New paper: learn CFG guidance schedules adversarially with GAN + RL
kwangmoo_yi · x · 2026-08-18
A new arXiv paper, "Adversarial Learning of Classifier-Free Guidance Schedules" (Pokle, Galashov et al., including DeepMind researchers Doucet and De Bortoli), tackles the classic CFG problem in text-to-image diffusion: a static, global guidance scale is suboptimal across timesteps, samples, and conditions. The authors frame learning time-, condition-, and sample-dependent guidance schedules as density ratio estimation: a discriminator estimates the time-dependent log-density ratio between the true and guided marginals, while a lightweight generator network predicts the optimal state-dependent guidance scale. Empirically, the method beats heuristic CFG schedules and prior dynamic-guidance learning approaches on text-to-image benchmarks. The sharer quips: are we going back to adversarial training?
Related event: DeepMind Paper: Adversarial Learning of CFG Guidance Schedules(2 posts)→
More from Multimodal
- MiniMax Music 3.0 users struggle to prompt dark, severe orchestral music — Silver-Spot-2763 · 2026-08-18
- Minimax H3 blurry video quality: is Turbo LoRA the cause? — Fit_Satisfaction2953 · 2026-08-18
- Artist showcases generative art experiment using pix2pix — memoakten · 2026-08-18
- Released: 80s Italian Street Photography LoRA for Krea 2 — StrainEquivalent996 · 2026-08-18
- MiniMax-H3 Spatial Physics LoRA Generates Realistic Physical Simulations — linoy_tsaban · 2026-08-18
- Seedance 2.5 Introduces Image-like Editing for Videos — CodeByPoonam · 2026-08-18