search

google / visualblocks

Visual Blocks for ML is a Google visual programming framework that lets you create ML pipelines in a no-code graph editor. You – and your users – can quickly prototype workflows by connecting drag-and-drop ML components, including models, user inputs, processors, and visualizations.
https://visualblocks.withgoogle.com
Apache License 2.0
1.18k stars 155 forks source link

Import models from Teachable Machines -> Tensorflow #52

Closed xavido closed 4 months ago

xavido commented 5 months ago

Hi, Is there any way to import models from Teachable Machine ? Thanks!

Xavi

mattsoulanille commented 4 months ago

Yes, but it depends on the model type. For example, here's a pipeline I made that predicts whether a cellphone is in the user's hand. It uses the teachable machine 224x224px model (based on MobileNetV2 I think). To host the model, I exported it as Tensorflow.js in teachable machine and uploaded it as a sharable link. When you export your model, you'll need to append /model.json to the link to run it in VisualBlocks.

Different model / project types might not have their pre and post processing pipelines supported at the moment.

ruofeidu commented 4 months ago

Thank you Matthew for the great example!!! Feel free to make a pull request to https://github.com/google/visualblocks/tree/main/pipelines and I'm happy to highlight your example on Community Page!

ruofeidu commented 4 months ago

Thank you Matthew for the great example!!! Feel free to make a pull request to https://github.com/google/visualblocks/tree/main/pipelines and I'm happy to highlight your example on Community Page!!

On Tue, Jun 18, 2024 at 10:32 AM Matthew Soulanille < @.***> wrote:

Yes, but it depends on the model type. For example, here's a pipeline I made that predicts whether a cellphone is in the user's hand https://github.com/user-attachments/files/15890339/teachable_machine_image.json. It uses the teachable machine 224x224px model (based on MobileNet). To host the model, I exported it as Tensorflow.js in teachable machine and uploaded it as a sharable link. When you export your model, you'll need to append /model.json to the link to run it in VisualBlocks.

Different model / project types might not have their pre and post processing pipelines supported at the moment.

— Reply to this email directly, view it on GitHub https://github.com/google/visualblocks/issues/52#issuecomment-2176624078, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAMLHAITLBPVPJU2TYKLIT3ZIBVJ7AVCNFSM6AAAAABIBSO64KVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDCNZWGYZDIMBXHA . You are receiving this because you are subscribed to this thread.Message ID: @.***>

xavido commented 4 months ago

Wow, thanks so much Mathew ! I will test it. Thanks so much! I'm using this workflow to work with primary and secondary teachers in rapid prototyping educational projects.

xavido commented 4 months ago

I close the issue as it's fixed.