IndexTTS-2.5#

  • Model Name: IndexTTS-2.5

  • Model Family: IndexTTS2

  • Abilities: ['text2audio', 'text2audio_zero_shot', 'text2audio_voice_cloning', 'text2audio_emotion_control']

  • Multilingual: True (Chinese, English, Japanese, Spanish, and Arabic)

Specifications#

  • Model ID: IndexTeam/IndexTTS-2.5

Execute the following command to launch the model:

xinference launch --model-name IndexTTS-2.5 --model-type audio

Upstream officially supports Python 3.10 and 3.11. A GPU is strongly recommended; enable use_bf16 at launch to reduce GPU memory use. Text-based emotion guidance additionally requires use_qwen_emo=True.