neuropoly / axondeepseg / 890

88%

push

travis-ci

Keras integration (#221)

* AxonDeePSeg to keras (#216)

* Notebook implementation of AxonDeepSeg(mini version)

* Notebook implementation of AxonDeepSeg(mini version)

* Incorpoation of Myelin and Axond dice coefficient and Cleaning of notebook

* Update requirements.txt

Updated Requirements.txt for Keras framework

* Tensorborad support for visualization and removed redundacy in pw_dice metric

* Data Augmentation support for improving the model

* Update unet-segmentation.ipynb

* Cleaning of Notebook and modular support for UNET

* Update Notebook

* Add files via upload

* Intergerated Segmentation support of Keras Model to ADS

* Add Keras SEM model to AxonDeepSeg/molel

* Fixed training error in Keras

* Fixes train_network.py path error and integration with travis.

* Fixes session closed issue on Travis.

* Remove unnecessary stub and files.

* Delete data_prep.py

* Delete model.py

* Delete TEM_striatum sample model.

* Remove notebook checkpoint folder and add gitignore rule

* Cleaning of apply_model, input_data, network_construction and train_network

* Temporary fix to hard coding of number of epochs (#217)

* Fixed saving best loss and accuracy in ckpt format instead of hdf5 (#221)

* Fixed saving best loss and accuracy in ckpt format instead of hdf5 (#221)

* Fixed training_guideline.ipynb (#221)

* Fixes cleaner visualization of U-net model graph in Tensorboard(#221)

* Fixes suppo

* Fixes cleaner visualization of U-net model graph in Tensorboard(#221)

* Fixes cleaner visualization of U-net model graph in Tensorboard(#221)

* Add TEM model (#221)

* Fixes docstring issue(#221)

* Fixes docstring issue(#221)

* Fixes documentation(RTD + changelog) issue(#221)

* Delete data_augmentation.py(#221)

* Update CHANGELOG file

* Revert GPU instructions

* Reinstate import AxonDeepSeg.ads_utils for bug tracking

* Remove debugging print call

* Remove trailing... (continued)

173 of 173 new or added lines in 8 files covered. (100.0%)

1644 of 1987 relevant lines covered (82.74%)

0.83 hits per line