fix: use full onnx model with 8 threads for fast local TTS

Benchmark: short=0.37s, long=1.06s with 8 CPU threads.
GPU not available in pip sherpa-onnx, CPU is fast enough.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

backend/main/xiaozhi-server/core/providers/tts/sherpa_tts.py

@@ -23,7 +23,7 @@ class TTSProvider(TTSProviderBase):
         if not os.path.exists(model_file):
             model_file = f"{model_dir}/model.onnx"
 
-        num_threads = int(config.get("num_threads", 4))
+        num_threads = int(config.get("num_threads", 8))
 
         tts_config = sherpa_onnx.OfflineTtsConfig(
             model=sherpa_onnx.OfflineTtsModelConfig(