Skip to main content
Use W&B to save your code, compare code across runs, view local changes, and capture Jupyter session history.

​
Enable code saving

Configure code saving for your team or organization. Team and organization controls are documented in Configure privacy settings.
By default, W&B disables code saving for all teams. Before you can turn it on for a team, an organization admin must enable it at the organization level. See the Organization section.

​
Team

To enable code saving for a team as a team admin, navigate to the Settings page, then go to the Privacy section, and configure Enable code saving by default for runs in that team. This option is available only when an organization admin has not enforced code saving restrictions for the whole organization. For navigation steps, see Configure privacy settings for a team.

​
Organization

To enable code saving for an organization as an organization admin, navigate to Settings, go to the Privacy section, and turn on Enforce default code saving restrictions so code saving stays off by default for every team. While this enforcement is on, team admins cannot turn on Enable code saving by default for a team. For the full list of organization controls, see Enforce privacy settings for all teams.

​
Save code

Capture source code that produced a run as an artifact. The code artifact is viewable within your project’s workspace. There are two ways to save code: a fine-grained approach and an automatic approach.

​
Fine grain control

Use wandb.Run.log_code() to log specific files or directories as an artifact. By default, W&B walks through the current directory and logs all files that end with .py. The following example shows how to use wandb.Run.log_code() to log the current directory: 
import wandb

with wandb.init() as run:
    # Log the current directory as a code artifact
    run.log_code(root=".")
The following example shows how to use wandb.Run.log_code() with include_fn= and exclude_fn= parameters to specify which files to include and exclude when you log code: 
import wandb

with wandb.init() as run:
    run.log_code(
         root="../",
         include_fn=lambda path: path.endswith(".py") or path.endswith(".ipynb"),
         exclude_fn=lambda path, root: os.path.relpath(path, root).startswith(
             "cache/"
         ),
    )
For more control over the types and locations of source code files that W&B saves, see the reference docs.

​
Automatically capture code

Use wandb.init(settings=wandb.Settings(code_dir=)) to automatically capture all code in the current directory and subdirectories. By default, W&B captures:
  • Files ending in .py
  • requirements.txt
  • Dockerfile
  • Excludes anything under wandb/ or .wandb/
The following example shows how capture all code in the current directory and subdirectories: 
import wandb

with wandb.init(settings=wandb.Settings(code_dir=".")) as run:
    # Your training code here
You can also specify a different directory by replacing . with the path to the directory that you want to capture.

​
View local changes

W&B generates diff files when you log code with staged or unstaged changes in your Git repository, and you can view the changes in the W&B UI. W&B names the diff files diff.patch or diff_<sha>.patch. These files contain local code changes that are not yet committed to Git, relative to HEAD. To view diff files:
  1. Navigate to your project’s workspace.
  2. In the left sidebar, click Workspace.
  3. In the run selector, select the run whose diff you want to view.
  4. On the run Overview page, select the Files tab.
  5. From the list of files, select the diff.patch or diff_<sha>.patch file.
  6. Select a file from the list to expand the diff for that file.
W&B supports two modes for viewing the diff across runs: unified and split.
  • Unified view shows the diff in a single pane:
    Diff patch unified view
  • Split view shows the diff in two panes, side-by-side:
    Diff patch split view
Toggle between the two views by selecting either the Unified or Split buttons in the upper right corner of the diff view, below the Download button.

​
Jupyter session history

W&B saves the history of code executed in your Jupyter notebook session. When you call wandb.init() inside of Jupyter, W&B adds a hook to automatically save a Jupyter notebook containing the history of code executed in your current session.
  1. Navigate to your project’s workspace that contains your code.
  2. Select the Artifacts tab in the project sidebar.
  3. Expand the code artifact.
  4. Select the Files tab.
Jupyter session history
This displays the cells that ran in your session along with any outputs created by calling IPython’s display method. This lets you see exactly what code ran within Jupyter in a given run. When possible, W&B also saves the most recent version of the notebook, which you find in the code directory as well.
Jupyter session output

​
Compare code across runs

Compare code used in different W&B runs:
  1. Select the Add panels button in the top right corner of the page.
  2. Expand TEXT AND CODE dropdown and select Code.
Code comparer panel