ENH: Dropout in transformer layers in DeepMIL #590

harshita-s · 2022-09-02T13:17:04Z

In this PR:

transformer_dropout parameter is added to TransformerPooling and TransformerPoolingBenchmark pooling layers.
Tests are updated with the transformer_dropout parameter.

codecov · 2022-09-02T13:25:37Z

Codecov Report

Merging #590 (bc096c4) into transfer_main (763dc7f) will increase coverage by 0.20%.
The diff coverage is 85.71%.

Flags with carried forward coverage won't be shown. Click here to find out more.

Impacted Files	Coverage Δ
...i-ml-cpath/src/health_cpath/utils/deepmil_utils.py	`59.70% <66.66%> (+0.61%)`	⬆️
.../src/health_ml/networks/layers/attention_layers.py	`96.33% <100.00%> (+14.08%)`	⬆️

hi-ml-cpath/src/health_cpath/utils/deepmil_utils.py

hi-ml/testhiml/testhiml/test_attentionlayers.py

Harshita Sharma added 3 commits September 2, 2022 13:05

add dropout to transformer pooling layers

dd0ee28

make default values consistent

a0802f7

update defaults

7acee21

Harshita Sharma added 2 commits September 2, 2022 13:35

add test for transformerpoolingbenchmark

1e32f81

add condition

ad37732

harshita-s marked this pull request as ready for review September 2, 2022 14:36

harshita-s requested a review from dccastro September 2, 2022 14:36

dccastro approved these changes Sep 2, 2022

View reviewed changes

Harshita Sharma added 3 commits September 5, 2022 07:57

PR comment kwargs

fb39f7c

PR comment insinstance and separate transformer pooling tests

5adc1b9

PR comment transformer dropout

bc096c4

harshita-s merged commit f7c21a6 into transfer_main Sep 5, 2022

harshita-s deleted the hsharma/transformer_dropout branch September 5, 2022 10:20