Mru tests #38
Open
Mru tests #38
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This code has not been properly tested, I'm just trying to get some results.
Commented out a line 'from timing import time_this' because it caused the import to fail on Colaboratory as this was a local import from another repository.
…d then reinstalled for commit changes to take effect.
…t to call train.train_segnet.train_segnet() with the time_log() decorator.
Using utils.timing.time_log(), to measure training times while running the notebook on Google Colab.
To train/train_segnet.train_segnet(). This to facilitate logging.
To enable easy logging of ideas.
Comments are now welcome when calling train_segnet()
I had accidentally saved them in / instead of /example_notebooks/
MRU-Net as multiresunet3.py
Added metrics defined by [Abdiel](https://github.com/Abdiel-EMT).
…implementing everything as close to the article as possible.
…tructure that will most likely change in future commits.
When exporting them from Colab, I had forgotten to correctly include the path.
The files were almost identical, this is my branch and the purpose of that file was to include the new metrics defined by Abdiel on the experiments done on my branch.
…it.edu/isbi_challenge/home. Trying to reproduce the original MultiResUNet paper's results.
…ed.create_train_test_generators, and a method to show images and segmentation maps side by side. I'm thinking of including thresholding because some of the images on the Chinese dataset are not binary.
…th a decorator that takes arguments, one cannot use the syntactic sugar @decorator(params)/method as instance attributes are not defined at the time of parsing the class. Instead the modification of the function should happen within __init__ as self.method = decorator(params)(self.method).
This contains enormous, gigantic leaps of the development of Segmed utility class.
Instead of casting non-JSON-serialisable with str(type(x)), they will now be casted using str(x) as this allows saving the name of functions, for instance, which is very useful.
Added a comment with an idea to improve legibility of the the JSON-Lines logs.
… logging decorators and the plot utility functions and the class will be complete.
This version is now explicit, specifying in the logs (timing.time_log) that the execution time is measured in seconds. 'time' -> 'execution time (s)'
…ate all files into a directory.
…lling the train method takes forever and does not seem to progress.
…me execution thus tests are on standby.
…s a text file. Removed a lot of legacy cells that were not even consistent with the current model implementation.
requirements.txt now includes gputil.
…u_tests I had been working on a Colab Jupyter Notebook to develop the Segmed class. I had not pulled from remote to my computer and then I modified a requitements.txt to add a dependency (gputil is needed by the class Segmed).
It enables experimentation with automatic logging.
…u_tests There is really no problem here, I shouln't have to comment on this merge.
concise.py
It is still unused, but the docstring for it has been written. It belongs to segnet.utils.Segmed.train
Found a bug while using the "TPU" option on Google Collab. The exception raised was: ValueError: invalid literal for int() with base 10: "NVIDIA-SMI has failed because it couldn't communicate with the NVIDIA driver. Make sure that the latest NVIDIA driver is installed and running."
These show the potential of using Segmed class.
Inside a comment in segmed.utils.Segmed.Segmed.train
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
I've created an example notebook for MultiResUNet.