We introduce HyperFields, a method for generating text-conditioned Neural Radiance Fields (NeRFs) with a single forward pass and (optionally) some fine-tuning. Key to our approach are: (i) a dynamic hypernetwork, which learns a smooth mapping from text token embeddings to the space of NeRFs; (ii) NeRF distillation training, which distills scenes encoded in individual NeRFs into one dynamic hypernetwork. These techniques enable a single network to fit over a hundred unique scenes. We further demonstrate that HyperFields learns a more general map between text and NeRFs, and consequently is capable of predicting novel in-distribution and out-of-distribution scenes -- either zero-shot or with a few finetuning steps. Finetuning HyperFields benefits from accelerated convergence thanks to the learned general map, and is capable of synthesizing novel scenes 5 to 10 times faster than existing neural optimization-based methods. Our ablation experiments show that both the dynamic architecture and NeRF distillation are critical to the expressivity of HyperFields.
@article{arxiv.2310.17075,
title = {HyperFields: Towards Zero-Shot Generation of NeRFs from Text},
author = {Sudarshan Babu and Richard Liu and Avery Zhou and Michael Maire and Greg Shakhnarovich and Rana Hanocka},
journal= {arXiv preprint arXiv:2310.17075},
year = {2024}
}
Comments
Accepted to ICML 2024, Project page: https://threedle.github.io/hyperfields/