question-mark
Stuck on an issue?

Lightrun Answers was designed to reduce the constant googling that comes with debugging 3rd party libraries. It collects links to all the places you might be looking at while hunting down a tough bug.

And, if you’re still stuck at the end, we’re happy to hop on a call to see how we can help out.

Weights are not logged when model is watched

See original GitHub issue

The gradients of the model are indeed logged but not the parameters.

Using: pytorch-lightning==1.0.4 torch==1.7.0 torchvision==0.8.1 wandb==0.10.9

mac os x 10.15.7

        logger = pl.loggers.WandbLogger(project=hparams.project_name, save_dir=save_dir)
        logger.log_hyperparams(vars(hparams))
        logger.watch(model, log='all', log_freq=1)

Issue Analytics

  • State:closed
  • Created 3 years ago
  • Reactions:1
  • Comments:8 (2 by maintainers)

github_iconTop GitHub Comments

4reactions
sboscommented, Mar 25, 2022

I’m experiencing the same problem. Gradients are logged fine, but parameters don’t show up in wandb panels. Is there any way to tell why this is silently failing?

0reactions
stathiuscommented, Dec 8, 2020

Hi @ariG23498 not really but it’s fine to close it.

Read more comments on GitHub >

github_iconTop Results From Across the Web

Weights are not logged when model is watched #1471 - GitHub
I believe by default it is supposed to log only gradients, but you can change it by using argument log. Please see the...
Read more >
Logging with Weights & Biases - Towards Data Science
Logging weight histograms is as simple as calling wandb.watch() and passing the network object as the argument.
Read more >
When is one supposed to run wandb.watch so that weights ...
wandb.watch will only start working once you call wandb.log after a backwards pass that touches the watched Module (docs).
Read more >
How to Use Weights & Biases with MMDetection - Wandb
Our goal today is simple: to use MMDetection to train an object detection model and see how Weights & Biases helps you log...
Read more >
Log (Almost) Anything with Weights & Biases - YouTube
In this video, Weights & Biases Deep Learning Educator Charles Frye demonstrates how to log rich media -- charts, videos, point clouds, ...
Read more >

github_iconTop Related Medium Post

No results found

github_iconTop Related StackOverflow Question

No results found

github_iconTroubleshoot Live Code

Lightrun enables developers to add logs, metrics and snapshots to live code - no restarts or redeploys required.
Start Free

github_iconTop Related Reddit Thread

No results found

github_iconTop Related Hackernoon Post

No results found

github_iconTop Related Tweet

No results found

github_iconTop Related Dev.to Post

No results found

github_iconTop Related Hashnode Post

No results found