Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

attempting to remove some speed issues #1482

Merged
merged 7 commits into from
Apr 15, 2020
Merged

attempting to remove some speed issues #1482

merged 7 commits into from
Apr 15, 2020

Conversation

williamFalcon
Copy link
Contributor

trying to match exact PT speed

@mergify mergify bot requested a review from a team April 14, 2020 00:10
@codecov
Copy link

codecov bot commented Apr 14, 2020

Codecov Report

Merging #1482 into master will decrease coverage by 0%.
The diff coverage is n/a.

@@          Coverage Diff           @@
##           master   #1482   +/-   ##
======================================
- Coverage      91%     91%   -0%     
======================================
  Files          67      67           
  Lines        3745    3742    -3     
======================================
- Hits         3403    3400    -3     
  Misses        342     342           

assert_speed_parity(pl_times, pl_times, num_epochs)


def assert_speed_parity(pl_times, pt_times, num_epochs):
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

do we really need this two times? Can't we just import the other one?

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

fair, i was just trying to find the threshold number before finalizing pr. will refactor this

@@ -112,10 +112,6 @@ def process_output(self, output, train=False):
num_gpus = self.num_gpus
callback_metrics = self.reduce_distributed_output(callback_metrics, num_gpus)

for k, v in callback_metrics.items():
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

why remove this?

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

it’s done twice. this is the first time but it’s done again at the end of the method

@mergify mergify bot requested a review from a team April 14, 2020 08:26
@Borda Borda added the feature Is an improvement or enhancement label Apr 14, 2020
benchmarks/test_rnn_parity.py Outdated Show resolved Hide resolved
benchmarks/test_trainer_parity.py Outdated Show resolved Hide resolved
@mergify mergify bot requested a review from a team April 14, 2020 09:00
williamFalcon and others added 4 commits April 14, 2020 07:41
Co-Authored-By: Jirka Borovec <Borda@users.noreply.github.com>
Co-Authored-By: Jirka Borovec <Borda@users.noreply.github.com>
@williamFalcon williamFalcon merged commit c96c6a6 into master Apr 15, 2020
@Borda Borda deleted the speed branch April 15, 2020 06:22
tullie pushed a commit to tullie/pytorch-lightning that referenced this pull request Jun 7, 2020
* removed some .items

* added speed tests

* added speed tests

* Update benchmarks/test_rnn_parity.py

Co-Authored-By: Jirka Borovec <Borda@users.noreply.github.com>

* Update benchmarks/test_trainer_parity.py

Co-Authored-By: Jirka Borovec <Borda@users.noreply.github.com>

* fix lost model reference

* added speed tests

Co-authored-by: Jirka Borovec <Borda@users.noreply.github.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
feature Is an improvement or enhancement
Projects
None yet
Development

Successfully merging this pull request may close these issues.

3 participants