OOD-Eval

OOD-Eval: Out-of-Domain Evaluation Prompts for Text-to-3D This repository contains the OOD-Eval dataset, a new collection of challenging out-of-domain (OOD) prompts specifically designed to facilitate rigorous evaluation of text-to-3D generation models. It was introduced in the paper MV-RAG: Retrieval Augmented Multiview Diffusion. This dataset helps assess how well text-to-3D approaches perform on rare or novel concepts, addressing a limitation where models often struggle to… See the full description on the dataset page:

Source: Hugging Face Hub (yosepyossi/OOD-Eval). Metadata imported from the dataset’s Hub tags.