Skip to content
Computer vision

Ivis-N1.2-Omni

арабо-англо-японская VLM на гибридной архитектуре

1 min readupdated

Ivis-N1.2-Omni 2B A1.3B VL от арабской компании Nepetai — мультимодальная модель для текста и изображений. Поддерживает арабский, английский и японский, а также программирование и математические рассуждения. Всего 2,01 млрд параметров, из них около 1,30 млрд активны на токен.

Ключевые особенности

  • Гибридная архитектура Transformer-Mamba-MoE: 24 слоя, из них 18 слоёв GQA-внимания и 6 слоёв Mamba Selective SSM.
  • Sparse MoE: 4 слоя, по 8 экспертов, Top-1 routing — на токен активируется один эксперт.
  • Поддержка изображений через VL vision encoder и генерация текстовых ответов.
  • Контекст до 128 000 токенов.
  • Словарь на 13 752 токена для арабского, английского, японского, кода и математики.
  • Лицензия Apache-2.0.

Пример

Сценарий использования: подать изображение и задать вопрос на арабском, английском или японском. Например: «Что изображено на фото? Ответь по-арабски.»

ПодробнееБенчмарков и примера кода в карточке нет.
Требования к VRAM не указаны; ориентируйтесь на 2,01 млрд параметров и контекст 128K.
Данные и детали обучения не раскрыты.
Лицензия Apache-2.0.
Discuss Telegram 0 81

Will it run?

Runs in F16

Fits without quantisation.

  • F16weights 5.3 GB · ready files6.3 GB
  • FP8 / INT8recommendedweights 2.6 GB3.4 GB
  • 4-bitweights 1.5 GB2.1 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 · F16 — from RTX 5060 Ti 8GB

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.

    # nepetai/Ivis-N1.2-Omni
    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 nepetai/Ivis-N1.2-Omni --local-dir ~/models/Ivis-N1.2-Omni
    # The model is in ~/models. How to run it is in the model card (link above).
    Add more LoRAs to Ivis-N1.2-Omni
  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

New posts about versions of Ivis-N1.2-Omni (quants, finetunes, LoRAs) will show up under the bell.

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

About the model

nepetai/Ivis-N1.2-Omni
Author
nepetai
Category
Computer vision
Released
Posted

On NeyroLearn

Hugging Face

810

Ask AI about this model

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

Comments

Computer vision

B6-27B

B6-27B
Computer vision

CoSTA-MIX

CoSTA-MIX