GSoC Week 10 Summary

Date: 29 July 2024 - 04 Aug 2024

Introduction

This week focused on refining the AGL Voice Agent by implementing auto-mode, testing integrations.

Tasks Completed This Week

# Fixed Recipes in meta-offline-voice-agent Layer

Addressed several issues within the meta-offline-voice-agent layer, including fixing the python3-pytorch recipe for arm64 and some patches.

# Fix agl-service-voiceagent.service by Removing Rasa

Removed the Rasa dependency from agl-service-voiceagent.service due to compatibility issues with newer Python versions.

# Implemented Auto-Mode in AGL Voice Assistant

Developed the auto-mode feature for the AGL voice assistant. When enabled, the voice assistant continuously listens for a wake word. Upon detecting the wake word, it starts voice recording for a set period, executes the spoken command, and then loops back to wake word detection.

# Tested Auto-Mode in agl-ivi-demo-flutter Image

Thoroughly tested the auto-mode feature in the agl-ivi-demo-flutter image.

# Upstream Accepted python3-pytorch recipe changes

The changes made to remove Intel-specific dependencies from the python3-pytorch recipe were accepted upstream. Upstream Pull Request

Code Pushed

• 30060: Integrate Whisper AI into AGL Voice Agent and Update Dependencies

Tasks for Next Week

• Update snips-nlu to improve natural language understanding capabilities.
• Update the UI of the Flutter app to enhance user experience and usability.

Conclusion

This week was productive, with key improvements and fixes enhancing the AGL Voice Agent's functionality. The coming week will focus on further refining the voice agent's capabilities, particularly NLU and Flutter app UI as well as testing on new hardware.