Release dev2 #141

data-hound · 2023-03-03T00:52:47Z

New release for DANNCE

New Features

Mirror augmentation added
- Mirror the images in half of the samples of the training set
- If mirror augmentation is used, the right_keypoints indices and left_keypoints indices must be specified as well.
  For the skeleton, ['RHand, 'LHand', 'RFoot', 'LFoot'],
  set right_keypoints: [0, 2] and
  left_keypoints: [1, 3] in the config file
Options for layer and instance normalization added.
More Losses available: Huber, logcosh, gaussian cross entropy
n-rand-views configured to choose a subset of camera views for training/testing on weights trained for a lesser number of views.
(Note: the number of views for which the weights have been trained should match the number of views specified with n-rand-views)
AVG+MAX training introduced where both the mean of heatmaps and the pmax of heatmaps are used for loss computation.
Saving prediction targets when running in MAX mode enabled

Codebase Improvements

Cluster scripts have now been remodelled to use yaml config files specific to each cluster instead of a set of shell scripts
- Sbatch submissions for dannce and com training and prediction have been implemented.
Logging statements added throughout the code base with options to change the log level in the config (closes Logging code type error #117, closes dannce-predict loss very small, but result same like normal but shift #138)
DANNCE now supports mat73 format files for label3d and predictions
Python based scripts for plotting predictions on videos added
Newer tests using function name parsing
Uniform nomenclature for results_dir across codebase
Callbacks moved to a new module
Code refactored and reformatted across the codebase
Trimming dependencies in multiple modules
Saving pred targets moved to end of training

Bug Fixes and Minor Enhancements

Allow multiple read attempts when loading video
Added sigmoid normalization for pmax in dannce-predict
Introduced closing videos from prev experiments
Remove GPU placement during inference
Model summary printed during training to help debug mismatches
Save training data into npy volumes and cleanup unused
Remove hardcoded file names in multi-gpu

…on in the configs.

…g in MAX mode

… networks from net.py

…d processing.py -adding backwards compatibility and defaulting to layer norm

updating jhw_dev with dev_fix_1.2 updates

This reverts commit 9058136. TF 2.6 doesn't work with built in cuda/cudnn installations on the cluster that we've been using. Reverting to keep tf2.4 branch using tf2.4 requirements.

Fixes #86

…to new cluster requirements

…e_dev_savedata Merge small changes to remote development into remote savedata branch

…nnce into release_development

…sso/dannce into release_development

…ease_development

…in interface to save prediction targets only when the corresponding flag is set to true in config.

… files to support matlab 7.3 files

…o release_dev2

…ving

spoonsso · 2023-05-03T15:04:14Z

@data-hound can you please test out the README install instructions in a fresh conda environment and then run all tests? thanks!

data-hound · 2023-05-09T21:32:15Z

Done. The install instructions work and all tests under /tests are passing

@data-hound can you please test out the README install instructions in a fresh conda environment and then run all tests? thanks!

diegoaldarondo and others added 30 commits July 16, 2021 15:31

Make scripts work with tf 2.4

0693770

Add mirror augmentation

a96ea18

Use Session instead of interactive session to execute in eager mode

d1b4130

formatting

51721aa

formatting

26581ad

Add multiple read attempts

b3ad8be

formatting

4091152

Add euclidean distance metric to logging

d99ebde

Add ability to specify left and right keypoints for mirror augmentati…

23367ae

…on in the configs.

formatting and modified comment

4737f7d

Load modules explicitly instead of through user scripts

125f17c

formatting

e4b77ae

Add psutil requirement to help callback memory leak debugging.

26c0196

Fix bug in com multipredict script

9f023b2

fix video loading bugs

b1b75c5

Added saveMaxPreds callback that enables distance metrics when runnin…

3f1d15b

…g in MAX mode

changing cluster scripts to work on DCC

df3eac1

added support for avg+max training and prediction; removed deprecated…

3b42606

… networks from net.py

testing instance normalization in nets.py

7d38596

added instance normalization to nets.py

e922c07

sigmoid normalization for p_max in dannce-predict

ff61469

fixed com_debug bug

0393bd3

integrating norm_method across nets.py, interface.py, __init__.py, an…

8b54c3b

…d processing.py -adding backwards compatibility and defaulting to layer norm

updated readme and setup.py to upgrade to tf2.6

9058136

testing cross entropy loss on max network

e2c43e1

fix video loading bugs

4a15f66

Local Changes to cluster scripts for making tf2.4 dcc compatible

a5ef7b7

fixed merge conflicts with tf2.4

2a16c03

Merge branch 'dev_fix_1.2' into jhw_dev

6e53d7a

updating jhw_dev with dev_fix_1.2 updates

Revert "updated readme and setup.py to upgrade to tf2.6"

c4a0588

This reverts commit 9058136. TF 2.6 doesn't work with built in cuda/cudnn installations on the cluster that we've been using. Reverting to keep tf2.4 branch using tf2.4 requirements.

Anshuman Sabath and others added 24 commits May 12, 2022 17:42

Changing to tf2.6

0dcdf71

Merge branch 'release' into as_logging

984842d

Merge pull request #108 from spoonsso/as_logging

9bcb3e3

Fixes #86

formatting, change n retries in multi inference, change holyoke.yaml …

136db21

…to new cluster requirements

fix bug in multimode n instances

e3936b7

Merge remote-tracking branch 'origin/release_development' into releas…

4244b5b

…e_dev_savedata Merge small changes to remote development into remote savedata branch

multi_gpu fixed hard coded io.yaml

a361e90

Fixing type errors in logging statements

d4f3781

Update README.md

3858380

Remove Intermediate Supervision Code

eed6621

Pulled in release_minorfixes

c32beec

Pulled in release branch to incorporate all the logging statements

cb70aeb

Merged and Tested release and release_minorfixes

ed70f14

Merge branch 'release_dev_savedata' of https://github.com/spoonsso/da…

1b65a7d

…nnce into release_development

Merge branch 'release_develpment_feature' of https://github.com/spoon…

558966d

…sso/dannce into release_development

Change merged print statement

96aa81b

Merge branch 'pdb_viz' of https://github.com/spoonsso/dannce into rel…

af8a161

…ease_development

Added Gaussian Cross Entropy Loss to custom objects. Added condition …

201e64b

…in interface to save prediction targets only when the corresponding flag is set to true in config.

Added new tests and ran the full test suite

b13cecb

Added new tests and ran the full test suite. Migrated the io for .mat…

f20cdf6

… files to support matlab 7.3 files

Merge branch 'release_dev2' of https://github.com/spoonsso/dannce int…

a242cb5

…o release_dev2

Update Readme

7d54aaf

mat73 integration all throughout. COM-addition to MAX preds before sa…

87fd513

…ving

Changing touchstone for max predictions

16aecc9

data-hound requested review from spoonsso and diegoaldarondo March 3, 2023 21:03

data-hound marked this pull request as ready for review March 15, 2023 23:42

Removing symlinks

9fc0ba4

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Release dev2 #141

Release dev2 #141

data-hound commented Mar 3, 2023 •

edited

Loading

spoonsso commented May 3, 2023

data-hound commented May 9, 2023

Release dev2 #141

Are you sure you want to change the base?

Release dev2 #141

Conversation

data-hound commented Mar 3, 2023 • edited Loading

New Features

Codebase Improvements

Bug Fixes and Minor Enhancements

spoonsso commented May 3, 2023

data-hound commented May 9, 2023

data-hound commented Mar 3, 2023 •

edited

Loading