-
Notifications
You must be signed in to change notification settings - Fork 52
Add 11 new knowledge base articles for Models and Weave #2431
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.
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
Merged
Merged
Changes from all commits
Commits
Show all changes
3 commits
Select commit
Hold shift + click to select a range
File filter
Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
There are no files selected for viewing
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
38 changes: 38 additions & 0 deletions
38
support/models/articles/anaconda-400-error-during-a-sweep.mdx
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,38 @@ | ||
| --- | ||
| title: "How do I fix an `anaconda 400 error` during a sweep?" | ||
| keywords: ["Sweeps", "Metrics"] | ||
| --- | ||
|
|
||
| The following error usually occurs when you do not log the metric that you are optimizing: | ||
|
|
||
| ``` | ||
| wandb: ERROR Error while calling W&B API: anaconda 400 error: | ||
| {"code": 400, "message": "TypeError: bad operand type for unary -: 'NoneType'"} | ||
| ``` | ||
|
|
||
| To fix this, make sure you are logging the exact metric specified in your sweep configuration. For example, if your sweep config specifies: | ||
|
|
||
| ```yaml | ||
| metric: | ||
| name: validation_loss | ||
| goal: minimize | ||
| ``` | ||
|
|
||
| Then your training code must call: | ||
|
|
||
| ```python | ||
| wandb.log({"validation_loss": value}) | ||
| ``` | ||
|
|
||
| The metric name must match exactly between your sweep configuration and your `wandb.log` call. | ||
|
|
||
| For more information, see [Sweeps troubleshooting](/models/sweeps/troubleshoot-sweeps/) and [Define sweep configuration](/models/sweeps/define-sweep-configuration/). | ||
|
|
||
| --- | ||
|
|
||
| {/* ---- AUTO-GENERATED: tab badges ---- | ||
| Managed by scripts/knowledgebase-nav/generate_tags.py from keywords in front matter. | ||
| Do not edit between these markers by hand. | ||
| ---- */} | ||
| <Badge stroke shape="pill" color="orange" size="md">[Sweeps](/support/models/tags/sweeps)</Badge><Badge stroke shape="pill" color="orange" size="md">[Metrics](/support/models/tags/metrics)</Badge> | ||
| {/* ---- END AUTO-GENERATED: tab badges ---- */} |
39 changes: 39 additions & 0 deletions
39
support/models/articles/cannot-link-artifact-to-registry-from-p.mdx
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,39 @@ | ||
| --- | ||
| title: "Why can't I link my artifact to the registry?" | ||
| keywords: ["Artifacts"] | ||
| --- | ||
|
|
||
| If you cannot link an artifact to a W&B Registry, the most common cause is that the artifact was logged with a personal entity instead of a team entity. Only artifacts logged within an organization's team can be linked to the organization's registry. | ||
|
|
||
| ## Fix: log artifacts with a team entity | ||
|
|
||
| Specify your team as the `entity` when you initialize a run: | ||
|
|
||
| ```python | ||
| import wandb | ||
|
|
||
| with wandb.init(entity="<team_entity>", project="<project_name>") as run: | ||
| artifact = wandb.Artifact(name="<artifact_name>", type="<type>") | ||
| # Add files and log the artifact | ||
| ``` | ||
|
|
||
| If you do not specify `entity`, the run uses your default entity, which may be your personal account. | ||
|
|
||
| ## Find your team entity | ||
|
|
||
| Your team entity is the same as your team name. Confirm it by navigating to your team's W&B profile page. The URL has the form `https://wandb.ai/<team>`, where `<team>` is your team entity. | ||
|
|
||
| ## Already logged with a personal entity? | ||
|
|
||
| If an artifact was already logged to your personal entity, you need to re-log it to a team entity within your organization. | ||
|
|
||
| For more information, see [Link a version to a collection](/models/registry/link_version/). | ||
|
|
||
| --- | ||
|
|
||
| {/* ---- AUTO-GENERATED: tab badges ---- | ||
| Managed by scripts/knowledgebase-nav/generate_tags.py from keywords in front matter. | ||
| Do not edit between these markers by hand. | ||
| ---- */} | ||
| <Badge stroke shape="pill" color="orange" size="md">[Artifacts](/support/models/tags/artifacts)</Badge> | ||
| {/* ---- END AUTO-GENERATED: tab badges ---- */} |
