ELEKTRONN elektronn3 issues

ELEKTRONN / elektronn3

A PyTorch-based library for working with 3D and 2D convolutional neural networks, with focus on semantic segmentation of volumetric biomedical image data

MIT License

161 stars 27 forks source link

issues

Newest

Newest Most commented Recently updated Oldest Least commented Least recently updated

Input normalization

#54 SebastienTs opened 1 year ago
0
Elektronn3

#53 SebDBI closed 1 year ago
0
UNet Implementation details

#52 SebastienTs opened 1 year ago
2
Sparse annotations for training

#51 SebastienTs opened 1 year ago
2
How to train the model with custom datasets

#50 Elmiar0642 opened 1 year ago
1
Error with 3D convolution on custom Kernels

#49 Elmiar0642 closed 1 year ago
2
Trainer scripts and metrics for GNN training

#47 hash-ir closed 2 years ago
0
Merge phil_convpoint into syconn

#46 pschubert closed 3 years ago
0
Knossos labels kzip

#45 riegerfr closed 3 years ago
0
Update requirements.txt

#44 flmuk closed 3 years ago
0
noise to zero_like, not empty_like

#43 riegerfr closed 4 years ago
0
introduced handling for target = None in RandomRotate2d

#42 riegerfr closed 4 years ago
1
implemented caching for KnossosRawData

#41 riegerfr closed 4 years ago
1
Various improvements by @Optiligence and @mdraw

#40 mdraw closed 5 years ago
0
Import error when running in SyConn environment

#39 AldoCP closed 5 years ago
1
Caconi ensemble

#38 caconi closed 3 years ago
0
Missing assignment to mode in train_unet_neurodata.py

#37 dmankins closed 5 years ago
1
train_unet_neurodata.py default seed is a seed of 0, not a random seed

#36 dmankins closed 4 years ago
2
Offline validation/evaluation of trained models

#35 mdraw closed 4 years ago
1
Inference example script

#34 mdraw closed 5 years ago
0
High-level documentation

#33 mdraw opened 5 years ago
0
NaN losses when training with torch.nn.CrossEntropyLoss

#32 mdraw closed 5 years ago
0
correct metrics with ignore class

#31 Optiligence closed 5 years ago
1
don’t allocate complete in- and output on the gpu

#30 Optiligence closed 5 years ago
0
adjust output shape as multiple of tile shape

#29 Optiligence closed 5 years ago
0
Reproducibility

#28 mdraw closed 5 years ago
0
fix metrics for i.e. negative gt batches

#27 Optiligence closed 5 years ago
0
ElasticTransform and RandomGaussianBlur transforms

#26 mdraw closed 5 years ago
1
Add Predictor, TripletNet, RandomFlip, 2D FCN

#25 pschubert closed 6 years ago
2
Enable automatic documentation build and hosting on readthedocs

#24 mdraw closed 6 years ago
0
New transforms API for datasets

#23 mdraw closed 6 years ago
0
Improve PatchCreator interface, add basic 2D training support

#22 mdraw closed 6 years ago
0
Automatic backup of training code and environment info (resolves #11)

#21 rasaxena closed 6 years ago
1
Ravil

#20 ravil-mobile closed 6 years ago
1
New data augmentation

#19 ravil-mobile closed 6 years ago
0
[WIP] Add local blur ("blob") augmentations

#18 mdraw closed 6 years ago
2
Large-scale prediction using trained models

#17 mdraw closed 5 years ago
1
readthedocs page with auto-generated API documentation

#16 mdraw closed 6 years ago
3
Docstrings and type information

#15 mdraw closed 5 years ago
1
Calculating and visualizing effective (empirical) receptive fields of network models

#14 mdraw closed 5 years ago
0
OOM when using the current PyTorch git master

#13 mdraw closed 6 years ago
2
Random HDF5 read errors

#12 mdraw closed 5 years ago
1
Back up training code to save_path for more reproducible experiments

#11 mdraw closed 6 years ago
2
"Invalid" targets with out-of-bounds elements

#10 mdraw closed 5 years ago
2
Add non-geometric augmentation methods

#9 mdraw closed 5 years ago
1
Plot overlay images to TensorBoard

#8 mdraw closed 6 years ago
0
Early stopping and dynamic learning rate reduction

#7 mdraw closed 5 years ago
1
Modularize and clean up StoppableTrainer

#6 mdraw closed 5 years ago
0
Support 2D data sets

#5 mdraw closed 5 years ago
1
Decide what to do with train.py

#4 mdraw closed 6 years ago
2