Uploaded May 2023 | Updated September 2026, 1 week ago

---

Pre-trained models are an important part of modern ML workflows, but it’s not always easy to find the best one for your needs. With the launch of Kaggle Models, we’re making thousands of models from Google Research, Deepmind, and other sources more accessible to the ML developer community alongside our ML resource hub of 200k+ public datasets. To showcase how to use Kaggle Models to improve ML projects, we’ll demo how to use an open source model from Kaggle and use it to submit to a competition.

Resources:
Kaggle Models → https://goo.gle/3MzI3HJ
Demo Notebook → https://goo.gle/3UbR9fd
TFHub + Kaggle blog post → https://goo.gle/40HpLZb
Find full set of ML resources here → https://g.co/ai/build

Speaker: Nate Keating

Watch more:
Watch all the Technical Sessions from Google I/O 2023 → https://goo.gle/IO23_sessions
Watch the AI/ML Sessions → https://goo.gle/IO23_ai_ml
All Google I/O 2023 Sessions → https://goo.gle/IO23_all

Subscribe to TensorFlow → https://goo.gle/TensorFlow

#GoogleIO


Products mentioned: TensorFlow - General
Event: Google I/O 2023
Speakers: Nate Keating