issues
search
funkelab
/
gunpowder
A library to facilitate machine learning on multi-dimensional images.
https://funkelab.github.io/gunpowder/
MIT License
78
stars
56
forks
source link
issues
Newest
Newest
Most commented
Recently updated
Oldest
Least commented
Least recently updated
reset dtype in normalize request
#115
abred
closed
4 years ago
0
graph: change directionality warning to debug
#114
abred
closed
4 years ago
0
train loop log: add current lr and set loss precision
#113
abred
closed
4 years ago
3
elastic aug: make debug log work in 2d and 3d
#112
abred
closed
4 years ago
0
print node if dtype consist. check fails
#111
abred
closed
4 years ago
0
add support for summaries with different write frequencies
#110
abred
closed
4 years ago
0
Simple augment transpose bugfix
#109
pattonw
closed
4 years ago
0
add optional automatic mixed precision and tf.data.Dataset
#108
abred
closed
4 years ago
2
Add NoiseAugment, ZarrWrite and DaisyRequestBlocks to API reference
#107
bentaculum
closed
4 years ago
0
Add hidden nodes to the API reference
#106
bentaculum
closed
4 years ago
2
DaisyRequestBlocks fails to create a daisy context
#105
bentaculum
closed
3 years ago
2
Fix Random Location Graphs
#104
riels89
closed
4 years ago
0
Hdf5LikeSource: Distinguish source offset vs passed array_specs offset
#103
bentaculum
closed
3 years ago
2
Contrib folder / gunpowder node exchange
#102
bentaculum
closed
3 years ago
0
Enable autoskip in crop.
#101
riels89
closed
4 years ago
0
Torch multiprocessing probelms
#100
pattonw
closed
4 years ago
2
Transpose around center in SimpleAugment
#99
riels89
closed
4 years ago
0
Fix SimpleAugment upstream request
#98
riels89
closed
4 years ago
0
Fix ROI issue when using both transpose and mirror
#97
riels89
closed
4 years ago
0
Fix/reject and stack
#96
riels89
closed
4 years ago
0
Add nodes for squeezing and unsqueezing arrays
#95
bentaculum
closed
4 years ago
0
RandomLocation: Preserve offsets for empty shift ROIs
#94
bentaculum
closed
4 years ago
4
Array: Add method for deep-copying
#93
bentaculum
closed
4 years ago
0
Revert "pytorch Train, Predict:"
#92
pattonw
closed
4 years ago
0
Add one more test for RandomLocation
#91
bentaculum
closed
4 years ago
2
Patch 1.1.3
#90
bentaculum
closed
4 years ago
0
Can't run the cremi example.
#89
skt9
closed
4 years ago
1
v1.1 Patch
#88
pattonw
closed
4 years ago
0
Augment tests
#87
pattonw
closed
4 years ago
0
Improved Graph Support
#86
pattonw
closed
4 years ago
0
Implement Placeholders
#85
pattonw
closed
4 years ago
0
Improve graph support
#84
pattonw
closed
4 years ago
0
Ensure centered
#83
pattonw
closed
4 years ago
0
Random seed
#82
pattonw
closed
4 years ago
0
Remove z5py
#81
pattonw
closed
4 years ago
0
v1.0 not on PyPI
#80
hanslovsky
closed
4 years ago
1
Implement graphs
#79
pattonw
closed
4 years ago
0
Node dependencies cleanup
#78
pattonw
closed
4 years ago
0
set random seed at start of each worker process
#77
abred
closed
4 years ago
1
Rewrite specified location node to include jitter, boundary check
#76
cmalinmayor
closed
4 years ago
0
Pytorch predict
#75
pattonw
closed
4 years ago
0
node_dependencies batch_id
#74
pattonw
closed
10 months ago
2
Bugfix upsample request larger upstream roi
#73
cmalinmayor
closed
4 years ago
1
Assert offset mult of voxel size
#72
cmalinmayor
closed
5 years ago
0
Set default klb reader threads to 1
#71
cmalinmayor
closed
5 years ago
0
Add option to set num threads for reading klbs
#70
cmalinmayor
closed
5 years ago
0
Avoid double counting channels in elastic augment reshape
#69
cmalinmayor
closed
5 years ago
0
fix permutation list for transpose augmentation
#68
abred
closed
4 years ago
1
SimpleAugment transposes axes in the ROI request, instead of transposing the axes in the array
#67
bentaculum
closed
4 years ago
1
Fix bug: Duplicating counts in ProfilingStats
#66
pattonw
closed
5 years ago
0
Previous
Next