Note, only relatively significant changes to user-visible functionality will be included here. Most recent changes at the top.
- Added the ability to mask both guidance and global changes. See the README section on masks.
- Added the
DiffuseHighParam
node and the ability to connect multiple VAEs, upscale models, noise generators and samplers as well as switch between them per iteration.
sigma_offset
YAML parameter removed - you can use schedule overrides to accomplish the same effect (see README).- Chunked sampling mode added, should make samplers that care about state (i.e. momentum or history like
dpmpp_2m
) work better for guidance steps. May change seeds, you can disable withchunked_sampling: false
in YAML parameters. - Added
sigma_dishonesty_factor
andsigma_dishonesty_factor_guidance
YAML parameters - can be used to increase detail. See README.
- Initial support for rectified flow models (Flux, SD3, SD3.5). Might slightly change seeds for other models.
- Improve contrast adaptive sharpening (hopefully). Will change seeds for workflows using
sharpen_mode: contrast_adaptive
, you can usesharpen_mode: contrast_adaptive_raw
for the old behavior.
- Added workaround for comfyanonymous/ComfyUI#2833 - may change seeds. You can disable it with
seed_rng: false
in YAML parameters. - Added
enable_gc
andenable_cache_clearing
options to YAML parameters. - Added
use_upscale_model
option to YAML parameters - allows disabling use of an upscale model even when connected. - Made
highres_sigmas
input optional, now defaults to the same schedule as the official implementation. - Added
schedule_override
option to YAML parameters, allows internally generating a schedule. See README for details.