issues
search
SteveOv
/
ebop_maven
EBOP Model Automatic input Value Estimation Neural network
GNU General Public License v3.0
1
stars
0
forks
source link
issues
Newest
Newest
Most commented
Recently updated
Oldest
Least commented
Least recently updated
add support for "adaptive wrap" of mags feature
#83
SteveOv
opened
2 weeks ago
1
add support for parameter scatter plots from JKTEBOP task 8 & 9 out files
#82
SteveOv
opened
2 weeks ago
0
new versions of TensorFlow/Keras leads to changed training result
#81
SteveOv
closed
3 weeks ago
3
Investigate predictions on "swap" synthetic test dataset
#80
SteveOv
opened
1 month ago
5
Rework trainset/dataset code so that it can flip the components/roll mags if secondary is deeper
#79
SteveOv
closed
1 month ago
3
investigate production of correlation plots between labels & estimates
#78
SteveOv
opened
2 months ago
0
Rework deb_example/datasets so only a single mags feature per mags_bins
#77
SteveOv
closed
2 months ago
0
Add ability to run jktebop task 8|9 fit to the interactive tester
#76
SteveOv
closed
2 months ago
4
interactive tester does not behave the same way as model_testing when using labels for control fit
#75
SteveOv
closed
2 months ago
1
Review default handling of qphot (and fitting flag) for control and test fits
#74
SteveOv
closed
2 months ago
1
Overnight model search stopped training after ~20 evals
#73
SteveOv
closed
2 months ago
5
add qphot to the train/test datasets
#72
SteveOv
closed
2 months ago
3
better handling of qphot
#71
SteveOv
closed
2 months ago
9
Rework model_search so that it is resumable
#70
SteveOv
closed
2 months ago
0
investigate learning rate scheduling
#69
SteveOv
closed
2 months ago
5
investigate a way of detecting if secondary will be deeper than primary when generating a train/test instance
#68
SteveOv
closed
2 months ago
5
Rework modelling.output_layer() so that we can support >1 activation
#67
SteveOv
closed
2 months ago
2
Investigate mitigations for "unphysical" prediction values
#66
SteveOv
closed
2 months ago
9
review re-scaling code in estimator.predict()
#65
SteveOv
closed
2 months ago
1
Create training dataset for test_estimator
#64
SteveOv
opened
2 months ago
0
investigate the use of numpy structured arrays in the output from Estimator
#63
SteveOv
closed
2 months ago
5
refine population synthesis for synthetic-mist-tess-dataset
#62
SteveOv
closed
2 months ago
5
Look to integrate FFI processing into ingest pipeline
#61
SteveOv
opened
3 months ago
2
Add a simple tutorial jupyter page for example maven
#60
SteveOv
opened
3 months ago
0
Potentially create a new deblib package from the libs folder of ebop_maven
#59
SteveOv
opened
3 months ago
1
update structure; move trainsets and datasets functionality to the root
#58
SteveOv
closed
3 months ago
3
Can we change the JKTEBOP negative L3 error to a warning (& set to zero)?
#57
SteveOv
closed
2 months ago
0
Review training set diststributions for k, J and e
#56
SteveOv
closed
3 months ago
6
check for consistent use to constrained or tight layout on plots
#55
SteveOv
closed
2 months ago
2
new plot: loss vs label value with support for > label on the same axes
#54
SteveOv
closed
3 months ago
2
model_testing: add boxplots for transiting/non-trans for the full synth test dataset
#53
SteveOv
closed
3 months ago
1
capture training log
#52
SteveOv
closed
3 months ago
0
interactive_tester incorrect format in models dropdown
#51
SteveOv
closed
3 months ago
0
code for creating JKTEBOP file names not working under venv
#50
SteveOv
closed
3 months ago
0
move the MIST isos out of the ebop_maven/libs directory structure
#49
SteveOv
closed
3 months ago
1
add ability to configure mags augmentations through the read_dataset() path
#48
SteveOv
closed
3 months ago
0
investigate using a matplotlib cmap on the H-R and pred-vs-label plots to indicate stars approximate colour
#47
SteveOv
opened
3 months ago
1
review best way to work with MIST models; isos or eeps?
#46
SteveOv
closed
3 months ago
2
reduce height of histogram plots - taking up too much space in paper
#45
SteveOv
closed
3 months ago
0
Update StellarModels to resolve the pandas FutureWarning
#44
SteveOv
closed
3 months ago
0
make_dataset_files() doesn't support setting test_ratio arg to 1.0
#43
SteveOv
closed
3 months ago
0
investigate discrepancy reported MAE/MSE values from model_testing
#42
SteveOv
closed
3 months ago
0
Create a new ebop_mave/pipeline module
#41
SteveOv
closed
4 months ago
1
Extend Estimator so that it's possible to access all MC Dropout predictions
#40
SteveOv
closed
4 months ago
0
Consider HO Tel for formal-test-dataset
#39
SteveOv
closed
4 months ago
1
add flatten() functionality to ingest pipeline
#38
SteveOv
closed
4 months ago
0
a more formal approach to omitting systems from formal-test-dataset
#37
SteveOv
closed
4 months ago
0
move deb_example from libs to root of ebop_maven
#36
SteveOv
closed
4 months ago
0
add post detrending clipping to the simple ingest pipeline (a la poc debfit pipeline)
#35
SteveOv
closed
4 months ago
0
Look into adding R2 Score to model_testing as this may be better for comparing models than MAE/MSE
#34
SteveOv
closed
4 months ago
0
Next