Create temporal consistent video with no flickering, without using face swap. Utilizing sparse control and openpose controlnet models, along with animatediff and others.Detailer and upscale options for fine tuning.

CONSISTENT VIDEO with sparse & openpose

Successfully tested with 1.5 and sdxl models. Selecting ALL models accordingly. Great results with Dreamshaper_8LCM, used in posted video. RealitiesEdgeXLLCM_TURBOXL has also generated good results.

Files with SHA256: b3ba1d5011abf024acf25054fc0e90f867d474a5fcf0563efea287c5ffc89b52