Name		Name	Last commit message	Last commit date
parent directory ..
audio		audio
graphics		graphics
llm		llm
multi-modal		multi-modal
other		other
vision		vision
webcam		webcam
README.md		README.md
pipelines_index.json		pipelines_index.json

README.md

Creator Guide for Visual Blocks

Thank you creators for contributing to Visual Blocks for ML! We firmly believe that: with your contribution, we can inspire more hackers, designers, and practitioners to unleash their creativity!

Contributing a New Pipeline

Step 1: CLA

Fork this repository and clone the forked repo to your laptop or workstation. Review and sign the Contributor License Agreement (CLA). Now you are ready to create a pull request :)

Step 2: Creating Your Pipeline

Visit https://visualblocks.withgoogle.com, click the Demo button. Next, click the Demo: Create Your Own tab and start dragging and dropping nodes from the node gallery and make your own pipelines. After you complete, come up with a descriptive title, and click the Export button on the top right corner to download. You may also click the Import button to load an existing pipeline.

Step 3: Exporting the Pipeline

Copy the exported pipeline to your local repository, commit and create a pull request by following the official GitHub guide. Make sure to add your pipeline to the corresponding subfolder within the pipeline folder (i.e. audio, vision, multi-modal, etc.). We strongly recommend contributors to also upload a GIF or JPG screenshot for your pipeline (width: 320px) to be featured in this guide.

Step 4: Describing Your Pipeline (Optional)

Write a .txt file that describes your pipeline. The description should be as clear and concise as possible. We encourage contributors to follow the two guidelines below:

D1: The description should explicitly explain what the pipeline does.
D2: The description should NOT explain how the pipeline does it.

For instance, for the weather summarizer pipeline:

☑ "Summarize the weather in San Francisco into in one sentence." - a good example

☒ "Weather summarizer" - a description that is too general and ambiguous [violating P1]

☒ "Get the weather in San Francisco first, and then use PaLM to summarize the generated content into one sentence." - a description that includes too many unnecessary technical details [violating P2]

Step 5: Naming your files

Please name your files in the following way:

.txt files should have the same name as the .json file. Example: pipeline_name.json can have a txt file with the name pipeline_name.txt.
Image files should either have the same name as the .json file, or have the same name with _highres appended to the end. Example: pipeline_name.json can have image files with the names pipeline_name.[image format] or pipeline_name_highres.[image format].

Step 6: Uploading

After you have uploaded your pipeline to the pipelines folder, you can easily share your amazing creation via URL like: https://visualblocks.withgoogle.com/#/edit/_?project_json=https:%2F%2Fraw.githubusercontent.com%2Fgoogle%2Fvisualblocks%2Fmain%2Fpipelines%2Fgraphics%2F3dphoto_portrait_depth.json

Example Pipelines

Please refer to graphics/3dphoto_portrait_depth.json and llm/palm2_weather_summarizer.json for two example pipelines and refer to the gallery below for interactive demos. Note that API keys and locally uploaded images are not exported for privacy and security.

Before you get started on the PaLM example, you need to obtain an API key first. Head to makersuite.google.com, sign up with your Google account, and click "Get an API key". Once you have the key, you can start using the API.