-
Notifications
You must be signed in to change notification settings - Fork 6
Issues: ML4GW/DeepClean
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Author
Label
Projects
Milestones
Assignee
Sort
Issues list
Correlate test metrics with parameter recovery
question
Further information is requested
research topic
Question about DeepClean optimization and interpretation
#65
opened Oct 13, 2022 by
alecgunny
in New feature or request
gwftools.find
, add logic for finding files that contain superset of data needed
enhancement
#60
opened Aug 2, 2022 by
alecgunny
Fix probabilistic behavior in PSD loss test in Something isn't working
documentation
Improvements or additions to documentation
good first issue
Good for newcomers
deepclean.trainer
bug
#56
opened Jun 2, 2022 by
alecgunny
How to name online framefiles when there is non-integer aggregation latency?
#40
opened Mar 26, 2022 by
saleem-muhammed
Images for the wiki
documentation
Improvements or additions to documentation
good first issue
Good for newcomers
#38
opened Mar 25, 2022 by
alecgunny
Build experiment tracking framework
research topic
Question about DeepClean optimization and interpretation
#35
opened Mar 25, 2022 by
alecgunny
Parameterize autoencoder architecture
hyperparameters
Research topic about hyperparameters used to train DeepClean
research topic
Question about DeepClean optimization and interpretation
#34
opened Mar 24, 2022 by
alecgunny
Can we make transformers work here?
architecture
Research topic pertaining to DeepClean architecture
research topic
Question about DeepClean optimization and interpretation
#33
opened Mar 24, 2022 by
alecgunny
Using convolutional kernels with sizes that are multiples of 8
architecture
Research topic pertaining to DeepClean architecture
research topic
Question about DeepClean optimization and interpretation
#32
opened Mar 24, 2022 by
alecgunny
Relationship between receptive field and learnable bandwidth
architecture
Research topic pertaining to DeepClean architecture
research topic
Question about DeepClean optimization and interpretation
#31
opened Mar 24, 2022 by
alecgunny
DeepClean receptive field
architecture
Research topic pertaining to DeepClean architecture
research topic
Question about DeepClean optimization and interpretation
#30
opened Mar 24, 2022 by
alecgunny
Shuffling produces empirically worse results
data
Research topic about data used to train DeepClean
research topic
Question about DeepClean optimization and interpretation
#28
opened Mar 24, 2022 by
alecgunny
Train stride tradeoffs
data
Research topic about data used to train DeepClean
research topic
Question about DeepClean optimization and interpretation
#27
opened Mar 24, 2022 by
alecgunny
Performance on data at the edge of kernels
loss function
Research topic pertaining to DeepClean loss function
research topic
Question about DeepClean optimization and interpretation
#26
opened Mar 24, 2022 by
alecgunny
Alternative loss functions
loss function
Research topic pertaining to DeepClean loss function
research topic
Question about DeepClean optimization and interpretation
#24
opened Mar 24, 2022 by
alecgunny
3 tasks
How many FFTs required for stable ASD loss?
loss function
Research topic pertaining to DeepClean loss function
research topic
Question about DeepClean optimization and interpretation
#23
opened Mar 24, 2022 by
alecgunny
Replacing No-data channels with dummy dc timeseries
#20
opened Mar 11, 2022 by
saleem-muhammed
2 tasks
Issue templates
documentation
Improvements or additions to documentation
good first issue
Good for newcomers
#14
opened Feb 11, 2022 by
alecgunny
Previous Next
ProTip!
What’s not been updated in a month: updated:<2024-12-09.