feat: add on-device LLM inference via LiteRT-LM#165
Open
utzcoz wants to merge 1 commit intoaaif-goose:mainfrom
Open
feat: add on-device LLM inference via LiteRT-LM#165utzcoz wants to merge 1 commit intoaaif-goose:mainfrom
utzcoz wants to merge 1 commit intoaaif-goose:mainfrom
Conversation
Adds an on-device model provider that runs Gemma 4 E2B/E4B locally using Google's LiteRT-LM (NPU/GPU accelerated). Users can download models from a built-in registry, manage them in settings, and select on-device inference from onboarding or the provider dropdown — no API key required. Changes: - New ON_DEVICE_LITERT provider with LiteRTProviderHandler wired into Agent.callLlm and tool resolution - Model registry loaded from assets/models_litert.json; downloads go through Android DownloadManager into app-private storage - ModelManagementScreen for browsing, downloading, and deleting on-device models; integrated into onboarding LLM configuration - OnDeviceModelManager scans for downloaded models on app startup (via GoslingApplication) so saved model identifiers resolve correctly across restarts/reinstalls - On-device system prompt preserves installed apps, screen resolution, and user memories so models can use tools instead of guessing URLs - Settings screen gracefully handles empty on-device model lists with a "No models downloaded" placeholder
utzcoz
force-pushed
the
feat/on-device-model-support
branch
from
06343e4 to
2d1097d
Compare
Contributor
Author
|
Hi @michaelneale, PTAL. |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Adds an on-device model provider that runs Gemma 4 E2B/E4B locally using Google's LiteRT-LM (NPU/GPU accelerated). Users can download models from a built-in registry, manage them in settings, and select on-device inference from onboarding or the provider dropdown — no API key required.
Changes:
Inspired by recent Google AI Edge Gallery + Gemma 4 E2B/E4B release.