feat: Add audio parameter support to gemini tts models #11287
Conversation
|
The latest updates on your projects. Learn more about Vercel for Git ↗︎
|
AyrennC
changed the title
- Add is_model_gemini_audio_model() method to detect TTS models - Include 'audio' parameter in supported params for TTS models - Map OpenAI audio parameter to Gemini speechConfig format - Add _extract_audio_response_from_parts() method to transform audio output to openai format
|
|
LiteLLM Mock Tests timed out after 8 minutes, all test were successful until time out. |
| ) | ||
|
|
||
| # Map OpenAI audio parameter to Gemini speech config | ||
| speech_config = {} |
There was a problem hiding this comment.
can we have these be typed dict's inside types/llms/vertex_ai.py - so any future updates are also tracked correctly
There was a problem hiding this comment.
added typed dict for SpeechConfig and its child in types/llms/vertex_ai.py
- simplified gemini tts model detection - moved gemini_tts test to test_litellm
Thanks! I went ahead and made the suggested changes:
|
|
Thanks @AyrennC would you mind contributing docs for the change, so people know how to use this? For VertexAI - here Contributing guide - https://docs.litellm.ai/docs/extras/contributing (although it's just an .md change, so i'm sure you can just do it on github as well) |
Title
Add 'audio' params support to all gemini tts models
Relevant issues
Fixes #11250
Fixes #11118
Pre-Submission checklist
Please complete all items before asking a LiteLLM maintainer to review your PR
tests/litellm/directory, Adding at least 1 test is a hard requirement - see detailsmake test-unitType
🆕 New Feature
🐛 Bug Fix
✅ Test
Changes