Add DashScope agent configuration files for VAD, LLM, TTS, and ASR services
- Introduced new YAML configuration files for DashScope, detailing agent behavior settings for VAD, LLM, TTS, and ASR. - Configured parameters including model paths, API keys, and service URLs for real-time processing. - Ensured compatibility with existing agent-side behavior management while providing specific settings for DashScope integration.
This commit is contained in:
47
engine/config/agents/dashscope.yaml
Normal file
47
engine/config/agents/dashscope.yaml
Normal file
@@ -0,0 +1,47 @@
|
||||
# Agent behavior configuration for DashScope realtime ASR/TTS.
|
||||
# This file only controls agent-side behavior (VAD/LLM/TTS/ASR providers).
|
||||
# Infra/server/network settings should stay in .env.
|
||||
|
||||
agent:
|
||||
vad:
|
||||
type: silero
|
||||
model_path: data/vad/silero_vad.onnx
|
||||
threshold: 0.5
|
||||
min_speech_duration_ms: 100
|
||||
eou_threshold_ms: 800
|
||||
|
||||
llm:
|
||||
# provider: openai | openai_compatible | siliconflow
|
||||
provider: openai_compatible
|
||||
model: deepseek-v3
|
||||
temperature: 0.7
|
||||
api_key: your_llm_api_key
|
||||
api_url: https://api.qnaigc.com/v1
|
||||
|
||||
tts:
|
||||
provider: dashscope
|
||||
api_key: your_tts_api_key
|
||||
api_url: wss://dashscope.aliyuncs.com/api-ws/v1/realtime
|
||||
model: qwen3-tts-flash-realtime
|
||||
voice: Cherry
|
||||
dashscope_mode: commit
|
||||
speed: 1.0
|
||||
|
||||
asr:
|
||||
provider: dashscope
|
||||
api_key: your_asr_api_key
|
||||
api_url: wss://dashscope.aliyuncs.com/api-ws/v1/realtime
|
||||
model: qwen3-asr-flash-realtime
|
||||
interim_interval_ms: 500
|
||||
min_audio_ms: 300
|
||||
start_min_speech_ms: 160
|
||||
pre_speech_ms: 240
|
||||
final_tail_ms: 120
|
||||
|
||||
duplex:
|
||||
enabled: true
|
||||
system_prompt: You are a helpful, friendly voice assistant. Keep your responses concise and conversational.
|
||||
|
||||
barge_in:
|
||||
min_duration_ms: 200
|
||||
silence_tolerance_ms: 60
|
||||
68
engine/config/agents/volcengine.yaml
Normal file
68
engine/config/agents/volcengine.yaml
Normal file
@@ -0,0 +1,68 @@
|
||||
# Agent behavior configuration (safe to edit per profile)
|
||||
# This file only controls agent-side behavior (VAD/LLM/TTS/ASR providers).
|
||||
# Infra/server/network settings should stay in .env.
|
||||
|
||||
agent:
|
||||
vad:
|
||||
type: silero
|
||||
model_path: data/vad/silero_vad.onnx
|
||||
threshold: 0.5
|
||||
min_speech_duration_ms: 100
|
||||
eou_threshold_ms: 800
|
||||
|
||||
llm:
|
||||
# provider: openai | openai_compatible | siliconflow
|
||||
provider: openai_compatible
|
||||
model: deepseek-v3
|
||||
temperature: 0.7
|
||||
# Required: no fallback. You can still reference env explicitly.
|
||||
api_key: your_llm_api_key
|
||||
# Optional for OpenAI-compatible endpoints:
|
||||
api_url: https://api.qnaigc.com/v1
|
||||
|
||||
tts:
|
||||
# provider: edge | openai_compatible | siliconflow | dashscope
|
||||
# dashscope defaults (if omitted):
|
||||
# api_url: wss://dashscope.aliyuncs.com/api-ws/v1/realtime
|
||||
# model: qwen3-tts-flash-realtime
|
||||
# dashscope_mode: commit (engine splits) | server_commit (dashscope splits)
|
||||
# note: dashscope_mode/mode is ONLY used when provider=dashscope.
|
||||
# volcengine defaults (if omitted):
|
||||
provider: volcengine
|
||||
api_url: https://openspeech.bytedance.com/api/v3/tts/unidirectional
|
||||
resource_id: seed-tts-2.0
|
||||
app_id: your_tts_app_id
|
||||
api_key: your_tts_api_key
|
||||
speed: 1.1
|
||||
voice: zh_female_vv_uranus_bigtts
|
||||
|
||||
asr:
|
||||
asr:
|
||||
provider: volcengine
|
||||
api_url: wss://openspeech.bytedance.com/api/v3/sauc/bigmodel
|
||||
app_id: your_asr_app_id
|
||||
api_key: your_asr_api_key
|
||||
resource_id: volc.bigasr.sauc.duration
|
||||
uid: caller-1
|
||||
model: bigmodel
|
||||
request_params:
|
||||
end_window_size: 800
|
||||
force_to_speech_time: 1000
|
||||
enable_punc: true
|
||||
enable_itn: false
|
||||
enable_ddc: false
|
||||
show_utterance: true
|
||||
result_type: single
|
||||
interim_interval_ms: 500
|
||||
min_audio_ms: 300
|
||||
start_min_speech_ms: 160
|
||||
pre_speech_ms: 240
|
||||
final_tail_ms: 120
|
||||
|
||||
duplex:
|
||||
enabled: true
|
||||
system_prompt: 你是一个人工智能助手,你用简答语句回答,避免使用标点符号和emoji。
|
||||
|
||||
barge_in:
|
||||
min_duration_ms: 200
|
||||
silence_tolerance_ms: 60
|
||||
Reference in New Issue
Block a user