AI Paper Atlas

Research map / Image generation / Fast diffusion sampling

DeepWeightFlow: Re-Basined Flow Matching for Generating Neural Network Weights

Saumya Gupta, Scott Biggs, Moritz Laber, Zohair Shafi, Robin Walters, Ayan Paul

ICLR 2026 · Image generation › Fast diffusion sampling · approach: Rectified flow straightening · 5 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add Git Re-Basin canonicalization of training weights when flow model capacity is low (ablation)DeepWeightFlow · MNIST MLP (flow hidden dim 64)accuracy (%): 25.54 -> 57.80Improved
add TransFusion canonicalization when flow model capacity is low (ablation)DeepWeightFlow · CIFAR-10 ViT-Small-192 (flow hidden dim 64)accuracy (%): 12.67 -> 43.13Improved
add Git Re-Basin canonicalization at sufficient flow capacity (ablation)DeepWeightFlow · MNIST MLP (flow hidden dim 512)accuracy (%): 96.19 -> 96.17No clear effect
add Git Re-Basin canonicalization at sufficient capacity (ablation)DeepWeightFlow · CIFAR-10 ResNet-18 (complete weights)accuracy (%): 93.47 -> 93.55No clear effect
add Re-Basin canonicalization (ablation)DeepWeightFlow · STL-10 ResNet-18accuracy (%): 62.50 -> 62.46No clear effect
add TransFusion canonicalization at larger flow capacity (ablation)DeepWeightFlow · CIFAR-10 ViT-Small-192accuracy (%): 82.58 -> 83.07Improved
recalibrate BN running statistics on generated weights instead of copying reference statistics (ablation)DeepWeightFlow · CIFAR-10 ResNet-18 (with Re-Basin)accuracy (%): 19.06 -> 93.05Improved
recalibrate BN running statistics (ablation)DeepWeightFlow · CIFAR-10 ResNet-20 (with Re-Basin)accuracy (%): 17.88 -> 74.57Improved
add incremental PCA compression of training weights (ablation)DeepWeightFlow · CIFAR-10 ResNet-20training time (min): 11.25 -> 1.23Improved
add incremental PCA compression of training weights (ablation)DeepWeightFlow · CIFAR-10 ResNet-20accuracy (%): 75.07 -> 75.96No clear effect
use Gaussian source distribution instead of Kaiming initialization (ablation)DeepWeightFlow · MNIST MLPaccuracy / diversityImproved
use DeepWeightFlow generated weights as transfer-learning initialization instead of FLoWNResNet-18 CIFAR-10 -> STL-10 · STL-10 (zero-shot)accuracy (%): 35.16 -> 48.32Improved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy