-
Python 3.6 will no longer receive security fixes [after 23 Dec 2021](https://endoflife.date/python). SCT relies exclusively on Python 3.6 because we use `tensorflow==1.5.0`. We use this version for us…
-
## Issue description
I followed tutorial for optimize hyperparameter in https://ivadomed.org/en/latest/tutorials/automate_training.html
### Current behavior
I got the error message
```
…
-
## Issue description
I’m testing the [training_curve.py](https://ivadomed.org/en/latest/scripts.html#ivadomed-training-curve) script and found an odd behavior. The script output a plot for each met…
-
## Issue description
I tried segmenting an image and received the following error:
```
>>> nii_lst, target_lst = imed_inference.segment_volume(path_model, input_filenames, options=options)
202…
-
### Motivation for the feature
While training some microscopy models, I encountered dice scores superior to 1 during the testing phase. Investigating this further, I noticed that the dice score com…
-
A recent SOTA [paper](https://arxiv.org/abs/2102.04306) and its [GH repository](https://github.com/Beckschen/TransUNet). It works with 3D images, and we have the potential of handling FLAIR TP1 & TP2 …
-
Trained a 2D multi-channel U-net model with (i) raw input images (located in `~/duke/projects/ivadomed/tmp_ms_challenge_2021`) and (ii) preprocessed input images (according to the latest preprocessing…
-
In #3646, we reworked the way we install dependencies. Part of these changes included modifying how `sct_check_dependencies` verifies requirements for frozen releases:
- Before, we only checked aga…
-
## Issue description
### Current behavior
Whenever I try to train an ivadomed model with RandomeAffine and ElasticDeform transforms I receive an error stating:
`Traceback (most recent ca…
-
Let's curate BIDS-compatible and `ivadomed`-compatible versions of the MS-Challenge 2016 dataset so they are ready for training. @charleygros mentioned before that he has trained some models on this, …