ELEKTRONN/elektronn3 issues and pull requests

#54 - Input normalization

Issue - State: open - Opened by SebastienTs about 1 year ago

#53 - Elektronn3

Issue - State: closed - Opened by SebDBI about 1 year ago

#52 - UNet Implementation details

Issue - State: open - Opened by SebastienTs about 1 year ago - 2 comments

#51 - Sparse annotations for training

Issue - State: open - Opened by SebastienTs about 1 year ago - 2 comments

#50 - How to train the model with custom datasets

Issue - State: open - Opened by Elmiar0642 over 1 year ago - 1 comment

#49 - Error with 3D convolution on custom Kernels

Issue - State: closed - Opened by Elmiar0642 over 1 year ago - 2 comments

#47 - Trainer scripts and metrics for GNN training

Pull Request - State: closed - Opened by hash-ir over 2 years ago

#46 - Merge phil_convpoint into syconn

Pull Request - State: closed - Opened by pschubert over 3 years ago

#45 - Knossos labels kzip

Pull Request - State: closed - Opened by riegerfr over 3 years ago

#44 - Update requirements.txt

Pull Request - State: closed - Opened by flmuk over 3 years ago

#43 - noise to zero_like, not empty_like

Issue - State: closed - Opened by riegerfr about 4 years ago

#42 - introduced handling for target = None in RandomRotate2d

Pull Request - State: closed - Opened by riegerfr over 4 years ago - 1 comment

#41 - implemented caching for KnossosRawData

Pull Request - State: closed - Opened by riegerfr over 4 years ago - 1 comment

#40 - Various improvements by @Optiligence and @mdraw

Pull Request - State: closed - Opened by mdraw almost 5 years ago

#39 - Import error when running in SyConn environment

Issue - State: closed - Opened by AldoCP about 5 years ago - 1 comment

#38 - Caconi ensemble

Pull Request - State: closed - Opened by caconi about 5 years ago

#37 - Missing assignment to mode in train_unet_neurodata.py

Issue - State: closed - Opened by dmankins over 5 years ago - 1 comment

#36 - train_unet_neurodata.py default seed is a seed of 0, not a random seed

Issue - State: closed - Opened by dmankins over 5 years ago - 2 comments

#35 - Offline validation/evaluation of trained models

Issue - State: closed - Opened by mdraw over 5 years ago - 1 comment
Labels: enhancement

#34 - Inference example script

Issue - State: closed - Opened by mdraw over 5 years ago
Labels: enhancement

#33 - High-level documentation

Issue - State: open - Opened by mdraw over 5 years ago
Labels: documentation

#32 - NaN losses when training with torch.nn.CrossEntropyLoss

Issue - State: closed - Opened by mdraw over 5 years ago
Labels: bug, high priority

#31 - correct metrics with ignore class

Pull Request - State: closed - Opened by Optiligence over 5 years ago - 1 comment

#30 - don’t allocate complete in- and output on the gpu

Pull Request - State: closed - Opened by Optiligence over 5 years ago

#29 - adjust output shape as multiple of tile shape

Pull Request - State: closed - Opened by Optiligence over 5 years ago

#28 - Reproducibility

Issue - State: closed - Opened by mdraw over 5 years ago

#27 - fix metrics for i.e. negative gt batches

Pull Request - State: closed - Opened by Optiligence over 5 years ago

#26 - ElasticTransform and RandomGaussianBlur transforms

Pull Request - State: closed - Opened by mdraw over 5 years ago - 1 comment

#25 - Add Predictor, TripletNet, RandomFlip, 2D FCN

Pull Request - State: closed - Opened by pschubert about 6 years ago - 2 comments

#24 - Enable automatic documentation build and hosting on readthedocs

Pull Request - State: closed - Opened by mdraw about 6 years ago

#23 - New transforms API for datasets

Pull Request - State: closed - Opened by mdraw over 6 years ago

#22 - Improve PatchCreator interface, add basic 2D training support

Pull Request - State: closed - Opened by mdraw over 6 years ago

#21 - Automatic backup of training code and environment info (resolves #11)

Pull Request - State: closed - Opened by rasaxena over 6 years ago - 1 comment

#20 - Ravil

Pull Request - State: closed - Opened by ravil-mobile over 6 years ago - 1 comment

#19 - New data augmentation

Pull Request - State: closed - Opened by ravil-mobile over 6 years ago

#18 - [WIP] Add local blur ("blob") augmentations

Pull Request - State: closed - Opened by mdraw over 6 years ago - 2 comments

#17 - Large-scale prediction using trained models

Issue - State: closed - Opened by mdraw over 6 years ago - 1 comment
Labels: enhancement, discussion

#16 - readthedocs page with auto-generated API documentation

Issue - State: closed - Opened by mdraw over 6 years ago - 3 comments
Labels: documentation

#15 - Docstrings and type information

Issue - State: closed - Opened by mdraw over 6 years ago - 1 comment
Labels: documentation

#14 - Calculating and visualizing effective (empirical) receptive fields of network models

Issue - State: closed - Opened by mdraw over 6 years ago
Labels: enhancement, discussion

#13 - OOM when using the current PyTorch git master

Issue - State: closed - Opened by mdraw over 6 years ago - 2 comments
Labels: bug

#12 - Random HDF5 read errors

Issue - State: closed - Opened by mdraw over 6 years ago - 1 comment
Labels: bug

#11 - Back up training code to save_path for more reproducible experiments

Issue - State: closed - Opened by mdraw over 6 years ago - 2 comments
Labels: enhancement

#10 - "Invalid" targets with out-of-bounds elements

Issue - State: closed - Opened by mdraw over 6 years ago - 2 comments
Labels: bug

#9 - Add non-geometric augmentation methods

Issue - State: closed - Opened by mdraw over 6 years ago - 1 comment
Labels: enhancement

#8 - Plot overlay images to TensorBoard

Issue - State: closed - Opened by mdraw over 6 years ago
Labels: enhancement

#7 - Early stopping and dynamic learning rate reduction

Issue - State: closed - Opened by mdraw over 6 years ago - 1 comment
Labels: enhancement

#6 - Modularize and clean up StoppableTrainer

Issue - State: closed - Opened by mdraw over 6 years ago
Labels: refactoring

#5 - Support 2D data sets

Issue - State: closed - Opened by mdraw over 6 years ago - 1 comment
Labels: enhancement

#4 - Decide what to do with train.py

Issue - State: closed - Opened by mdraw over 6 years ago - 2 comments
Labels: discussion

#3 - Elastic deformations

Issue - State: closed - Opened by mdraw over 6 years ago - 1 comment
Labels: enhancement

#2 - Refactor augmentations

Issue - State: closed - Opened by mdraw over 6 years ago
Labels: refactoring

#1 - Rewrite data handling

Pull Request - State: closed - Opened by mdraw over 6 years ago

GitHub / ELEKTRONN/elektronn3 issues and pull requests