{"ID":2895107,"CreatedAt":"2026-06-01T04:54:23.091178241Z","UpdatedAt":"2026-06-01T04:54:23.091178241Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2507.09441","arxiv_id":"2507.09441","title":"RectifiedHR: High-Resolution Diffusion via Energy Profiling and Adaptive Guidance Scheduling","abstract":"High-resolution image synthesis with diffusion models often suffers from energy instabilities and guidance artifacts that degrade visual quality. We analyze the latent energy landscape during sampling and propose adaptive classifier-free guidance (CFG) schedules that maintain stable energy trajectories. Our approach introduces energy-aware scheduling strategies that modulate guidance strength over time, achieving superior stability scores (0.9998) and consistency metrics (0.9873) compared to fixed-guidance approaches. We demonstrate that DPM++ 2M with linear-decreasing CFG scheduling yields optimal performance, providing sharper, more faithful images while reducing artifacts. Our energy profiling framework serves as a powerful diagnostic tool for understanding and improving diffusion model behavior.","short_abstract":"High-resolution image synthesis with diffusion models often suffers from energy instabilities and guidance artifacts that degrade visual quality. We analyze the latent energy landscape during sampling and propose adaptive classifier-free guidance (CFG) schedules that maintain stable energy trajectories. Our approach in...","url_abs":"https://arxiv.org/abs/2507.09441","url_pdf":"https://arxiv.org/pdf/2507.09441v2","authors":"[\"Ankit Sanjyal\"]","published":"2025-07-13T01:21:10Z","proceeding":"cs.GR","tasks":"[\"cs.GR\",\"cs.CV\"]","methods":"[\"Diffusion Model\"]","has_code":false}