29 changes: 29 additions & 0 deletions
29
support/models/articles/commerror-run-does-not-exist-during-swee.mdx
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,29 @@ | ||
| --- | ||
| title: "How do I fix `CommError, Run does not exist` during a sweep?" | ||
| keywords: ["Sweeps", "Runs"] | ||
| --- | ||
|
|
||
| If you see both `CommError, Run does not exist` and `ERROR Error uploading` during a sweep, the most likely cause is that you are setting a run ID manually in your code: | ||
|
|
||
| ```python | ||
| wandb.init(id="some-string") | ||
| ``` | ||
|
|
||
| You can not set a run ID for W&B Sweeps because W&B automatically generates random, unique IDs for runs created by sweeps. Run IDs must be unique within a project. | ||
|
|
||
| To fix this, remove the `id` parameter from `wandb.init()`. If you want to set a custom name that appears on tables and graphs, use the `name` parameter instead: | ||
|
|
||
| ```python | ||
| wandb.init(name="a helpful readable run name") | ||
| ``` | ||
|
|
||
| For more information, see [Sweeps troubleshooting](/models/sweeps/troubleshoot-sweeps/). | ||
|
|
||
| --- | ||
|
|
||
| {/* ---- AUTO-GENERATED: tab badges ---- | ||
| Managed by scripts/knowledgebase-nav/generate_tags.py from keywords in front matter. | ||
| Do not edit between these markers by hand. | ||
| ---- */} | ||
| <Badge stroke shape="pill" color="orange" size="md">[Sweeps](/support/models/tags/sweeps)</Badge><Badge stroke shape="pill" color="orange" size="md">[Runs](/support/models/tags/runs)</Badge> | ||
| {/* ---- END AUTO-GENERATED: tab badges ---- */} | ||
54 changes: 54 additions & 0 deletions
54
support/models/articles/cuda-out-of-memory-during-a-sweep.mdx
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change | ||||
|---|---|---|---|---|---|---|
| @@ -0,0 +1,54 @@ | ||||||
| --- | ||||||
| title: "How do I fix `Cuda out of memory` during a sweep?" | ||||||
| keywords: ["Sweeps", "Run Crashes"] | ||||||
| --- | ||||||
|
|
||||||
| If you see `Cuda out of memory` during a sweep, refactor your code to use process-based execution. Rewrite your code as a Python script and call the sweep agent from the CLI instead of the Python SDK. | ||||||
|
|
||||||
| 1. Add your training logic to a Python script (for example, `train.py`): | ||||||
|
|
||||||
| ```python | ||||||
| if __name__ == "__main__": | ||||||
| train() | ||||||
| ``` | ||||||
|
|
||||||
| 2. Reference the script in your YAML sweep configuration: | ||||||
|
|
||||||
| ```yaml | ||||||
| program: train.py | ||||||
| method: bayes | ||||||
| metric: | ||||||
| name: validation_loss | ||||||
| goal: maximize | ||||||
|
||||||
| goal: maximize | |
| goal: minimize |
38 changes: 38 additions & 0 deletions
38
support/models/articles/enterprise-license-not-recognized.mdx
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,38 @@ | ||
| --- | ||
| title: "Why is my enterprise license not recognized?" | ||
| keywords: ["Administrator"] | ||
| --- | ||
|
|
||
| If your enterprise license is not recognized or features are not available after setting the license, check the following: | ||
|
|
||
| ### License not recognized | ||
|
|
||
| - Verify the license key is correctly formatted with no extra spaces or characters. | ||
| - Ensure the license has not expired. | ||
| - Check that the license is set in the correct configuration location. | ||
|
|
||
| ### Features not available after setting the license | ||
|
|
||
| - Restart your W&B services after setting the license. | ||
| - Verify the license includes the specific features you are trying to access. | ||
| - Check system logs for any license validation errors. | ||
|
|
||
| ### License expiration warnings | ||
|
|
||
| - Monitor the System Settings page for expiration notifications. | ||
| - Set up alerts for license expiration in your monitoring system. | ||
|
|
||
| ### Get support | ||
|
|
||
| For assistance with enterprise licenses, contact your account team, [Sales](mailto:sales@wandb.ai), or [Support](mailto:support@wandb.ai). | ||
|
|
||
| For more information, see [Enterprise licenses](/platform/hosting/enterprise-licenses). | ||
|
|
||
| --- | ||
|
|
||
| {/* ---- AUTO-GENERATED: tab badges ---- | ||
| Managed by scripts/knowledgebase-nav/generate_tags.py from keywords in front matter. | ||
