Skip to content
Snippets Groups Projects

preliminary source_classification ML integration

  • source_classification is based on ML estimates
  • classifiers read from emfollow/data to produce estimates per event basis

CC @shasvath.kapadia @shaon.ghosh

Edited by Leo Pound Singer

Merge request reports

Loading
Loading

Activity

Filter activity
  • Approvals
  • Assignees & reviewers
  • Comments (from bots)
  • Comments (from users)
  • Commits & branches
  • Edits
  • Labels
  • Lock status
  • Mentions
  • Merge request status
  • Tracking
  • Leo Pound Singer
  • Leo Pound Singer
  • Leo Pound Singer
  • Those two test pickle files are pretty big. Can they be slimmed down? Remember test data need not be physically realistic; it must only satisfy the requirements of your format.

  • Leo Pound Singer
  • Leo Pound Singer
  • Leo Pound Singer
  • Leo Pound Singer
  • What does scaler_set_all mean? Is it a misspelling of scalar?

  • Leo Pound Singer
  • Deep Chatterjee added 7 commits

    added 7 commits

    Compare with previous version

  • @leo-singer The pickle files in the test are serving dual purpose. It firstly provides us the data for the unittests. It is also a fallback option for the EM-Bright inference, in case the connectivity fails for some reason (see: https://git.ligo.org/deep.chatterjee/gwcelery/blob/source-classification-ML/gwcelery/tasks/em_bright.py#L51-60 and https://git.ligo.org/lscsoft/p-astro/blob/master/ligo/em_bright.py#L68-76).

    This might not be the best option. I will be happy hear any better suggestion.

    Edited by Shaon Ghosh
  • Leo Pound Singer
  • @leo-singer The pickle files in the test are serving dual purpose. It firstly provides us the data for the unittests. It is also a fallback option for the EM-Bright inference, in case the connectivity fails for some reason (see: https://git.ligo.org/deep.chatterjee/gwcelery/blob/source-classification-ML/gwcelery/tasks/em_bright.py#L51-60 and https://git.ligo.org/lscsoft/p-astro/blob/master/ligo/em_bright.py#L68-76).

    This might not be the best option. I will be happy hear any better suggestion.

    You could keep the files in the p-astro repository.

  • Where are we with this?

  • Deep Chatterjee added 16 commits

    added 16 commits

    Compare with previous version

  • added 1 commit

    • 9bc85552 - preliminary source_classification unittests

    Compare with previous version

  • added 1 commit

    • 21fcb54d - preliminary source_classification unittests

    Compare with previous version

  • Deep Chatterjee added 4 commits

    added 4 commits

    Compare with previous version

  • Deep Chatterjee changed the description

    changed the description

  • Shaon Ghosh added 1 commit

    added 1 commit

    • 7878d197 - puttting back the old source classification scheme from OPA for non-gstLAL triggers

    Compare with previous version

  • Shaon Ghosh added 1 commit

    added 1 commit

    • 7e5fd1fd - Added rem mass computation for the point estimate case. Fixed naming of the...

    Compare with previous version

  • Shaon Ghosh added 1 commit

    added 1 commit

    Compare with previous version

  • added 1 commit

    • 6e269116 - don't pass coinc or psd to classifier tasks anymore

    Compare with previous version

  • Shaon Ghosh
  • added 1 commit

    • 11912a8c - added and pinned p-astro version in requirements

    Compare with previous version

  • Deep Chatterjee added 10 commits

    added 10 commits

    • 11912a8c...49112f42 - 4 commits from branch emfollow:master
    • 69d0dfc9 - preliminary source_classification unittests
    • adb5accc - puttting back the old source classification scheme from OPA for non-gstLAL triggers
    • 9f47d24a - Added rem mass computation for the point estimate case. Fixed naming of the...
    • c39241ad - Fixed syntax error
    • 634e7fb7 - don't pass coinc or psd to classifier tasks anymore
    • f4b67ed8 - added and pinned p-astro version in requirements

    Compare with previous version

  • Leo Pound Singer
  • added 1 commit

    Compare with previous version

  • Deep Chatterjee added 11 commits

    added 11 commits

    • 8fe17e70...f29f0ce2 - 4 commits from branch emfollow:master
    • e12a057e - preliminary source_classification unittests
    • f37c9b6e - puttting back the old source classification scheme from OPA for non-gstLAL triggers
    • b8cf1774 - Added rem mass computation for the point estimate case. Fixed naming of the...
    • 8a43f293 - Fixed syntax error
    • 1c930cce - don't pass coinc or psd to classifier tasks anymore
    • 23dcd477 - added and pinned p-astro version in requirements
    • c9085a0f - increased coverage

    Compare with previous version

  • added 1 commit

    Compare with previous version

  • Shaon Ghosh added 1 commit

    added 1 commit

    Compare with previous version

  • Shaon Ghosh added 1 commit

    added 1 commit

    Compare with previous version

  • added 1 commit

    Compare with previous version

  • Leo Pound Singer unmarked as a Work In Progress

    unmarked as a Work In Progress

  • Leo Pound Singer
  • Leo Pound Singer
  • Leo Pound Singer
  • Leo Pound Singer
  • Leo Pound Singer
  • Leo Pound Singer
  • Leo Pound Singer
  • Leo Pound Singer
  • Loading
  • Loading
  • Loading
  • Loading
  • Loading
  • Loading
  • Loading
  • Loading
  • Loading
  • Loading
  • Please register or sign in to reply
    Loading