config.yaml

# * Settings marked with * are advanced settings that won't appear in the Streamlit page and can only be modified manually in config.py
## ======================== Basic Settings ======================== ##
# API settings
api:
  key: 'YOUR_API_KEY'
  base_url: 'https://yunwu.zeabur.app'
  model: 'claude-3-5-sonnet-20240620'

# Replicate API settings
replicate_api_token: 'YOUR_REPLICATE_API_TOKEN'

# Language settings, written into the prompt, can be described in natural language
target_language: '简体中文'

whisper:
  # Whisper settings [whisperx, whisperxapi]
  method: 'whisperx'
  # Whether to perform UVR processing before transcription
  uvr_before_transcription: true
  # Whisper specified recognition language [en, zh, auto] auto for automatic detection, en for forced translation to English
  language: 'en'

# Video resolution [0x0, 640x360, 1920x1080]  0x0 will generate a 0-second black video placeholder
resolution: '640x360'

## ======================== Advanced Settings ======================== ##
# Web display language, auto for automatic detection based on system [zh_CN, en_US, auto]
display_language: 'auto'

# * HTTP proxy settings
http_proxy:
  use: false
  address: "http://127.0.0.1:7890"

# *Default resolution for downloading YouTube videos [360, 1080, best]
ytb_resolution: '360'

subtitle:
  # *Maximum length of each subtitle line in characters
  max_length: 70
  # *Translated subtitles are slightly larger than source subtitles, affecting the reference length for subtitle splitting
  target_multiplier: 1.1

# *Number of LLM multi-threaded accesses
max_workers: 8
# *Maximum number of words for the first rough cut, below 18 will cut too finely affecting translation, above 22 is too long and will make subsequent subtitle splitting difficult to align
max_split_length: 20

# *Whether to pause after extracting professional terms and before translation, allowing users to manually adjust the terminology table output\log\terminology.json
pause_before_translate: false

## ======================== Dubbing Settings ======================== ##
# TTS selection [openai_tts, gpt_sovits, azure_tts, fish_tts]
tts_method: 'fish_tts'

# OpenAI TTS-1 API configuration
openai_tts:
  voice: 'alloy'
  api_key: 'YOUR_OPENAI_API_KEY'
  base_url: 'https://yunwu.zeabur.app'

# Azure configuration
azure_tts:
  key: 'YOUR_AZURE_API_KEY'
  region: 'eastasia'
  voice: 'zh-CN-YunfengNeural'

# SoVITS configuration
gpt_sovits:
  character: 'Huanyuv2'
  refer_mode: 3

# FishTTS configuration
fish_tts:
  api_key: 'YOUR_FISH_API_KEY'
  character: 'AD学姐'
  character_id_dict:
    'AD学姐': '7f92f8afb8ec43bf81429cc1c9199cb1'
    '丁真': '54a5170264694bfc8e9ad98df7bd89c3'
    '赛马娘': '0eb38bc974e1459facca38b359e13511'
    '蔡徐坤': 'e4642e5edccd4d9ab61a69e82d4f8a14'
    '雷军': '738d0cc1a3e9430a9de2b544a466a7fc'

# *Audio speed range
speed_factor:
  min: 1
  max: 1.4
  normal: 1.2  # *Considered normal speech rate

# *Merge audio configuration
min_subtitle_duration: 3
min_trim_duration: 2.50

# Volume settings
original_volume: 0.1  # Original voice volume in dubbed video (0.1 = 10% or 0)
dub_volume: 1.5  # *Dubbed audio volume (1.5 = 150%, most original dubbing audio is relatively quiet)






## ======================== Additional settings 请勿修改======================== ##
# Whisper model directory
model_dir: './_model_cache'

# Supported upload video formats
allowed_video_formats:
- 'mp4'
- 'mov'
- 'avi'
- 'mkv'
- 'flv'
- 'wmv'
- 'webm'

# LLMs that support returning JSON format
llm_support_json:
- 'deepseek-coder'
- 'gpt-4o'

# Spacy models
spacy_model_map:
  en: 'en_core_web_md'
  ru: 'ru_core_news_md'
  fr: 'fr_core_news_md'
  ja: 'ja_core_news_md'
  es: 'es_core_news_md'
  de: 'de_core_news_md'
  it: 'it_core_news_md'
  zh: 'zh_core_web_md'

# Languages that use space as separator
language_split_with_space:
- 'en'
- 'es'
- 'fr'
- 'de'
- 'it'
- 'ru'

# Languages that do not use space as separator
language_split_without_space:
- 'zh'
- 'ja'

video:
  resolution: '0x0'