feat: add FunASR speech recognition backend

[LauraGPT]

Summary

Adds FunASR as a new speech recognition backend for LocalAI.

FunASR is an industrial-grade ASR toolkit (16K+ GitHub stars) from Alibaba DAMO Academy, featuring:

• SenseVoice: 170x realtime speed, emotion & audio event detection, 50+ languages
• Paraformer: Highest accuracy for Chinese ASR, non-autoregressive
• Built-in VAD, punctuation restoration, and speaker diarization

Changes

• backend/python/funasr/ — New Python gRPC backend wrapping FunASR's AutoModel API
• backend/index.yaml — Meta definition and image entries for cpu/cuda12/cuda13/rocm/metal/l4t platforms

Usage

name: funasr-sensevoice
backend: funasr
parameters:
  model: iic/SenseVoiceSmall

The backend supports the standard /v1/audio/transcriptions endpoint with optional language parameter.

Test plan

• Verify pip install funasr installs cleanly in the backend venv
• Test SenseVoice model loading and transcription via gRPC
• Test with CPU and CUDA backends
• Verify existing backends are unaffected (index.yaml meta anchor is additive)