A Collection of Google's On-Device Models

Help us complete the list

  • To contribute go to C:\Users\YOUR_PC_USER\AppData\Local\Google\Chrome\User Data\optimization_guide_model_store
  • If you find a new non-empty folder not listed here please upload it to this repo

List of All Available Models

Following is the complete list of machine learning models in Chrome many of which are on your device. They are located in your User Data folder and you can easily check to see which ones you have as they are all in numbered folders.

  1. Unknown Optimization – Used when the optimization target is not specified.
  2. Painful Page Load Prevention – Triggers optimizations when page load times are expected to cause user frustration.
  3. Language Detector – Optimizes for detecting the language of the page content.
  4. Page Topics Classifier – Analyzes and classifies the topics present on a page.
  5. New Tab User Targeting – Segments users based on their interaction with new tab pages.
  6. Content Sharing Optimizer – Targets users who frequently share content.
  7. Voice Interaction Optimizer – Segments users based on voice interaction with the web.
  8. Model Validation – Ensures models are working as intended during execution.
  9. Page Entities Classifier – Detects and classifies entities (e.g., people, places, things) on a page.
  10. Notification Permission Predictor – Predicts whether a user is likely to allow notification permissions.
  11. Dummy Data Collection – Used for running experiments and collecting data client-side.
  12. Android Chrome Start Page Targeting – Segments users who frequently use Chrome’s Android start page.
  13. Query Tiles Targeting – Segments users interacting with query tiles.
  14. Page Visibility Detector – Optimizes based on the visibility of a page’s content.
  15. Page Topics Classifier V2 – A refined version of the page topics classifier.
  16. Low Engagement User Targeting – Segments users with low engagement on Chrome.
  17. Feed User Targeting – Targets users who prefer using content feeds.
  18. Price Tracking Action – Determines if price tracking should be shown as a contextual action.
  19. Text Classifier – Optimizes for smart text selection and entity extraction.
  20. Geolocation Permission Predictor – Predicts if a user will allow geolocation permissions.
  21. Shopping User Targeting – Segments users interested in shopping.
  22. Android Chrome Start Page Targeting V2 – Updated targeting for Chrome’s Android start page users.
  23. Search User Targeting – Segments users who frequently use search.
  24. Omnibox Tail Suggestions – Optimizes on-device suggestions in the Chrome Omnibox.
  25. Client-Side Phishing Protection – Protects against phishing attacks by analyzing client-side data.
  26. Omnibox URL Scoring – Optimizes URL suggestions in the Chrome Omnibox.
  27. Device Switcher Targeting – Targets users who switch between devices.
  28. Adaptive Toolbar Optimizer – Optimizes the toolbar based on user behavior.
  29. Tablet Productivity User Targeting – Segments tablet users who focus on productivity tasks.
  30. Phishing Image Protection – Embeds images to protect against phishing attacks.
  31. History Clusters Ranking – Ranks history clusters for the New Tab Page module.
  32. Web App Install Promotion – Promotes web app installations.
  33. Text Embedder – Embeds text for efficient storage and retrieval.
  34. Visual Search Classifier – Classifies and extracts search-related images from a webpage.
  35. Bottom Toolbar Targeting – Segments users based on bottom toolbar interaction.
  36. Autofill Field Classifier – Optimizes autofill suggestions by classifying form fields.
  37. iOS Module Ranker – Ranks modules on the iOS start page.
  38. Desktop NTP Module Optimizer – Determines which modules to display on a desktop’s New Tab Page.
  39. Preloading Heuristics – Optimizes preloading of links based on speculation rules.
  40. Text Safety Detector – Ensures the safety of text content on a page.
  41. Android Home Module Ranker – Ranks home modules for Android users.
  42. Compose On-Device Optimizer – Supports running the Compose feature on devices.
  43. Passage Embedder – Embeds passages for retrieval and analysis.
  44. Phrase Segmentation – Breaks up text into meaningful phrases.
  45. Compose Promotion Targeting – Targets users for promoting the Compose feature.
  46. URL Visit Resumption Ranker – Ranks URL visits for features that allow resuming visits.
  47. Camera Background Segmentation – Segments video frames to optimize camera backgrounds.
  48. History Search Optimizer – Optimizes search within browser history.
  49. Prompt API Optimizer – Configures optimizations for the Prompt API.
  50. Metrics Clustering Optimizer – Segments users based on metrics clustering.
  51. Summarize API Optimizer – Configures the API that summarizes content.
  52. Password Manager Form Classifier – Classifies forms to optimize password management.
  53. Suspicious Notification Detection – Detects potentially suspicious notification content.
  54. History Query Intent Optimizer – Classifies the intent behind history queries.

Source: DEJAN

Downloads last month

-

Downloads are not tracked for this model. How to track
Inference API
Unable to determine this model's library. Check the docs .