Skip to content
Video

MiniMax H3 INT4 ConvRot

запуск на 12 ГБ VRAM

1 min read

Полная квантизованная версия модели MiniMax H3 от автора MM744, переведённая в формат INT4 ConvRot (сжатие весов до 4 бит с сохранением архитектуры). Конвертация значительно уменьшает размер файлов и снижает нагрузку на видеопамять и оперативную память без потери основных возможностей оригинала. Модель ориентирована на владельцев видеокарт с 12 ГБ VRAM, которым не хватает ресурсов для запуска полноразмерного чекпоинта.

Ключевые особенности

  • Сжатие весов в формат INT4 ConvRot при сохранении исходной архитектуры MiniMax H3.
  • Существенное снижение требований к объёму VRAM (от 12 ГБ) и оперативной памяти.
  • Уменьшенный объём передачи данных в памяти, что ускоряет загрузку и инференс на совместимом железе.
  • Возможны незначительные отличия в качестве генерации по сравнению с полноразмерной версией.

Что можно сделать

  • Запустить генерацию на базе MiniMax H3 локально на потребительских видеокартах с 12 ГБ VRAM.
  • Сократить время загрузки чекпоинта и ускорить работу пайплайнов на совместимых GPU.
  • Использовать как облегчённую альтернативу оригиналу для тестов и повседневных задач без апгрейда железа.

Пример

Загрузите чекпоинт MiniMax H3 INT4 ConvRot в совместимый локальный интерфейс для генерации на базе MiniMax H3 на системе с 12 ГБ VRAM — триггер-слова для активации не требуются.

ПодробнееРекомендованный объём видеопамяти: 12 ГБ VRAM.
Формат квантизации: INT4 ConvRot.
Базовая модель: MiniMax H3.
Discuss 124 5.1K

Will it run?

Working out the memory needs

The bot is still collecting this model's weight sizes; the estimate will appear here in a few minutes. Meanwhile you can rent a GPU with headroom.

Massed ComputeOpen Massed Compute

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.

    # civitai/2830162-minimax-h3-int4-convrot-12gb-vram
    # ===== Tokens: paste yours between the quotes =====
    export CIVITAI_TOKEN=""   # civitai.com/user/account → API Keys, for Civitai LoRAs
    # ==========================================
    mkdir -p ~/models
    vid=$(curl -fsSL "https://civitai.com/api/v1/models/2830162" | python3 -c 'import json,sys;print(json.load(sys.stdin)["modelVersions"][0]["id"])')
    cd ~/models && curl -fL -OJ -H "Authorization: Bearer $CIVITAI_TOKEN" "https://civitai.com/api/download/models/$vid"
    # 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

No known relatives of this model yet.

About the model

civitai/2830162-minimax-h3-int4-convrot-12gb-vram
Author
MM744
Category
Video
Released
Posted

On NeyroLearn

Civitai

5.1K124

Ask AI about this model

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

Comments