Skip to content
Video

Frieren H3 LoRA

аниме-стиль и голоса каста для MiniMax H3

2 min readupdated

Rank-16 LoRA для MiniMax H3, которая выучила внешность, состав персонажей и голоса «Sousou no Frieren» на 672 трёхсекундных диалоговых клипах и 721 стилле. Обучение шло сразу на изображениях, видео и аудио с нулевого шага; карточка также содержит полный гайд по сборке датасета и тренировке.

Ключевые особенности

  • Rank 16, alpha 16, targets attention;mlp;audio — аудиослои добавлены, чтобы выучить голоса.
  • Стиль, персонажи и голоса Frieren; для старта промпта нужен триггер 2d anime style, Frieren style,.
  • 672 клипа по 73 кадра (3,04 с) со звуком и 721 стилл 1920×1080; валидация — 30 клипов и 30 стиллов.
  • Три чекпойнта: step4200 рекомендуемый, step3900 — безопасная альтернатива, step3150 — мягкая палитра и лучшее следование свету.
  • Работает на базе H3 FL2VA и на Ref2VA с character sheets как референсами.

Пример

Положите .safetensors в ComfyUI/models/loras/, загрузите на модели H3 при strength 1.0 и начните промпт с 2d anime style, Frieren style,. Например: «Химмель и Фрирен, средний план, мягкий дневной свет»; для Ref2VA можно приложить character sheets.

ПодробнееРекомендуемый step4200 — лучшие лица и чистые линии, самый низкий validation loss.
step3900 — лучший fine-detail и костюмы, step3150 — мягкая палитра и свет.
4 200 шагов (~5,5 эпох) заняли ~14 ч на одной RTX 5090.
Данные: 672 клипа по 73 кадра со звуком + 721 стилл; база — H3 FL2VA.
Примеры и видео
Discuss Telegram 24

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.

    # tazmannner/Frieren-H3-LoRA
    # ===== Tokens: paste yours between the quotes =====
    export HF_TOKEN=""        # huggingface.co/settings/tokens, for gated models
    # ==========================================
    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 tazmannner/Frieren-H3-LoRA --local-dir ~/models/Frieren-H3-LoRA
    # The model is in ~/models. How to run it is in the model card (link above).
    Add more LoRAs to Frieren-H3-LoRA
  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 Frieren-H3-LoRA (quants, finetunes, LoRAs) will show up under the bell.

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

About the model

tazmannner/Frieren-H3-LoRA
Author
tazmannner
Category
Video
Released
Posted

On NeyroLearn

Hugging Face

024

Ask AI about this model

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

Comments