Commits

Breadcrumbs

History for

onjdev

User selector

All users

All time

Commit history

Commits on Mar 25, 2018

• ensure shot is only cut if it remains longer than RNN length after cutting

  

  Julian Kates-Harbeck

  committed
  4dc0edbView commit details

  Copy full SHA for 4dc0edb

  View code at this pointBrowse repository at this point

Commits on Mar 24, 2018

• fix small error in variable ordering

  

  Julian Kates-Harbeck

  committed
  4054c03View commit details

  Copy full SHA for 4054c03

  View code at this pointBrowse repository at this point

• fix bugs related to T_min warn becoming too large. performance computation was wrong when T_max_warn was smaller. Shots can't be cut if they are shorter than T_min_warn

  

  Julian Kates-Harbeck

  committed
  bbfa51bView commit details

  Copy full SHA for bbfa51b

  View code at this pointBrowse repository at this point

Commits on Mar 7, 2018

• make normalizer resilient for when conf does not contain norm_stat_range

  

  Julian Kates-Harbeck

  committed
  ac7df4eView commit details

  Copy full SHA for ac7df4e

  View code at this pointBrowse repository at this point

Commits on Feb 9, 2018

• not saving from within preprocessor class

  

  Julian Kates-Harbeck

  authored and

  Julian Kates-Harbeck

  committed
  374e069View commit details

  Copy full SHA for 374e069

  View code at this pointBrowse repository at this point

• merged

  

  Julian Kates-Harbeck

  authored and

  Julian Kates-Harbeck

  committed
  1827332View commit details

  Copy full SHA for 1827332

  View code at this pointBrowse repository at this point

Commits on Feb 7, 2018

• backwards compatibility for normalizer (check if signal has is_positive attribute). Plotting options for performance analyzer

  

  Julian Kates-Harbeck

  committed
  85d4b8bView commit details

  Copy full SHA for 85d4b8b

  View code at this pointBrowse repository at this point

Commits on Jan 24, 2018

• added option for partial multiplication factor in bleed in

  

  Julian Kates-Harbeck

  committed
  dae8bf0View commit details

  Copy full SHA for dae8bf0

  View code at this pointBrowse repository at this point

Commits on Jan 11, 2018

• Resolve merge conflicts

  ASvyatkovskiy

  committed
  4823984View commit details

  Copy full SHA for 4823984

  View code at this pointBrowse repository at this point

• bleed in

  

  Julian Kates-Harbeck

  committed
  ec4ed03View commit details

  Copy full SHA for ec4ed03

  View code at this pointBrowse repository at this point

• bleed in option of adding multiple times

  

  Julian Kates-Harbeck

  authored and

  Julian Kates-Harbeck

  committed
  0f3d904View commit details

  Copy full SHA for 0f3d904

  View code at this pointBrowse repository at this point

• changed bleed in to sample until num disruptive shots have been sampled

  

  Julian Kates-Harbeck

  authored and

  Julian Kates-Harbeck

  committed
  e61732dView commit details

  Copy full SHA for e61732d

  View code at this pointBrowse repository at this point

Commits on Dec 8, 2017

• Clip to good statistical tange after normalization

  ASvyatkovskiy

  committed
  eb67aaeView commit details

  Copy full SHA for eb67aae

  View code at this pointBrowse repository at this point

Commits on Dec 7, 2017

• Keep all_signals for preprocessing, adjust all_signals in the conf_parser.py instead

  ASvyatkovskiy

  committed
  d69e64aView commit details

  Copy full SHA for d69e64a

  View code at this pointBrowse repository at this point

• all_signals -> use_signals

  ASvyatkovskiy

  committed
  e09122fView commit details

  Copy full SHA for e09122f

  View code at this pointBrowse repository at this point

• all_signals -> use_signals

  ASvyatkovskiy

  committed
  4f40889View commit details

  Copy full SHA for 4f40889

  View code at this pointBrowse repository at this point

• Reduce verbosity

  ASvyatkovskiy

  authored
  d77d570View commit details

  Copy full SHA for d77d570

  View code at this pointBrowse repository at this point

Commits on Dec 6, 2017

• Signal positivity constraint, rebase

  ASvyatkovskiy

  committed
  b2156fcView commit details

  Copy full SHA for b2156fc

  View code at this pointBrowse repository at this point

• Fix ShotList type mismatch

  ASvyatkovskiy

  committed
  d8193b7View commit details

  Copy full SHA for d8193b7

  View code at this pointBrowse repository at this point

• Fix NameError: global name 'conf' is not defined

  ASvyatkovskiy

  authored
  ebf33eeView commit details

  Copy full SHA for ebf33ee

  View code at this pointBrowse repository at this point

Commits on Dec 4, 2017

• using signal hiding to compute influence

  

