Skip to content
Images

nathanma2

LoRA-стиль для Krea 2

2 min readupdated

Это DreamBooth-LoRA (небольшая дообучаемая надстройка) для генератора изображений Krea 2. Она добавляет к промпту заданный стиль, текстуру и характер рендера: в примерах карточки один и тот же токен применяется к киберпанк-коту, осеннему пейзажу и космической сцене. LoRA обучена на Krea 2 RAW, а примеры показаны на Krea 2 Turbo за 8 шагов. Активируется токеном nma. Пригодится иллюстраторам и авторам промптов, которым нужен готовый стилевой адаптер поверх Krea 2.

Ключевые особенности

  • Триггер nma включает концепт в промпте.
  • Обучена на Krea 2 RAW, демонстрируется на Krea 2 Turbo.
  • Примеры на Turbo сгенерированы за 8 шагов при guidance_scale=0.0.
  • Загружается через diffusers как LoRA-вес ��оверх пайплайна Krea 2.
  • В карточке показаны разные сюжеты: персонаж, пейзаж и фантастическая сцена.

Что можно сделать

  • Добавлять стиль nma к своим сценам — от персонажей до пейзажей.
  • Делать быстрые черновики на Krea 2 Turbo с 8 шагами и затем отбирать удачные варианты.
  • Встроить LoRA в Python-пайплайн на diffusers для пакетной генерации.
  • Проверять один и тот же стиль на разных жанрах и сюжетах, как в примерах карточки.

Пример

import torch
from diffusers import Krea2Pipeline

pipe = Krea2Pipeline.from_pretrained("krea/Krea-2-Turbo", torch_dtype=torch.bfloat16).to("cuda")
pipe.load_lora_weights("mobbrain/nathanma2")
image = pipe("A futuristic cyborg cat with glowing neon circuits sitting on a rainy cyberpunk street, cinematic lighting, nma style.", num_inference_steps=8, guidance_scale=0.0).images[0]
image.save("output.png")
ПодробнееБаза обучения — Krea 2 RAW, генерация в примерах — на Krea 2 Turbo.
В примерах: 8 шагов, guidance_scale=0.0, токен nma.
Данных о VRAM, бенчмарках и ограничениях в карточке нет.
Примеры
Discuss Telegram 0 8

Will it run?

Runs in FP8 / INT8

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

  • BF16weights 25 GB · ready files29 GB
  • FP8 / INT8recommendedweights 12 GB15 GB
  • 4-bitweights 6.9 GB9.1 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 5080 · BF16 — from RTX 5090

Massed Compute hourly prices, November 2026. The discounted price applies with the coupon at checkout.

This is a LoRA: counted together with its base model Krea-2-Raw.

Estimate: weights plus working memory. Long videos, high resolutions and long contexts need more.

Run it on Massed Compute

A ready script for the machine we advise above: paste it and the model downloads and starts.

  1. 1. Rent A30

    $0.26 an hour with the coupon. Pick the Ubuntu image when you create the machine.

    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.

    #!/usr/bin/env bash
    # nathanma2
    set -e
    # NVIDIA driver: CUDA 13 needs 580+. If older or missing, the current nvidia-open is installed (keep yours: KEEP_DRIVER=1)
    DRV=$(nvidia-smi --query-gpu=driver_version --format=csv,noheader 2>/dev/null | head -1 | cut -d. -f1)
    case "$DRV" in ""|*[!0-9]*) DRV=0 ;; esac
    if [ "$DRV" -lt 580 ] && [ -z "${KEEP_DRIVER:-}" ]; then
      . /etc/os-release && \
      curl -fsSLO "https://developer.download.nvidia.com/compute/cuda/repos/ubuntu${VERSION_ID//./}/x86_64/cuda-keyring_1.1-1_all.deb" && \
      sudo dpkg -i cuda-keyring_1.1-1_all.deb && rm -f cuda-keyring_1.1-1_all.deb && \
      sudo apt-get update && \
      sudo DEBIAN_FRONTEND=noninteractive apt-get install -y nvidia-open || \
      { echo "Could not install the NVIDIA driver: see the error above. To keep the current one, run with KEEP_DRIVER=1."; exit 1; }
      echo "NVIDIA driver updated. Reboot the machine (sudo reboot) and run this same script again."
      exit 0
    fi
    curl -LsSf https://astral.sh/uv/install.sh | sh
    source "$HOME/.local/bin/env"
    # CUDA 13+ → int8_convrot (Comfy-Org's advice), else fp8_scaled
    CU=$(nvidia-smi | grep -oP 'CUDA Version: \K[0-9]+' || echo 12)
    [ "$CU" -ge 13 ] && TORCH=cu130 || TORCH=cu128
    [ -d ~/ComfyUI ] || git clone https://github.com/comfyanonymous/ComfyUI ~/ComfyUI
    cd ~/ComfyUI
    uv venv --python 3.12 --allow-existing && source .venv/bin/activate
    uv pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/$TORCH
    uv pip install -r requirements.txt "huggingface_hub[cli]"
    # Base files: Krea 2, 8-bit
    if [ "$CU" -ge 13 ]; then
      hf download Comfy-Org/Krea-2 diffusion_models/krea2_raw_int8_convrot.safetensors text_encoders/qwen3vl_4b_int8_convrot.safetensors vae/qwen_image_vae.safetensors --local-dir models
    else
      hf download Comfy-Org/Krea-2 diffusion_models/krea2_raw_fp8_scaled.safetensors text_encoders/qwen3vl_4b_fp8_scaled.safetensors vae/qwen_image_vae.safetensors --local-dir models
    fi
    # LoRAs (1)
    hf download mobbrain/nathanma2 checkpoint-1500/pytorch_lora_weights.safetensors --local-dir models/loras/mobbrain__nathanma2
    # Start ComfyUI. From your computer: ssh -L 8188:localhost:8188 user@IP, then http://localhost:8188
    python main.py --listen 127.0.0.1 --port 8188
    Add more LoRAs to Krea-2-Raw
  3. 3. Use it

    From your computer: ssh -L 8188:localhost:8188 user@machine-IP, then open http://localhost:8188. Or open that address in the browser on the machine's desktop. The model's workflow is usually in its Hugging Face card.

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.
  • Nunchaku (SVDQuant 4-bit)v1.2.1from January 25, 2026

    ≈3× faster and 3.5× less memory

    4-bit versions of popular image models with almost no quality loss (FP4 on RTX 50).

    How to enable: Ready weights for supported models; in ComfyUI use ComfyUI-nunchaku.

  • 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

Requires the base model

krea/Krea-2-Raw

This model is a LoRA / adapter. Download the base model first, then load this one on top.

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

Quantised versions 12All on Hugging Face
Fine-tunes 12All on Hugging Face
LoRAs and adapters 12All on Hugging Face

About the model

mobbrain/nathanma2
Author
mobbrain
Category
Images
Released
Posted

On NeyroLearn

Hugging Face

80

Ask AI about this model

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

Comments

Images

verbie

verbie