HyperNet Fields: Efficiently Training Hypernetworks without Ground Truth by Learning Weight Trajectories

TL;DR

This paper introduces HyperNet Fields, a novel approach to train hypernetworks by modeling the entire weight training trajectory as a neural field, eliminating the need for ground truth weights for each sample.

Contribution

The method enables training hypernetworks without per-sample ground truth by learning weight trajectories as neural fields, simplifying training and broadening applicability.

Findings

Achieves competitive results in personalized image generation.

Demonstrates effective 3D shape reconstruction from images and point clouds.

Eliminates the need for ground truth weights in hypernetwork training.

Abstract

To efficiently adapt large models or to train generative models of neural representations, Hypernetworks have drawn interest. While hypernetworks work well, training them is cumbersome, and often requires ground truth optimized weights for each sample. However, obtaining each of these weights is a training problem of its own-one needs to train, e.g., adaptation weights or even an entire neural field for hypernetworks to regress to. In this work, we propose a method to train hypernetworks, without the need for any per-sample ground truth. Our key idea is to learn a Hypernetwork `Field` and estimate the entire trajectory of network weight training instead of simply its converged state. In other words, we introduce an additional input to the Hypernetwork, the convergence state, which then makes it act as a neural field that models the entire convergence pathway of a task network. A critical benefit in doing so is that the gradient of the estimated weights at any convergence state must then match the gradients of the original task -- this constraint alone is sufficient to train the Hypernetwork Field. We demonstrate the effectiveness of our method through the task of personalized image generation and 3D shape reconstruction from images and point clouds, demonstrating competitive results without any per-sample ground truth.