This dataset contains 100 original procedurally modeled abstract objects, each with a unique nonce name and a fixed, entirely fictional maker record. It provides 10,000 RGB images at 512×512, one portable GLB mesh and one preview per object, and paired data for visual understanding and text-to-image generation. The objects and all maker records are newly constructed for this dataset; no CLEAR identities, biographies, images, or external meshes are included. Each object has 60 training views, 20 held-out camera views, and 20 held-out appearance views with changed background and lighting. A held-out appearance view may share a camera direction with training. The same geometric object appears…
Independent publisher
Chengyang
xzz789
Models in Library0
Datasets in Library1
Models on Hugging Face2
Followers—
Compare the published evidence
This profile connects the 1 model and dataset record shown below to their upstream cards and sources. The library counts describe SAVRN's coverage, rather than this publisher's entire output.
Published tasks: Image to text.
Compare model specifications, source dates, and hardware estimates on the individual records. Publisher identity verification is separate from a model's quality, deployment readiness, or suitability for your work.