Skip to content
Video

MiniMax-H3-ref2va-V5xs-INT8

облегчённая видеомодель Hailuo 3.0 для ComfyUI

2 min read

Однофайловый чекпоинт на базе модели генерации видео MiniMax H3 (Hailuo 3.0) в варианте ref2va. Версия прошла прунинг (удаление лишних параметров для облегчения сети) и квантование в формат INT8 через инструмент DaSiWa Station, что позволяет запускать генерацию видео по референсу на менее требовательном железе. В отличие от тяжелых оригинальных весов, этот релиз упакован в один готовый файл специально для быстрой установки в ComfyUI. Подойдет тем, кто хочет протестировать архитектуру Hailuo 3.0 локально без сложной настройки окружения.

Ключевые особенности

  • Построена на архитектуре MiniMax H3 (Hailuo 3.0) в модификации ref2va для задач image-to-video.
  • Оптимизирована с помощью прунинга (pruned + graviton) и построчного квантования INT8 row-wise ConvRot (runtime).
  • Поставляется в виде единого файла (single-file checkpoint), готового к работе в ComfyUI без ручной сборки компонентов.
  • Квантование выполнено через пайплайн DaSiWa Station (сборка от 8 ноября 2026 года).

Что можно сделать

  • Генерировать видеоролики на основе входного изображения (image-to-video) в локальных пайплайнах ComfyUI.
  • Сравнить скорость работы и качество движения облегчённой INT8-сборки с референсной реализацией от MiniMax-AI.

Пример

Скачайте файл модели из репозитория и поместите его в директорию ComfyUI/models/diffusion_models/, после чего выберите чекпоинт в стандартном узле загрузки диффузионной модели в интерфейсе ComfyUI для генерации видео по референсному кадру.

ПодробнееБазовая архитектура: MiniMax H3 (Hailuo 3.0) ref2va.
Тип сжатия: pruned + graviton, INT8 row-wise ConvRot (runtime).
Лицензия: MiniMax H3 Community License Agreement.
Discuss Telegram 1

Will it run?

Runs in 4-bit

Full precision does not fit, a quantised version does: quality barely changes down to 8 bits, more noticeably at 4 bits.

  • INT8recommendedweights 20 GB · ready files25 GB
  • 4-bitweights 11 GB16 GB

Tip: go 8-bit.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: INT8 — from RTX 5090

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.

    # zzb1420/MiniMax-H3-ref2va-V5xs-INT8
    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 zzb1420/MiniMax-H3-ref2va-V5xs-INT8 --local-dir ~/models/MiniMax-H3-ref2va-V5xs-INT8
    # 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.

  • SageAttention 3v2.2.0from October 28, 2025

    even faster on RTX 50 / Blackwell

    FP4 attention for Blackwell. Faster than SageAttention 2, but the authors suggest v2 when accuracy matters.

    How to enable: Build from the sageattention3_blackwell folder of the repo; in ComfyUI via KJNodes.

    Earlier generations (2)
    • SageAttention 2++ — Quantised (INT8/FP8) attention with no visible quality loss. The main speed-up for image and video generation.
    • SageAttention 1 — First version: INT8 for Q·K. Still works, but 2++ is faster and more accurate.
  • 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 (3)
    • 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.
    • xFormers — The older way to speed up attention, works almost everywhere. Sage or Flash are faster where available.
  • cache-dit (DBCache)v1.5.6from November 20, 2026new release

    1.5–3× at nearly the same quality

    Reuses block outputs between diffusion steps. Supports most DiT models.

    How to enable: pip install cache-dit; one line cache_dit.enable_cache(pipe) for diffusers.

    Earlier generations (2)
    • MagCache — Skips steps by magnitude of change. More accurate than TeaCache, no calibration.
    • TeaCache — The first popular diffusion step cache. Works, newer methods are more accurate.
  • LightX2V step distill0.6.1from November 9, 2026new release

    5–10× fewer steps

    Distilled LoRAs and an engine: video in 4–8 steps instead of 30–50.

    How to enable: Load their step-distill LoRA, CFG 1.0, 4–8 steps.

  • 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.

MiniMax H3 prompts

All prompts

Ready prompts with the resulting video: copy and try.

Model family

New posts about versions of MiniMax-H3-ref2va-V5xs-INT8 (quants, finetunes, LoRAs) will show up under the bell.

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

About the model

zzb1420/MiniMax-H3-ref2va-V5xs-INT8
Author
zzb1420
Category
Video
Released
Posted

On NeyroLearn

Hugging Face

01

Ask AI about this model

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

Comments