| Do not edit between these markers by hand. | ||
| ---- */} | ||
| <Badge stroke shape="pill" color="orange" size="md">[Administrator](/support/models/tags/administrator)</Badge> | ||
| {/* ---- END AUTO-GENERATED: tab badges ---- */} |
31 changes: 31 additions & 0 deletions
31
support/models/articles/process-hangs-when-using-hydra-with-wan.mdx
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,31 @@ | ||
| --- | ||
| title: "Why does my process hang when using Hydra with W&B?" | ||
| keywords: ["Experiments", "Run Crashes"] | ||
| --- | ||
|
|
||
| If your process hangs when started with Hydra, this is likely caused by a multiprocessing conflict between Hydra and W&B. | ||
|
|
||
| To fix this, change W&B's multiprocessing protocol to `"thread"`. You can do this in one of two ways: | ||
|
|
||
| **Option 1**: Pass a settings parameter to `wandb.init()`: | ||
|
|
||
| ```python | ||
| wandb.init(settings=wandb.Settings(start_method="thread")) | ||
| ``` | ||
|
|
||
| **Option 2**: Set a global environment variable: | ||
|
|
||
| ```bash | ||
| export WANDB_START_METHOD=thread | ||
| ``` | ||
|
|
||
| For more information, see the [Hydra integration guide](/models/integrations/hydra/) and [Distributed training troubleshooting](/models/track/log/distributed-training/). | ||
|
|
||
| --- | ||
|
|
||
| {/* ---- AUTO-GENERATED: tab badges ---- | ||
| Managed by scripts/knowledgebase-nav/generate_tags.py from keywords in front matter. | ||
| Do not edit between these markers by hand. | ||
| ---- */} | ||
| <Badge stroke shape="pill" color="orange" size="md">[Experiments](/support/models/tags/experiments)</Badge><Badge stroke shape="pill" color="orange" size="md">[Run Crashes](/support/models/tags/run-crashes)</Badge> | ||
| {/* ---- END AUTO-GENERATED: tab badges ---- */} |
50 changes: 50 additions & 0 deletions
50
support/models/articles/rate-limit-exceeded-on-metric-logging.mdx
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,50 @@ | ||
| --- | ||
| title: "How do I fix `Rate limit exceeded` errors when logging metrics?" | ||
| keywords: ["Metrics", "Experiments"] | ||
| --- | ||
|
|
||
| If you receive an HTTP `429 Rate limit exceeded` error when calling `wandb.log()`, you are exceeding the rate limit quota for your project. W&B applies rate limits per project, and paid plans have higher limits than free plans. | ||
|
|
||
| ## How to stay under the rate limit | ||
|
|
||
| 1. **Update your W&B SDK**: The latest version includes optimized mechanisms for retrying requests and managing quota usage. | ||
|
|
||
| ```shell | ||
| pip install --upgrade wandb | ||
| ``` | ||
|
|
||
| 2. **Reduce logging frequency**: Log metrics less often. For example, log every N epochs instead of every epoch: | ||
|
|
||
| ```python | ||
| for epoch in range(100): | ||
| if epoch % 5 == 0: | ||
| wandb.log({"acc": accuracy, "loss": loss}) | ||
| ``` | ||
|
|
||
| 3. **Sync data manually**: If you are rate limited, W&B stores your run data locally. You can sync it later with: | ||
|
|
||
| ```shell | ||
| wandb sync <run-file-path> | ||
| ``` | ||
|
|
||
| ## Rate limit headers | ||
|
|
||
| When rate limited, the response includes these headers: | ||
|
|
||
| | Header | Description | | ||
| | --- | --- | | ||
| | `RateLimit-Remaining` | Quota remaining in the current window (0–1000 scale) | | ||
| | `RateLimit-Reset` | Seconds until the current quota resets | | ||
|
|
||
| If `RateLimit-Remaining` is `0`, wait for the number of seconds in `RateLimit-Reset` before retrying. | ||
|
|
||
| For more information, see [Experiments limits and performance](/models/track/limits/). | ||
|
|
||
| --- | ||
|
|
||
| {/* ---- AUTO-GENERATED: tab badges ---- | ||
| Managed by scripts/knowledgebase-nav/generate_tags.py from keywords in front matter. | ||
| Do not edit between these markers by hand. | ||
| ---- */} | ||
| <Badge stroke shape="pill" color="orange" size="md">[Metrics](/support/models/tags/metrics)</Badge><Badge stroke shape="pill" color="orange" size="md">[Experiments](/support/models/tags/experiments)</Badge> | ||
| {/* ---- END AUTO-GENERATED: tab badges ---- */} |
Oops, something went wrong.
Oops, something went wrong.
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Minor grammar: "can not" should be "cannot" here.