New Energy-Guided Flow Matching Boosts Image Gen Efficiency, FID to 1.55
burny_tech · x · 2026-08-09
This research introduces Energy-Guided Flow Matching (EG-FM), improving flow matching by introducing sample-adaptive, coarse-to-fine generative trajectories.
Using a heat-kernel-filtered endpoint, the method progressively reveals high-frequency details based on spectral energy, requiring almost no changes to the backbone or inference cost. On the ImageNet 256x256 class-conditional image generation task, EG-FM achieves lower FID with fewer epochs (1.55 at 200 epochs, 1.45 at 600 epochs). It also scores 0.85 on GenEval for text-to-image generation.
Related event: EG-FM Method Enhances Pixel-Level Image Generation Efficiency(2 posts)→
More from Multimodal
- Kandinsky Lab open-sources S2V 1.0: audio-driven avatar video from one still, MIT licensed — lmoroney · 2026-10-16
- Meme: Flux 3 Klein 5.6B steals the spotlight from Flux 2 Klein 9B — xb1n0ry · 2026-10-16
- Pichai amplifies user demo: Gemini Argon builds Asuka-period pagoda in 11 hours — sundarpichai · 2026-10-16
- Midjourney style code --sref 1127756965 yields bold collage-style illustrations — sergeantsref · 2026-10-16
- A cinematic ocean-to-cliff drone-style prompt worth trying on video models — umesh_ai · 2026-10-16
- Flux.3 klein vs Qwen Image 2.1 turbo on RTX 4070: 10s vs 3s per 1MP edit — Ant_6431 · 2026-10-16