Skip to content
This repository has been archived by the owner on Nov 16, 2023. It is now read-only.

Issues: microsoft/seismic-deeplearning

53 Open 152 Closed
53 Open 152 Closed
Author
Filter by author
Loading
Label
Filter by label
Loading
Use alt + click/return to exclude labels
or ⇧ + click/return for logical OR
Projects
Filter by project
Loading
Milestones
Filter by milestone
Loading
Assignee
Filter by who’s assigned
Sort
Sort by
Newest Oldest Most commented Least commented Recently updated Least recently updated Best match
Most reactions
👍 👎 😄 🎉 😕 ❤️ 🚀 👀

Issues list

fix seeds in train.py and test.py to make results perfectly reproducible on a single GPU - enhances the robustness of results Investment: High how long it takes to do a work item Investment: Medium how long it takes to do a work item Type: Enhancement This an enhancement to an existing feature
#382 opened Jun 24, 2020 by maxkazmsft V0.2.1 [July release]
add a clean-up job to remove output directories from build VM Prior: Low Type: Enhancement This an enhancement to an existing feature
#377 opened Jun 19, 2020 by fazamani V0.2.1 [July release]
cache Penbscot and Dutch F3 data on build VM and download weekly Prior: Medium Type: Enhancement This an enhancement to an existing feature
#376 opened Jun 19, 2020 by fazamani V0.2.1 [July release]
operator1 and operator2 close-out Prior: High
#374 opened Jun 15, 2020 by maxkazmsft V0.2 [June release]
bug bash - facilitates gathering feedback from stakeholders Prior: High
#373 opened Jun 15, 2020 by maxkazmsft V0.2 [June release]
Add data QC module Investment: Low how long it takes to do a work item Prior: High Type: Feature New feature or request
#370 opened Jun 15, 2020 by fazamani V0.2.1 [July release]
Dynamic global data information, reduces user-required specifications Investment: Low how long it takes to do a work item Prior: Low Type: Feature New feature or request
#369 opened Jun 15, 2020 by fazamani V0.2.1 [July release]
Azure ML training pipeline improvement comments - facilitates better user onboarding for the pipeline Prior: High
#366 opened Jun 11, 2020 by maxkazmsft V0.2.1 [July release]
add pre-trained models to test section of the HRNet notebooks - facilitates faster onboarding with more pre-trained models Prior: Low Type: Enhancement This an enhancement to an existing feature
#364 opened Jun 11, 2020 by maxkazmsft V0.2.1 [July release]
add UNet pre-trained model - facilitates faster onboarding with UNet model Prior: Low Type: Enhancement This an enhancement to an existing feature
#363 opened Jun 11, 2020 by maxkazmsft V0.2.1 [July release]
make pre-run notebooks for segyconverter utility and SEResNet available in main README - facilitates faster quick-start experience Prior: High Type: Enhancement This an enhancement to an existing feature Type: Gap Not feature but a missing item to an existing feature
#361 opened Jun 11, 2020 by maxkazmsft V0.2.1 [July release]
enable Azure ML pipeline from CSE team - adds AML training pipeline to codebase (models can be trained on AML) Prior: High Type: Feature New feature or request
#360 opened Jun 11, 2020 by maxkazmsft V0.2.1 [July release]
Remove depth channel from Tensorboard visualization - makes grayscale images appear as such in TensorBoard logging Prior: High Type: Feature New feature or request
#355 opened Jun 8, 2020 by yalaudah V0.2 [June release]
add GPU profile information to multi-GPU train and test scripts - facilitates ability to view possible code bottlenecks Prior: High Type: Feature New feature or request
#348 opened Jun 7, 2020 by maxkazmsft V0.2.1 [July release]
@yalaudah
debug test.py not limiting test output in debug mode - facilitates faster build runtimes Prior: High Type: Bug Something isn't working Type: Enhancement This an enhancement to an existing feature
#331 opened May 28, 2020 by maxkazmsft Outstanding items
debug test.py patch padding before model scoring - facilitates scoring correctness and higher scoring accuracy Prior: Medium Type: Accuracy related to increasing performance accuracy Type: Bug Something isn't working
#330 opened May 28, 2020 by maxkazmsft V0.2.1 [July release]
add unit test to check that each input slice has uniform distribution - facilitates uniform sampling of training data for better model accuracy Prior: High Type: Accuracy related to increasing performance accuracy
#319 opened May 26, 2020 by maxkazmsft V0.2.1 [July release]
1
add code coverage badge to github front page - facilitates visual certainty in unit test code coverage Prior: High Type: Feature New feature or request
#316 opened May 26, 2020 by maxkazmsft V0.2.1 [July release]
Add serialization/deserialization unit tests - checks whether writing model snapshots to disc actually works properly Prior: Medium
#303 opened May 18, 2020 by yalaudah V0.2.1 [July release]
@yalaudah
Add unit tests for _patch_label_2d()
#301 opened May 11, 2020 by yalaudah
Generated data splits should be tracked along with model outputs, and not stored with data. Type: Enhancement This an enhancement to an existing feature
#286 opened Apr 23, 2020 by yalaudah V0.2.1 [July release]
Host DutchF3 on Azure open datasets - overcomes reliability problem of data availability Prior: Low Type: Enhancement This an enhancement to an existing feature
#279 opened Apr 21, 2020 by sharatsc Outstanding items
@yalaudah
read model parameters from configuration file for texture_net - facilitates easy model architecture changes Prior: Low Type: Enhancement This an enhancement to an existing feature
#277 opened Apr 20, 2020 by maxkazmsft Outstanding items
make output transform unpad and scale down mask in penobscot metrics - facilitates correctly reported inline mIoU Prior: Medium Type: Accuracy related to increasing performance accuracy Type: Enhancement This an enhancement to an existing feature
#276 opened Apr 20, 2020 by maxkazmsft V0.2.1 [July release]
add logging to data.py methods - facilitates better logging Prior: Low Type: Enhancement This an enhancement to an existing feature
#271 opened Apr 20, 2020 by maxkazmsft Outstanding items
ProTip! Add no:assignee to see everything that’s not assigned.