  Julian Kates-Harbeck

  committed
  30df67bView commit details

  Copy full SHA for 30df67b

  View code at this pointBrowse repository at this point

Commits on Dec 1, 2017

• Merge branch 'master' of https://github.com/PPPLDeepLearning/plasma-python

  

  Julian Kates-Harbeck

  authored and

  Julian Kates-Harbeck

  committed
  0b3cb25View commit details

  Copy full SHA for 0b3cb25

  View code at this pointBrowse repository at this point

• added functionality to compute signal influence by hiding part of the signal in time and performing inference with that modified signal

  

  Julian Kates-Harbeck

  authored and

  Julian Kates-Harbeck

  committed
  c56430eView commit details

  Copy full SHA for c56430e

  View code at this pointBrowse repository at this point

Commits on Nov 27, 2017

• During augmentation, replace signal with noise instead of add (note: this is done after normalization)

  ASvyatkovskiy

  committed
  baf06faView commit details

  Copy full SHA for baf06fa

  View code at this pointBrowse repository at this point

Commits on Nov 11, 2017

• only update normalizer with those machines that are missing

  

  Julian Kates-Harbeck

  authored and

  Julian Kates-Harbeck

  committed
  ea375aeView commit details

  Copy full SHA for ea375ae

  View code at this pointBrowse repository at this point

• Merge branch 'master' of https://github.com/PPPLDeepLearning/plasma-python

  

  Julian Kates-Harbeck

  committed
  25a8122View commit details

  Copy full SHA for 25a8122

  View code at this pointBrowse repository at this point

• only update normalizer with those machines that are missing

  

  Julian Kates-Harbeck

  authored and

  Julian Kates-Harbeck

  committed
  3a16cacView commit details

  Copy full SHA for 3a16cac

  View code at this pointBrowse repository at this point

• only update normalizer with those machines that are missing

  

  Julian Kates-Harbeck

  authored and

  Julian Kates-Harbeck

  committed
  77d90d3View commit details

  Copy full SHA for 77d90d3

  View code at this pointBrowse repository at this point

• fixed overflow error in normalization

  

  Julian Kates-Harbeck

  committed
  d4a3041View commit details

  Copy full SHA for d4a3041

  View code at this pointBrowse repository at this point

Commits on Nov 10, 2017

• changed handling of shot cutting. Shots are only cut during training but not during inference.

  

  Julian Kates-Harbeck

  authored and

  Julian Kates-Harbeck

  committed
  e5529bbView commit details

  Copy full SHA for e5529bb

  View code at this pointBrowse repository at this point

• path changes based on group of signals used for selecting valid shots. If all_signals is changed, the group of shots considered valid (because all signals are present) and their lengths (because of…

  Show description for 25adba3
  

  Julian Kates-Harbeck

  authored and

  Julian Kates-Harbeck

  committed
  25adba3View commit details

  Copy full SHA for 25adba3

  View code at this pointBrowse repository at this point

• small bug fixes

  jnkh

  committed
  3b5ead4View commit details

  Copy full SHA for 3b5ead4

  View code at this pointBrowse repository at this point

Commits on Nov 9, 2017

• small bug in preprocessing path

  jnkh

  committed
  470e0aaView commit details

  Copy full SHA for 470e0aa

  View code at this pointBrowse repository at this point

• saving preprocessed data by signal group. If all_signals is changed, shots have to be re-processed.

  

  Julian Kates-Harbeck

  committed
  21c88edView commit details

  Copy full SHA for 21c88ed

  View code at this pointBrowse repository at this point

• preprocessing files by which group of all signals was used for downloading. This is relevant because if we change all_signals, the amount of shots that are valid might change. So we can remove rare…

  Show description for 40e64ef
  

  Julian Kates-Harbeck

  committed
  40e64efView commit details

  Copy full SHA for 40e64ef

  View code at this pointBrowse repository at this point