Model reference · open weights
ice-012-audio is an open-weight audio or speech model from fanuonai. AxForge deploys and operates it for you on dedicated EU-owned hardware — with the licence handled where one is required.
Available as managed deployment — configured and operated for you on dedicated EU hardware, quoted per deployment.
What it is
| Released by | fanuonai |
|---|---|
| Type | Audio & music |
| Task | Text→speech |
| Parameters (lead) | 714M |
| Runs with | iceAudio |
| Released | 2026-08-29 |
| Popularity | 801 downloads / month |
| Licence | Commercial licence needed |
About
ICE-012 Audio is a multilingual TTS model with streaming output and reference-based voice cloning. It introduces an active acoustic adapter — conditioning codec embeddings before the backbone and refining hidden states after it.
| Property | Value |
|---|---|
| Languages | 590 names / variants |
| Voice cloning | Reference-audio based |
| Voice controls | Gender · Age · Pitch · Accent · Style · Speed |
| License | cc-by-nc-4.0 |
from transformers import AutoModelForCausalLM
import torch
model = AutoModelForCausalLM.from_pretrained(
"darkps/ice-012-audio",
trust_remote_code=True,
device_map="auto",
)
python 02.py --language "arabic" --gender female --style "young adult" --speed 1.0 --output out.wav
python 02.py --text-file input.txt --language "en" --gender male --style "elderly, low pitch"
python 02.py --language "egyptian lahgtna" --gender female --auto-voice
| Argument | Default | Description |
|---|---|---|
--model | ./ice-012-Audio | Path to model directory |
--output | tts_stream.wav | Output WAV file |
--text-file | — | Read text from file |
--language | Arabic | Name or 2–3-letter ID |
--gender | — | male / female |
--style | — | Comma-separated modifiers |
--speed | 1.0 | Speed multiplier |
--auto-voice | off | Model selects voice automatically |
python 03.py --reference ref.wav --output clone.wav
python 03.py --reference ref.wav --text-file input.txt --reference-text "transcript"
| Argument | Default | Description |
|---|---|---|
--model | ./ice-012-Audio | Path to model directory |
--reference | required | Reference WAV file |
--output | clone_stream.wav | Output WAV file |
--text-file | — | Read text from file |
--reference-text | — | Reference transcript (improves quality) |
--speed | 1.0 | Speed multiplier |
Use only audio you own or have explicit permission to process.
| Value | |
|---|---|
male | Male voice |
female | Female voice |
--style)| Category | Values |
|---|---|
| Age | child · teenager · young adult · middle-aged · elderly |
| Pitch | very low pitch · low pitch · moderate pitch · high pitch · very high pitch |
| Style | whisper |
| Accent | american accent · british accent · australian accent · canadian accent · indian accent · chinese accent · japanese accent · korean accent · russian accent · portuguese accent |
Combining modifiers:
--gender female --style "young adult, high pitch"
--gender male --style "elderly, low pitch"
--gender female --style "teenager, british accent"
--speed)0.5 · 1.0 (default) · 1.5 · 2.0
Accepts a full name (case-insensitive) or a 2–3-letter ID. Unrecognized values fall back to language-agnostic mode.
# All three are equivalent:
language = "palestinian lahgtna"
language = "Palestinian Lahgtna"
language = "ps"
🎙️ Sample 01 🎙️ Sample 02 🎙️ Sample 03 🎙️ Sample 04
| Name | ID |
|---|---|
| Egyptian Lahgtna | eg |
| Saudi Lahgtna | sa |
| Moroccan Lahgtna | ma |
| Bahraini Lahgtna | bh |
| Sudanese Lahgtna | sd |
| Iraqi Lahgtna | iq |
| Lebanese Lahgtna | lb |
| Syrian Lahgtna | sy |
| Libyan Lahgtna | ly |
| Palestinian Lahgtna | ps |
| Tunisian Lahgtna | tn |
| Algerian Lahgtna | dz |
| Yemeni Lahgtna | ye |
| # | Display name | Accepted language name | ID |
|---|---|---|---|
| 1 | Egyptian Lahgtna | egyptian lahgtna | eg |
| 2 | Saudi Lahgtna | saudi lahgtna | sa |
| 3 | Moroccan Lahgtna | moroccan lahgtna | ma |
| 4 | Bahraini Lahgtna | bahraini lahgtna | bh |
| 5 | Sudanese Lahgtna | sudanese lahgtna | sd |
| 6 | Iraqi Lahgtna | iraqi lahgtna | iq |
| 7 | Lebanese Lahgtna | lebanese lahgtna | lb |
| 8 | Syrian Lahgtna | syrian lahgtna | sy |
| 9 | Libyan Lahgtna | libyan lahgtna | ly |
| 10 | Palestinian Lahgtna | palestinian lahgtna | ps |
| 11 | Tunisian Lahgtna | tunisian lahgtna | tn |
| 12 | Algerian Lahgtna | algerian lahgtna | dz |
| 13 | Yemeni Lahgtna | yemeni lahgtna | ye |
| 14 | Abkhazian | abkhazian | ab |
| 15 | Afrikaans | afrikaans | af |
| 16 | Albanian | albanian | sq |
| 17 | Algerian Arabic | algerian arabic | arq |
| 18 | Algerian Saharan Arabic | algerian saharan arabic | aao |
| 19 | Amharic | amharic | am |
| 20 | Armenian | armenian | hy |
| 21 | Assamese | assamese | as |
| 22 | Asturian | asturian | ast |
| 23 | Azerbaijani | azerbaijani | az |
| 24 | Baharna Arabic | baharna arabic | abv |
| 25 | Banjar | banjar | bjn |
| 26 | Bashkir | bashkir | ba |
| 27 | Basque | basque | eu |
| 28 | Batak Mandailing | batak mandailing | btm |
| 29 | Belarusian | belarusian | be |
| 30 | Bengali | bengali | bn |
| 31 | Betawi | betawi | bew |
| 32 | Bhojpuri | bhojpuri | bho |
| 33 | Bosnian | bosnian | bs |
| 34 | Breton | breton | br |
| 35 | Buginese | buginese | bug |
| 36 | Bukharic | bukharic | bhh |
| 37 | Bulgarian | bulgarian | bg |
| 38 | Bulu (Cameroon) | bulu (cameroon) | bum |
| 39 | Bundeli | bundeli | bns |
| 40 | Burmese | burmese | my |
| 41 | Burushaski | burushaski | bsk |
| 42 | Cajatambo North Lima Quechua | cajatambo north lima quechua | qvl |
| 43 | Cameroon Pidgin | cameroon pidgin | wes |
| 44 | Campidanese Sardinian | campidanese sardinian | sro |
| 45 | Cantonese | cantonese | yue |
| 46 | Catalan | catalan | ca |
| 47 | Cebuano | `cebua |
From the published model card. Full card on the HuggingFace links in the sidebar.
Using it via the API
Once AxForge deploys ice-012-audio for you, it answers on the OpenAI-compatible API — the same base URL and keys as every other model. (ice-012-audio below is illustrative; you get the exact model name on deployment.)
$ curl -sS https://api.axforge.ai/v1/audio/transcriptions \ -H "Authorization: Bearer $AXFORGE_API_KEY" \ -F model="ice-012-audio" -F file=@audio.mp3
Create an account — your API key is available in the console. 3M free tokens every 30 days with every new account.