleanup
FabianIsensee Aug 29, 2016
4daee1e
minor bug fix
FabianIsensee Aug 29, 2016
13938aa
forgot to comment loading of pretrained params
FabianIsensee Aug 29, 2016
1c83015
renamed layers for clarity
FabianIsensee Aug 29, 2016
5412a68
python code for downloading and extracting pretrained weights, added …
FabianIsensee Jan 12, 2017
d88900b
Unet now uses relu and He's initialization scheme as in the original …
FabianIsensee Jan 12, 2017
570b18e
new pretrained weights
FabianIsensee Jan 16, 2017
d254617
This commit adds the UNet segmentation network to lasagne's recipes. …
FabianIsensee Aug 24, 2016
b8ee2d1
Merge branch 'master' of https://github.com/FabianIsensee/Recipes
FabianIsensee Jan 16, 2017
4b6cec2
fixed a bug with the seg_output. There wehe some kwargs missing leadi…
FabianIsensee Mar 17, 2017
cf9da6c
fixed a bug in seg_output
FabianIsensee Mar 17, 2017
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
3 changes: 2 additions & 1 deletion examples/UNet/massachusetts_road_segm.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -124,7 +124,8 @@ def main():
updates = lasagne.updates.adam(loss, params, learning_rate=learning_rate)

# create a convenience function to get the segmentation
seg_output = lasagne.layers.get_output(net["output_segmentation"], x_sym, deterministic=True)

seg_output = lasagne.layers.get_output(net["output_segmentation"], x_sym, deterministic=True, batch_norm_update_averages=False, batch_norm_use_averages=False)
seg_output = seg_output.argmax(1)

train_fn = theano.function([x_sym, seg_sym, w_sym], [loss, acc_train], updates=updates)
Expand Down