Skip to content
Other

Gabriel Dropout LoRA для MOSS-Image-Base

аниме-стиль в rank 64

1 min read

Это bf16 style LoRA от quarterturn для модели MOSS-Image-Base, сделанная под изображения в стиле Gabriel Dropout. Обучена на 683 аниме-кадрах из сериала: 12 000 шагов через ai-toolkit.

Ключевые особенности

  • Rank 64, bf16; базовая модель — MOSS-Image-Base.
  • Датасет: 683 аниме-кадра из Gabriel Dropout; captions в JSON-формате.
  • Обучение через ai-toolkit: 12 000 шагов, batch_size 1, gradient accumulation 1.
  • Тренировался только UNet, text encoder заморожен; flowmatch, adamw8bit, lr 1e-4, constant scheduler, gradient checkpointing.
  • Имена для промптов: gabriel, haniel, machiko, master, raphiel, satanichia, tanaka, tapris, vignette — по карточке, должны давать сходство с персонажем и состоянием одежды.
  • Для данных использовали ccip_grounding с face detection, downsampling до ~100 репрезентаций на персонажа и ollama-captioner на Strata и Qwen3.8-flash-next.

Пример

Промпт: gabriel, raphiel, satanichia — имена из обучения. Добавьте описание одежды, чтобы получить нужный образ: карточка указывает, что имя и state of dress должны влиять на сходство.

ПодробнееБенчмарков и данных о VRAM в карточке нет.
12 000 шагов, batch_size 1, bf16, rank 64; обучался только UNet, text encoder заморожен.
Оптимизатор adamw8bit, lr 1e-4, constant scheduler, flowmatch, gradient checkpointing.
Датасет — 683 стоп-кадра Gabriel Dropout; список персонажей ограничен именами из карточки.
Discuss Telegram 1 115

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.

    # quarterturn/moss-image-base-gabriel-dropout-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 quarterturn/moss-image-base-gabriel-dropout-lora --local-dir ~/models/moss-image-base-gabriel-dropout-lora
    # 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.

  • 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 moss-image-base-gabriel-dropout-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

quarterturn/moss-image-base-gabriel-dropout-lora
Author
quarterturn
Category
Other
Released
Posted

On NeyroLearn

Hugging Face

1151

Ask AI about this model

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

Comments