Skip to content

Conversation

sozercan
Copy link
Collaborator

@sozercan sozercan commented Jul 23, 2025

Reason for Change:

Adds AIKit as part of integrations to extend KAITO with CPU-based inference, and increased custom model support with a variety of different formats.

Requirements

  • added unit tests and e2e tests (if applicable).

Issue Fixed:
Fixes #1212

Notes for Reviewers:
I'll update versioned docs after LGTM

Signed-off-by: Sertac Ozercan <sozercan@gmail.com>
Copy link

Title

Add AIKit Integration Documentation


Description

  • Added AIKit integration documentation

  • Updated sidebar to include AIKit category


Changes walkthrough 📝

Relevant files
Documentation
sidebars.js
Add AIKit to sidebar integrations                                               

website/sidebars.js

  • Added a new category for 'Integrations'
  • Included 'aikit' under the Integrations category
  • +8/-0     
    aikit.md
    Add AIKit integration documentation                                           

    website/docs/aikit.md

  • Created new markdown file for AIKit integration
  • Documented overview, deployment steps, and custom model creation
  • +117/-0 

    Need help?
  • Type /help how to ... in the comments thread for any questions about PR-Agent usage.
  • Check out the documentation for more information.
  • Copy link

    PR Reviewer Guide 🔍

    Here are some key observations to aid the review process:

    ⏱️ Estimated effort to review: 3 🔵🔵🔵⚪⚪
    🧪 No relevant tests
    🔒 No security concerns identified
    ⚡ Recommended focus areas for review

    Typo

    There is a typo in the text: "Create a KAITO workspacec configuration file" should be "Create a KAITO workspace configuration file".

    Create a KAITO workspacec configuration file to deploy your model. Here's a complete example:
    Missing Link

    The link to the installation guide in step 41 is incorrect. It points to "installation.md#install-workspace-controller", but the correct path should be verified as it might not exist or be incorrectly referenced.

    - Install [KAITO workspace controller](installation.md#install-workspace-controller) on your cluster

    Copy link

    PR Code Suggestions ✨

    Explore these optional code suggestions:

    CategorySuggestion                                                                                                                                    Impact
    General
    Adjust category collapse setting

    Ensure that the collapsed property is consistent with the intended user experience.
    Consider setting it to true if you want the category to be collapsed by default.

    website/sidebars.js [43-50]

     {
         type: 'category',
         label: 'Integrations',
    -    collapsed: false,
    +    collapsed: true,
         items: [
             'aikit',
         ],
     },
    Suggestion importance[1-10]: 6

    __

    Why: The suggestion asks to verify or ensure a change in the collapsed property, which is a minor improvement affecting user experience but not critical functionality.

    Low
    Fix spelling error

    Correct the typo in "workspacec" to "workspace".

    website/docs/aikit.md [45]

    -Create a KAITO workspacec configuration file to deploy your model.
    +Create a KAITO workspace configuration file to deploy your model.
    Suggestion importance[1-10]: 5

    __

    Why: Correcting the typo in "workspacec" to "workspace" is a minor improvement that enhances readability but does not affect functionality.

    Low

    Signed-off-by: Sertac Ozercan <sozercan@gmail.com>
    Signed-off-by: Sertac Ozercan <sozercan@gmail.com>
    @zhuangqh
    Copy link
    Collaborator

    This feature is awesome! Let's add a e2e test for it.

    - "run"
    - "--address=:5000"
    ```

    Copy link
    Collaborator

    Choose a reason for hiding this comment

    The reason will be displayed to describe this comment to others. Learn more.

    could you add a hint here? users may need to check the memory requirement before they deploy a cpu model and add these requirements to the spec.

    OOM is the most common problem we met during deploying a llm model.

    Copy link
    Collaborator Author

    Choose a reason for hiding this comment

    The reason will be displayed to describe this comment to others. Learn more.

    thanks, added a note. it is hard to get exact requirements, but i added general guidelines, and a way to get more accurate estimation.

    ```

    After building the image, you can use it in your KAITO workspace configuration by updating the `image` field.

    Copy link
    Collaborator

    Choose a reason for hiding this comment

    The reason will be displayed to describe this comment to others. Learn more.

    Can you also call out the limitations here. What are supported runtimes or anything that works in local AI but not included in this integration.

    Copy link
    Collaborator Author

    Choose a reason for hiding this comment

    The reason will be displayed to describe this comment to others. Learn more.

    thanks, added a note at the end.

    Signed-off-by: Sertac Ozercan <sozercan@gmail.com>
    @sozercan
    Copy link
    Collaborator Author

    sozercan commented Aug 1, 2025

    @zhuangqh thanks! we can follow up with e2e after docs pr is merged, wdyt?

    Copy link
    Collaborator

    @zhuangqh zhuangqh left a comment

    Choose a reason for hiding this comment

    The reason will be displayed to describe this comment to others. Learn more.

    LGTM

    @Fei-Guo Fei-Guo merged commit fa41991 into kaito-project:main Aug 2, 2025
    15 of 16 checks passed
    Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
    Projects
    Status: Done
    Development

    Successfully merging this pull request may close these issues.

    support for cpu based inference
    4 participants