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
PythonMIT
Issues
- 0
Input normalization
#54 opened by SebastienTs - 2
UNet Implementation details
#52 opened by SebastienTs - 0
Elektronn3
#53 opened by SebDBI - 2
Sparse annotations for training
#51 opened by SebastienTs - 0
Refactor augmentations
#2 opened by mdraw - 2
Error with 3D convolution on custom Kernels
#49 opened by Elmiar0642 - 1
How to train the model with custom datasets
#50 opened by Elmiar0642 - 0
noise to zero_like, not empty_like
#43 opened by riegerfr - 2
- 1
Offline validation/evaluation of trained models
#35 opened by mdraw - 0
Inference example script
#34 opened by mdraw - 1
Add non-geometric augmentation methods
#9 opened by mdraw - 1
Import error when running in SyConn environment
#39 opened by AldoCP - 1
- 0
Reproducibility
#28 opened by mdraw - 0
High-level documentation
#33 opened by mdraw - 1
Elastic deformations
#3 opened by mdraw - 0
- 1
- 0
Modularize and clean up StoppableTrainer
#6 opened by mdraw - 1
Docstrings and type information
#15 opened by mdraw - 1
Random HDF5 read errors
#12 opened by mdraw - 0
Calculating and visualizing effective (empirical) receptive fields of network models
#14 opened by mdraw - 2
"Invalid" targets with out-of-bounds elements
#10 opened by mdraw - 1
Support 2D data sets
#5 opened by mdraw - 1
Large-scale prediction using trained models
#17 opened by mdraw - 3
- 2
- 0
Plot overlay images to TensorBoard
#8 opened by mdraw - 2
Decide what to do with train.py
#4 opened by mdraw - 2
OOM when using the current PyTorch git master
#13 opened by mdraw