Skip to content
[Jenkins]Jenkins
log in
Success

Changes

Summary

  1. QuadTune: Display maps of the first 4 SVD patterns using dash, rather than make_subplots. (details)
  2. QuadTune: Create 2 new plots. (details)
  3. QuadTune: Break out loss function contributions by metric and plot the metrics with the greatest improvements and degradations. (details)
  4. QuadTune: Break out loss function contributions by metric and plot the metrics with the greatest improvements and degradations. (details)
  5. Small improvements to bindiff script. (details)
Commit 81fa0e3bef38ab4a734aa95e5f8b4dacb0776743 by Vince Larson
QuadTune: Display maps of the first 4 SVD patterns using dash, rather than make_subplots.

Also, show maps of the sensitivity to individual parameters.

For #910.
The file was modified utilities/sens_matrix/create_figs.py (diff)
Commit 0bb668e3a21dce9af466abc9f0e747fbb7b36eab by Vince Larson
QuadTune: Create 2 new plots.

1) colored matrix plot of SVD vh matrix; and
2) bar chart of importance of parameters

For #910.
The file was modified utilities/sens_matrix/create_figs.py (diff)
Commit e525bd0e3c6333eb4b9945f6fe81d7a9e3cfd3e3 by Vince Larson
QuadTune: Break out loss function contributions by metric and plot the metrics with the greatest improvements and degradations.

Also plot signed sum of contributions, in order to see how biases affect which parameters are tuned.

For #910.
The file was modified utilities/sens_matrix/sens_matrix_dashboard.py (diff)
Commit d7d04ed1c8a65801bfc91bbbe4227ef39f7f7292 by Vince Larson
QuadTune: Break out loss function contributions by metric and plot the metrics with the greatest improvements and degradations.

Also plot signed sum of contributions, in order to see how biases affect which parameters are tuned.

For #910.
The file was modified utilities/sens_matrix/create_figs.py (diff)
Commit 2651012ea532b1d38ef6006d7457571806e34908 by Gunther Huebler
Small improvements to bindiff script.
The file was modified run_scripts/run_bindiff_all.py (diff)