Skip to content

Conversation

Balladie
Copy link
Contributor

adding CFG++ to the samplers are straightforward so did as requested here

checked on Flux Krea / SD3.5 / SDXL that it produces reasonable results

example results from Flux Krea:

dpmpp_2m_sde, cfg=1.0 dpmpp_2m_sde_cfg_pp, cfg=1.0 dpmpp_2m_sde, cfg=1.5 dpmpp_2m_sde_cfg_pp, cfg=1.5
flux_krea_midpoint_cfg1 0 flux_krea_midpoint_cfgpp_cfg1 0 flux_krea_midpoint_cfg1 5 flux_krea_midpoint_cfgpp_cfg1 5

@nestflow
Copy link

nestflow commented Sep 13, 2025

I can confirm both of dpmpp_2m_sde_cfg_pp and dpmpp_2m_sde_heun_cfg_pp works well for SDXL models (NoobAI and IllustriousXL to be specific I tested).

Though, it should be noted that these will not give the same results as res_multistep with CFG++ counterparts, and the reason has been represented in #9444.

@Kosinkadink Kosinkadink added the Good PR This PR looks good to go, it needs comfy's final review. label Sep 16, 2025
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
Good PR This PR looks good to go, it needs comfy's final review.
Projects
None yet
Development

Successfully merging this pull request may close these issues.

4 participants