[Simple question] Any available logger for test or validation steps? @Logger availability #10862
Unanswered
Hikaru-13915
asked this question in
code help: CV
Replies: 1 comment
-
I think I can reproduce it in Kaggle notebook with |
Beta Was this translation helpful? Give feedback.
0 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
-
As all of you knows, we can use Loggers by setting like
from pytorch_lightning.loggers import 'YOURLOGGER'
trainer = model.Trainer(logger=YOURLOGGER(save_dir="logs/", name="your_train"))
Now I'm trying to use logger, CSVLogger, but it only tells me the information as to the training steps like bellow.
train_loss train_acc epoch step valid_loss valid_acc test_loss test_acc
0 0.626578 0.875000 0 49 NaN NaN NaN NaN
1 0.514625 0.875000 0 99 NaN NaN NaN NaN
2 0.334852 0.906250 0 149 NaN NaN NaN NaN
3 0.121677 0.953125 0 199 NaN NaN NaN NaN
4 0.135196 0.968750 0 249 NaN NaN NaN NaN
The example above doesn't reveal "valid_loss" and "valid_acc", "test_loss", "test_acc", though I defined these values at their own steps in the model.
Is there any possible way to log a test or validation steps??
Please teach me any available Loggers for test steps and how to implement that
Beta Was this translation helpful? Give feedback.
All reactions