Workflow name: Flux generates consistent multi-views of game characters
[Workflow introduction]
Open the workflow, first upload an image with multiple poses, and then describe the posture of this image in text at the prompt word on the left. Use other text prompt words combined with MJ expansion and Tile model to control the character style and posture. Generate these key information on one image, which well maintains the consistency of the character. The flux model also perfectly solves the effect of hand and face collapse. The effect is stable, so try it now!

[Usage scenario]
Generate the image effect of the style you want based on your favorite pose, achieve character consistency, and greatly improve creative inspiration.

[Key nodes]
SetUnionControlNetType, ControlNetApply SD3 and HunyuanDiT

[Model version]
FLUX
Model name: flux1-dev-Q4 0.gguf

[LoRA model]
Model name: filmfotos.safetensors

[ControlNet application]
FLUX.1-dev-ControlNet-Union-Pro-Shakker-Labs.safetensors
Tile tile preprocessor
ControlNetApply SD3 and HunyuanDiT
Start time: 0
End time: 0.35
Intensity: 0.53

[K sampler]
CFG: 1
Sampling method: euler
Scheduler: normal
Noise reduction: 1