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.

Sign up for GitHub

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

Jump to bottom

fix log epoch #1986

Merged
merged 8 commits into from
Feb 22, 2023
Merged

fix log epoch #1986

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
8 commits
Select commit Hold shift + click to select a range
9bba16a
fix logs
mvpatel2000 Feb 21, 2023
9a10644
Merge branch 'dev' into mvpatel2000/logger-tweaks
mvpatel2000 Feb 21, 2023
63ba44c
fix tests
mvpatel2000 Feb 21, 2023
c376e2c
Merge branch 'mvpatel2000/logger-tweaks' of github.com:mvpatel2000/co…
mvpatel2000 Feb 21, 2023
a85dcab
remove dead loggers
mvpatel2000 Feb 21, 2023
ffff010
fix tests
mvpatel2000 Feb 21, 2023
f1f031c
add prints
mvpatel2000 Feb 22, 2023
e046118
fix logs
mvpatel2000 Feb 22, 2023
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
5 changes: 1 addition & 4 deletions composer/trainer/trainer.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1921,7 +1921,7 @@ def _train_loop(self) -> None:
try:
if int(self.state.timestamp.batch_in_epoch) == 0:
self.engine.run_event(Event.EPOCH_START)
self.logger.log_metrics({'epoch': int(self.state.timestamp.epoch)})
self.logger.log_metrics({'trainer/epoch': int(self.state.timestamp.epoch)})

dataloader = self.state.dataloader
if isinstance(dataloader, DataLoader) and isinstance(dataloader.sampler, DistributedSampler):
Expand Down Expand Up @@ -2760,9 +2760,6 @@ def _eval_loop(

self.engine.run_event(Event.EVAL_BATCH_END)

self.logger.log_metrics({'epoch': self.state.timestamp.epoch.value})
self.logger.log_metrics({'trainer/global_step': self.state.timestamp.batch.value})

self._compute_and_log_metrics(dataloader_label=dataloader_label, metrics=metrics)

self.engine.run_event(Event.EVAL_END)
Expand Down
3 changes: 2 additions & 1 deletion tests/loggers/test_cometml_logger.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -194,7 +194,8 @@ def test_comet_ml_logging_train_loop(monkeypatch, tmp_path):

# Check that basic metrics appear in the comet logs
assert len([
metric_msg for metric_msg in msg_type_to_msgs['metric_msg'] if metric_msg['metric']['metricName'] == 'epoch'
metric_msg for metric_msg in msg_type_to_msgs['metric_msg']
if metric_msg['metric']['metricName'] == 'trainer/epoch'
]) == 2

# Check that basic params appear in the comet logs
Expand Down
3 changes: 1 addition & 2 deletions tests/metrics/test_current_metrics.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -110,13 +110,12 @@ def test_current_metrics(eval_interval: str,):
num_expected_calls += (train_subset_num_batches + 1) * num_epochs
# computed at eval end
if compute_val_metrics:
num_calls_per_eval = 3 # metrics + epoch + trainer/global_step
num_evals = 0
if eval_interval == '1ba':
num_evals += train_subset_num_batches * num_epochs
if eval_interval == '1ep':
num_evals += num_epochs
num_expected_calls += (num_calls_per_eval) * num_evals
num_expected_calls += num_evals

num_actual_calls = len(mock_logger_destination.log_metrics.mock_calls)

Expand Down