Tempo3D: Efficient Temporal-Aware Fine-Tuning and Multi-View Latent Aggregation for 3D Generation
Tempo3D is proposed, a resource-efficient paradigm for high-fidelity VecSet-based 3D generation via efficient temporal-aware fine-tuning and multi-view latent aggregation that outperforms state-of-the-art baselines in generation quality, geometric precision, and editability under a highly efficient computational budget.