Skip to content
Audio

Humaneness Voice Large 1.7 Singing

an experimental singing and speech generation model

1 min readupdated

This new release from LAION combines speech synthesis with song generation. Built on the MOSS-TTS speech transformer, the model was trained on a massive dataset covering songs in English, German, Dutch, French, and Spanish, alongside humming and whistling. The adapter is already merged into the base BF16 weights, so you can load the model using standard methods. It works well for experimenting with complex vocals, intonation, and musical sketches.

Key features

  • Vocal performance and singing support across five European languages
  • Generates humming, whistling, and wordless melodies in addition to lyrics
  • Total parameter count of 4.13 billion
  • The LoRA adapter is already merged into the base weights, requiring no extra setup

What you can do

  • Generate short vocal pieces or lullabies in multiple languages
  • Create an experimental track using a text prompt and an instruction file
  • Test speech generation with specific musical and emotional cues

Example

hf download laion/Humaneness-Voice-Large-1.7-Singing-Experimental --local-dir experimental-singing
cd experimental-singing
python -m pip install -r requirements.txt
python infer.py --text "An original short lyric." --prompt-file direction.txt --language en --seconds 20 --output example.wav
Read moreDetails
Architecture: MOSS-TTS local transformer v1.5
Training: 3 epochs on 392,256 examples across 8 GH200 GPUs
Adapter parameters: rank/alpha 128 across 146 attention projections
License: CC BY 4.0
Discuss Telegram 0 40

Will it run?

Runs in BF16

Fits without quantisation.

  • BF16weights 7.7 GB · ready files9.0 GB
  • FP8 / INT8recommendedweights 3.8 GB4.7 GB
  • 4-bitweights 2.2 GB2.9 GB

Tip: go 8-bit.FP8 / INT8 is within about a percent of full precision in quality, takes half the memory and runs faster on RTX 40/50 and H100. Full precision only pays off for fine-tuning.

What to get for this model

Massed ComputeOpen Massed Compute

Your own GPU: FP8 / INT8 — from RTX 2060 · BF16 — from RTX 3080

Massed Compute hourly prices, November 2026. The discounted price applies with the coupon at checkout.

Estimate: weights plus working memory. Long videos, high resolutions and long contexts need more.

Run it on Massed Compute

The script downloads the model to the rented machine. How to run it is in the model card.

  1. 1. Rent a machine with a fitting GPU

    Open Massed Compute
  2. 2. Paste the script into the machine's terminal

    Connect over SSH (the Massed Compute panel shows the command) or open a terminal on the machine's desktop.

    HF_TOKEN / CIVITAI_TOKENThe script starts with token lines: paste yours between the quotes. The Hugging Face token is for gated models, the Civitai one for Civitai models.

    # laion/Humaneness-Voice-Large-1.7-Singing-Experimental
    mkdir -p ~/models
    curl -LsSf https://astral.sh/uv/install.sh | sh
    source "$HOME/.local/bin/env"
    uv tool install "huggingface_hub[cli]"
    hf download laion/Humaneness-Voice-Large-1.7-Singing-Experimental --local-dir ~/models/Humaneness-Voice-Large-1.7-Singing-Experimental
    # The model is in ~/models. How to run it is in the model card (link above).
  3. 3. Use it

    The model is in ~/models. How to run it is in the model card (link above).

Stop the machine in the panel when you are done: it bills by the minute while it runs.

How to speed it up

Picked for this model type. Choose your hardware above to see what fits you exactly.

  • FlashAttention 44.0.0b40from November 18, 2026new release

    fastest exact attention on B200

    Kernels for Blackwell (beta). Exact attention without quantisation — fine for everything, LLMs included.

    How to enable: pip install flash-attn-4; vLLM and SGLang pick it up themselves.

    Earlier generations (2)
    • FlashAttention 3 — The Hopper (H100/H200) version with FP8.
    • FlashAttention 2 — The standard for RTX 30/40/50 and A100: exact attention without the memory cost of long contexts.
  • torch.compile2.15.0from October 28, 2026new release

    10–30% for free

    Compiles the model for your GPU. Slower first run, faster afterwards.

    How to enable: model = torch.compile(model); in ComfyUI the TorchCompileModel node.

Model family

  1. moss-tts-local-transformer-4.55b-voice-acting-v2family root
  2. moss-tts-local-transformer-4.55b-voice-acting-v2-sft3
  3. Humaneness-Voice-Large-1.2-Alpha
  4. Humaneness-Voice-Large-1.3-Alpha
  5. Humaneness-Voice-Large-1.7
  6. Humaneness-Voice-Large-1.7-Singing-Experimental

New posts about versions of moss-tts-local-transformer-4.55b-voice-acting-v2 (quants, finetunes, LoRAs) will show up under the bell.

No known relatives of this model yet. All derivatives on Hugging Face

About the model

laion/Humaneness-Voice-Large-1.7-Singing-Experimental
Author
laion
Category
Audio
Released
Posted

On NeyroLearn

Hugging Face

400

Ask AI about this model

Answered by Gemini 4 · the chat is kept only in this tab

Comments