Commits

Branch selector

 master

User selector

All users

Datepicker

All time

Commit History

Commits on Jul 14, 2021

• Add files via upload

  ge-dong

  authored
  101d810

  Copy full SHA for 101d810

Commits on Jul 13, 2020

• Add files via upload

  ge-dong

  authored
  a86db90

  Copy full SHA for a86db90

Commits on Dec 19, 2019

• Merge branch 'master' of github.com:ge-dong/plasma-python

  

  Ge Dong

  committed
  d0357fc

  Copy full SHA for d0357fc

• new file: plasma/models/tcn.py

  

  Ge Dong

  committed
  aa937b3

  Copy full SHA for aa937b3

Commits on Dec 10, 2019

• Merge branch 'master' into master

  felker

  authored
  f70440c

  Copy full SHA for f70440c

Commits on Nov 20, 2019

• modified: data/signals.py

  Show description for 35806ec
  

  Ge Dong

  committed
  35806ec

  Copy full SHA for 35806ec

Commits on Nov 13, 2019

• Restrict Conda TensorFlow dependency to <2.0.0

  Show description for 9935ff7
  felker

  committed
  9935ff7

  Copy full SHA for 9935ff7

• new file: plasma/models/torch_runner_dist.py

  

  Ge Dong

  committed
  a26466b

  Copy full SHA for a26466b

• modified: plasma/models/mpi_runner.py

  

  Ge Dong

  committed
  176ea6b

  Copy full SHA for 176ea6b

Commits on Nov 12, 2019

• modified: plasma/conf_parser.py

  Show description for 299674b
  

  Ge Dong

  committed
  299674b

  Copy full SHA for 299674b

Commits on Nov 7, 2019

• Merge pull request #40 from PPPLDeepLearning/hotfix/warnings

  Show description for 2150150
  felker

  authored
  2150150

  Copy full SHA for 2150150

• Use 1-based indexing for epochs

  felker

  committed
  df88d1a

  Copy full SHA for df88d1a

• Make it explicit that diagnostic within train_epoch() refers to local progress

  felker

  committed
  b036c0b

  Copy full SHA for b036c0b

• Fix and improve diagnostics within main mpi_train() loop over epochs

  Show description for 9000727
  felker

  committed
  9000727

  Copy full SHA for 9000727

• Use write_all() for ModelBuilder.load_model_weights() diagnostics

  Show description for 4b549ef
  felker

  committed
  4b549ef

  Copy full SHA for 4b549ef

• Make write_all() safe for use in non-MPI contexts

  felker

  committed
  425ff0b

  Copy full SHA for 425ff0b

• Share TF version as a global variable

  Show description for af09d4f
  felker

  committed
  af09d4f

  Copy full SHA for af09d4f

• Suppress most TensorFlow deprecation warnings with external Keras

  felker

  committed
  aa2fd04

  Copy full SHA for aa2fd04

Commits on Nov 6, 2019

• Workaround for "WARNING: Logging before flag parsing goes to stderr."

  Show description for 66e1c5c
  felker

  committed
  66e1c5c

  Copy full SHA for 66e1c5c

• Completely sort out order of initial Keras and/or TensorFlow init msgs

  Show description for c5dfa9d
  felker

  committed
  c5dfa9d

  Copy full SHA for c5dfa9d

• Add fn for flushing BOTH stderr and stdout, in MPI rank order

  felker

  committed
  5f0cb09

  Copy full SHA for 5f0cb09

• Fix bug in output--- was passing part of printed obj as "end"

  felker

  committed
  b4bc1d1

  Copy full SHA for b4bc1d1

• Cleanup setup.cfg comments

  felker

  committed
  fec3e5a

  Copy full SHA for fec3e5a

• Only specify filename, not relative path, when using per-file-ignores

  Show description for 676621b
  felker

  committed
  676621b

  Copy full SHA for 676621b

• Call new MPI init fn in mpi_learn.py; call print_unique() about signals

  felker

  committed
  2c86e39

  Copy full SHA for 2c86e39

• Conditionally import deprecated TensorFlow V1 APIs

  Show description for 64bfa8a
  felker

  committed
  64bfa8a

  Copy full SHA for 64bfa8a

• Suppress duplicate output from 2x calls to Normalizer.train()

  Show description for 2fc8089
  felker

  committed
  2fc8089

  Copy full SHA for 2fc8089

• Always set recompute_normalization to False inside Normalizer after compute

  felker

  committed
  706f44d

  Copy full SHA for 706f44d

Commits on Nov 5, 2019

• Try ensuring that computing normalizer is only performed by master rank

  felker

  committed
  a571715

  Copy full SHA for a571715

• Shorten line

  felker

  committed
  409d5d5

  Copy full SHA for 409d5d5

• Suppress duplicated stdout output from normalize.py

  felker

  committed
  df41970

  Copy full SHA for df41970

• Suppress duplicated stdout output from guarantee_preprocessed()

  felker

  committed
  1680750

  Copy full SHA for 1680750

• Forgot to pass conf to MPI init fn

  felker

  committed
  8a30ddd

  Copy full SHA for 8a30ddd

• Extract all MPI variables to separate, global/shared module

  Show description for 39709fd
  felker

  committed
  39709fd

  Copy full SHA for 39709fd

• Prepare to call print_unique() or analog in normalize.py

  felker

  committed
  55639ce

  Copy full SHA for 55639ce

Pagination

Previous

Next