Best practices to organize hyperparameter searches
When you run many hyperparameter searches, tag your runs so you can group and find related experiments later. Tag runs i โฆ
Can you group runs by tags?
A run can have multiple tags, so grouping by tags isnโt supported. Instead, add a value to the config object for these r โฆ
How can I categorize runs without the Group feature?
You can categorize runs by tags or custom metadata. To organize runs by these values, use the Group button in the projec โฆ
How can I change how frequently to log system metrics?
To configure the frequency to log system metrics, set _stats_sampling_interval to a number of seconds, expressed as a fl โฆ
How can I change the colors of each run in the same group?
You canโt change the colors of individual runs within a group. All runs in the same group share a common color.
How can I delete multiple runs in bulk instead of one at a time?
From the Runs tab in your Workspace, select the checkbox next to each run to delete, then click Delete. To filter runs b โฆ
How can I disable logging of system metrics to W&B?
To disable logging of system metrics, set _disable_stats to True:
How can I log a metric that doesn't change over time such as a final evaluation accuracy?
To record a single, unchanging value such as a final evaluation accuracy, log it once with run.log. For example, run.log โฆ
How can I log additional metrics after a run completes?
To add metrics to a run that has already finished, W&B offers several approaches depending on the complexity of your wor โฆ
How can I recover deleted runs?
Learn how to recover recently deleted runs in the W&B App for up to 7 days.
How do I change the group of a completed run?
You can change the group assigned to a completed run using the API. This feature doesnโt appear in the web UI. To update โฆ
How do I download the console log file from a run?
W&B stores your scriptโs stdout and stderr as output.log (or multipart chunks under logs/). Where to retrieve it depends โฆ
How do I fix `CommError, Run does not exist` during a sweep?
If you see both CommError, Run does not exist and ERROR Error uploading during a sweep, the most likely cause is that yo โฆ
How do I fix the error `resume='must' but run ([RUN-ID]) doesn't exist`?
If you encounter the error resume=โmustโ but run ( RUN-ID ) doesnโt exist, the run youโre attempting to resume doesnโt e โฆ
How do I log NLP metrics and text outputs in W&B?
You can log corpus-level NLP scores (BLEU, ROUGE, perplexity) with wandb.log() and per-example outputs with wandb.Table. โฆ
How do I log gradients and model weights with wandb.watch()?
wandb.watch() hooks into a PyTorch modelโs parameters and gradients and logs histograms of their values at regular inter โฆ
How do I log runs launched by continuous integration or internal tools?
To launch automated tests or internal tools that log to W&B, create a Service Account on the team settings page. This ac โฆ
How do I move a run to another project?
You can move a run from one project to another to reorganize your work, consolidate related experiments, or relocate a r โฆ
How do I page through large API results in W&B?
You can page through API result using the standard lazy-iterator pattern and per_page parameter. Additionally, you can u โฆ
How do I rerun a grid search?
If a grid search completes but some W&B Runs crashed, delete those runs to rerun them. Then select the Resume button on โฆ
How do I resolve permission errors when logging a run?
Permission errors when logging a run typically indicate a mismatch between your code, your credentials, or your deployme โฆ
How do I set up W&B alerts and notifications?
You can set up alerts in and notifications using the W&B Settings page. W&B supports automated alerts in User Settings ( โฆ
How do I update run config, tags, and notes via the W&B API?
After a run finishes, use the Public API guide to edit config, display name, tags, and notes without re-running the expe โฆ
How do I use W&B with JAX?
W&B has no JAX-specific integration. However, you can convert JAX device arrays to Python scalars and then use wandb.log โฆ
How do I use the parallel coordinates chart in W&B?
The parallel coordinates chart shows how hyperparameters relate to metrics across many runs. Each line is one run; each โฆ
How many runs can I create per project?
For optimal performance, limit each project to approximately 10,000 runs.
Why are steps missing from a CSV metric export?
Export limits can prevent you from exporting the entire run history as a CSV or with the run.history API. To access the โฆ
Why do my workspace settings not persist between sessions?
Workspace layout (panels, filters, grouping) persists only when you save a view. Unsaved changes stay in your browser se โฆ
Why is console output not captured for my run?
W&B captures your scriptโs stdout and stderr and stores it as output.log on the runโs Files tab. By default, that file u โฆ
Why is my W&B run slow to initialize or upload?
Slow wandb.init() or sluggish metric uploads are usually caused by network latency, large media payloads, high logging f โฆ
Why is my run showing as crashed?
W&B marks a run as Crashed when it stops receiving heartbeats from the process that called wandb.init(), without the pro โฆ