Fish Audio Drama 3 vs Ming-3.0-tts-0.5B

Fish Audio Drama 3

A local voice cousin for directed speech, multi-character scenes, voice cloning and word-level edits in about 80 languages. It needs an NVIDIA GPU, about 29 GB for the int4 build or at least 48 GB for FP8, through SGLang-Omni or fish-speech. The main risk is the licence, revocable, capped at USD 500,000 revenue for commercial use and barring hosted services.

Ming-3.0-tts-0.5B

A small MIT-licensed speech cousin for offline narration, voice cloning and voice design in seven languages and two dialects. It needs about 4 GB but runs only through inclusionAI's own Python scripts on CUDA, with no streaming. The main risk is a loading path that runs the publisher's code, with no safety evaluation or security contact.

Fish Audio Drama 3

1. Install SGLang-Omni from source with uv pip install --prerelease=allow -e ".[fishaudio]", then sgl-omni serve --model-path fishaudio/drama-3-int4 --port 8000 (or fishaudio/drama-3 for FP8 on 48 GB or more)
2. POST JSON to /v1/audio/speech with input, optional references (audio_path and its transcript) and previous_takes; set response_format and stream for the output
3. Start each line with <|speaker:N|> (N from 0 to 15) for several voices, and separate lines with <|audio_separator|>
4. Direct with point tags like [sighs], scoped tags like <whisper>...</whisper>, a bracketed style prefix, or Direction: and Line: for a director's note
5. The int4 build needs compute capability 8.0 or newer; only the triton attention backend is accepted
6. Keep each reference clip under 60 seconds and the prompt under 8,192 rows, about 21.5 rows per second of reference audio
7. Check the licence before commercial use or exposing it to other users, and only clone voices with consent
8. Treat text from untrusted sources as a script that can carry directions and speaker tags

Ming-3.0-tts-0.5B

1. Clone github.com/inclusionAI/Ming-3.0-tts and pip install -r requirements.txt (torch 2.6.0, transformers 4.46.3), since the Hugging Face repository alone won't load
2. Start with python infer.py --model_path inclusionAI/Ming-3.0-tts-0.5B --task IP --text "The garden will remain open until sunset." --ip US_James --use_zero_spk_emb --no_text_normalization --output output/tts.wav
3. Pass --no_text_normalization unless pynini 2.1.6 is installed, since text normalisation imports it
4. Pick one task per call and pass only that task's controls, since controls from another task are rejected
5. Run python tools/preflight.py --model_path <dir> to check the checkpoint before the first call
6. inclusionAI tested only NVIDIA H800 and H20 with CUDA 12.4, and no doc covers PyTorch on ROCm, so the AMD reference machine is untested
7. On Apple silicon, mlx-audio 0.6.1 or later loads mlx-community's conversion, which isn't inclusionAI's
8. Clone only voices you have consent for and label the audio as synthetic, as the card asks