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

Use eval_batch_size for AutoEncoder loss stats #861

Merged
merged 8 commits into from
Apr 17, 2023
Merged

Use eval_batch_size for AutoEncoder loss stats #861

Changes from all commits
Commits
Show all changes
8 commits
Select commit Hold shift + click to select a range
55838d2
fix string formatting in debug lines
efajardo-nv Apr 10, 2023
ab01bfc
use eval_batch_size for loss stats
efajardo-nv Apr 10, 2023
6bc5b88
Merge branch 'branch-23.07' into use-eval-batch-for-loss-stats
efajardo-nv Apr 10, 2023
7925010
Merge branch 'branch-23.07' into use-eval-batch-for-loss-stats
efajardo-nv Apr 13, 2023
c64e478
use lazy formatting for logger messages
efajardo-nv Apr 14, 2023
857a3d0
Merge branch 'branch-23.07' of https://github.com/nv-morpheus/Morpheu…
efajardo-nv Apr 14, 2023
f96b9dd
style fix
efajardo-nv Apr 14, 2023
774f2bf
Merge branch 'use-eval-batch-for-loss-stats' of https://github.com/ef…
efajardo-nv Apr 14, 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
15 changes: 7 additions & 8 deletions morpheus/models/dfencoder/autoencoder.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -905,15 +905,14 @@ def _fit_centralized(self, df, epochs=1, val=None, run_validation=False, use_val

# Early stopping
current_net_loss = net_loss
LOG.debug('The Current Net Loss:', current_net_loss)
LOG.debug('The Current Net Loss: %s', current_net_loss)

if current_net_loss > last_loss:
count_es += 1
LOG.debug('Early stop count:', count_es)
LOG.debug('Early stop count: %s', count_es)

if count_es >= self.patience:
LOG.debug('Early stopping: early stop count({}) >= patience({})'.format(
count_es, self.patience))
LOG.debug('Early stopping: early stop count(%s) >= patience(%s)', count_es, self.patience)
break

else:
Expand Down Expand Up @@ -1617,15 +1616,15 @@ def get_anomaly_score_losses(self, df):
"""
self.eval()

n_batches = len(df) // self.batch_size
if len(df) % self.batch_size > 0:
n_batches = len(df) // self.eval_batch_size
if len(df) % self.eval_batch_size > 0:
n_batches += 1

mse_loss_slices, bce_loss_slices, cce_loss_slices = [], [], []
with torch.no_grad():
for i in range(n_batches):
start = i * self.batch_size
stop = (i + 1) * self.batch_size
start = i * self.eval_batch_size
stop = (i + 1) * self.eval_batch_size

df_slice = df.iloc[start:stop]
data_slice = self.prepare_df(df_slice)
Expand Down