Idol

Use it to create a 3D Gaussian-splat object from a photo, render, or generated image.

Availability

Hugging Face repository, GitHub SDK, and Hugging Face Spaces demo

Pricing model

Not established

Deployment

Runs locally on an NVIDIA GPU with CUDA

Privacy

Not established

Licence

Apache-2.0

• Idol turns one picture of an object into a 3D Gaussian splat. ^[1]

  Idol turns one picture of an object into a 3D Gaussian splat.

• It generates up to 1,048,576 Gaussians and writes a standard 3DGS `.ply` file. ^[1]

  Give it a render, a generated image or a photo: it removes the background, frames the object and generates up to 1,048,576 Gaussians, including the sides the picture never showed.

• The model weights are available from the `speridlabs/idol` Hugging Face repository. ^[1]

  | `--weights` | `speridlabs/idol` | A Hugging Face repo or a local folder.

• A demo is available on Hugging Face Spaces. ^[1]

  terms with your Hugging Face account, then log in once with `hf auth login` (or set `HF_TOKEN`):

• The documented setup requires Python 3.10+, PyTorch 2.5+, and an NVIDIA GPU with CUDA. ^[1]

  # Python 3.10+, PyTorch 2.5+, an NVIDIA GPU with CUDA

• The project states that it uses 7.2 GB of peak VRAM. ^[1]