Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

Learning to Solve Generative ODEs Beyond the Linear Span

About

Diffusion and flow generative models sample by integrating a learned ODE, but high quality still requires many sequential model evaluations. Solver learning reduces this cost by adapting scalar coefficients, timesteps, or both, while keeping the backbone model fixed. In this work, we identify a structural bottleneck in this update family: each step remains span-limited. Since the scalar-coefficient update lies in the span of buffered velocity evaluations, it can fit only the in-span component while leaving any out-of-span residual unreachable by scalar recombination alone. We propose SpanLift, a lightweight neural solver that augments scalar-coefficient updates with a spatial residual operator. SpanLift keeps a fixed base solver as an in-span prior and learns a spatial residual operator over the state and velocity buffer. The operator is trained by endpoint teacher matching, preserves the pretrained backbone, and adds no model NFEs. Empirically, the learned correction transfers across base solvers and is predominantly out-of-span. Across pixel-space diffusion, latent flow matching, and precipitation nowcasting, SpanLift achieves state-of-the-art few-step sampling. With only 3 NFE, it improves CIFAR-10 FID from 8.16 to 5.69 and ImageNet FID from 17.37 to 11.83.

Sihyeon Kim, Seunghun Lee, Vikas Singh, Hyunwoo J. Kim• 2026

Related benchmarks

TaskDatasetResultRank
Conditional Image GenerationImageNet 64x64
FID3.78
149
Unconditional Image GenerationCIFAR-10 32 x 32
FID2.41
112
Unconditional Image GenerationFFHQ 64 x 64
FID3.11
84
Image GenerationAFHQ 64x64 v2 (unconditional)
FID2.42
54
Unconditional GenerationAFHQ 64 x 64 v2
FID2.42
48
Unconditional Image GenerationFFHQ unconditional 64x64
FID3.11
45
Precipitation nowcastingSEVIR (test)
CSI (16)63.12
23
Text-to-Image GenerationMJHQ (25K)
FID7.51
9
Showing 8 of 8 rows

Other info

Follow for update