Skip to content
3D

Idol Diffusers

3D Gaussian splat из одного изображения

1 min readupdated

Это Diffusers-конвертация speridlabs/idol от OzzyGT для генерации 3D Gaussian splat по одному изображению. В репозитории только веса и конфиги: denoiser, octree decoder, Gaussian decoder и два энкодера — DINOv3 и FLUX.2 VAE. Размер — около 0.37B параметров.

Ключевые особенности

  • Генерирует 3D Gaussian splat по одному изображению — single-image image-to-3D.
  • Модульный pipeline Idol Modular Diffusers: код и веса разделены между репозиториями.
  • На выходе — активированные Gaussians (positions, colors, scales, rotations, opacities) и предсказанная камера, а не готовый файл.
  • Вход: подготовленное изображение 1024×1024, объект вырезан с запасом и на чёрном фоне.
  • Запуск на CUDA в bfloat16; в примере используется trust_remote_code=True.

Пример

import torch
from diffusers import ModularPipeline
from diffusers.utils import load_image

pipe = ModularPipeline.from_pretrained("OzzyGT/idol_custom_blocks", trust_remote_code=True)
pipe.load_components(dtype=torch.bfloat16)
pipe.to("cuda")
result = pipe(image=load_image("input.png"), generator=torch.Generator("cuda").manual_seed(42), output=["gaussians", "camera_latents"])
ПодробнееВеса: Apache-2.0 от speridlabs/idol; DINOv3 — отдельная лицензия, FLUX.2 VAE — Apache-2.0 без изменений.
Репозиторий содержит только веса и конфиги; для запуска нужен репозиторий блоков OzzyGT/idol_custom_blocks.
Пример в карточке: seed 42, 50 шагов, 262 144 Gaussians; результат сохраняется в .ply через отдельный recipe.
Требования к VRAM в карточке не указаны; trust_remote_code=True выполняет код с Hub — используйте только доверенные репозитории.
Видео-пример
Discuss Telegram 1

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.

    # OzzyGT/idol_diffusers
    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 OzzyGT/idol_diffusers --local-dir ~/models/idol_diffusers
    # The model is in ~/models. How to run it is in the model card (link above).
    Add more LoRAs to idol_diffusers
  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 idol_diffusers (quants, finetunes, LoRAs) will show up under the bell.

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

About the model

OzzyGT/idol_diffusers
Author
OzzyGT
Category
3D
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