Skip to content
3D

TripoSplat-diffusers-3d

image-to-3D без remote code

1 min readupdated

Конвертация VAST-AI/TripoSplat в пайплайн diffusers-3d от suvadityamuk. Модель на ~0,37 млрд параметров превращает изображение в Gaussian splat; компоненты оформлены как обычные папки Diffusers и Transformers, поэтому удалённый код не требуется.

Ключевые особенности

  • Пайплайн AutoPipelineForImageTo3D: обычные папки компонентов Diffusers/Transformers, без remote code.
  • Все сети работают в plain PyTorch на CPU или GPU; рендер splats требует опциональный gsplat.
  • DINOv3 и FLUX.2 VAE не включены: грузятся из оригинальных репозиториев на закреплённой ревизии, нужны лицензии и hf auth login.
  • Если у входа нет alpha, фон вырезается BiRefNet; pipeline.disable_background_removal() отключает шаг.
  • Выход — GaussianSplatAsset (SH degree 0, куб [-0.5, 0.5]^3, Z-up); по умолчанию 20 шагов, guidance 3.0, 262144 Gaussians.

Пример

import torch
from diffusers_3d import AutoPipelineForImageTo3D, ImageCondition

pipeline = AutoPipelineForImageTo3D.from_pretrained('suvadityamuk/TripoSplat-diffusers-3d').to('cuda')
splat = pipeline(ImageCondition(image=rgba), generator=torch.Generator('cuda').manual_seed(42)).objects[0]
ПодробнееБенчмарков в карточке нет. Запуск: CPU или GPU на чистом PyTorch; для рендера Gaussian splats нужен опциональный gsplat, для PBR-меша TRELLIS.2 — CuMesh.
DINOv3 и FLUX.2 VAE не входят в репозиторий: нужны лицензии и hf auth login. Веса TripoSplat — MIT; сторонние компоненты — под своими лицензиями.
По умолчанию 20 шагов, guidance 3.0, 262144 Gaussians; без dtype компоненты грузятся в родной точности, для CPU можно dtype=torch.float32.
Discuss Telegram 0 17

Will it run?

Runs in F16

Fits without quantisation.

  • F16weights 0.7 GB · ready files1.3 GB
  • FP8 / INT8recommendedweights 0.3 GB0.9 GB
  • 4-bitweights 0.2 GB0.7 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 2060 · F16 — from RTX 2060

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

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

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.

    # suvadityamuk/TripoSplat-diffusers-3d
    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 suvadityamuk/TripoSplat-diffusers-3d --local-dir ~/models/TripoSplat-diffusers-3d
    # 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 TripoSplat (quants, finetunes, LoRAs) will show up under the bell.

Quantised versions 5All on Hugging Face
Fine-tunes 3All on Hugging Face

About the model

suvadityamuk/TripoSplat-diffusers-3d
Author
suvadityamuk
Category
3D
Released
Posted

On NeyroLearn

Hugging Face

170

Ask AI about this model

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

Comments