Unet Loader (GGUF)

SamplerCustomAdvanced

DualCLIPLoader

Empty Latent Image

CLIP Text Encode (Prompt)

BasicGuider

LayerUtility: Purge VRAM

Load VAE

Tiled VAE Decode

Random Prompts

{
 __emo__ __adjectives__ 
 __begin__ __cameraanimal__ 
    {
__remington__ __mucha__ |
__whistler__  __remington__ |
__remington__ space cowboy |
__whistler__  __remington__ __rosetti__ space cowgirl |
__whistler__  __remington__ __rosetti__ intergalactic rodeo  |
__whistler__  __remington__ __rosetti__ __meteo__ |
__remington__ space cowboy |
__whistler__ __rosetti__ space cowgirl |
 __remington__ __rosetti__ intergalactic rodeo  |
__whistler__  __remington__ __meteo__ |
     }
}

__end__
32k closeup landscape outdoor award winning photography off center intricate highly detailed cinematic dramatic vivid complementary colors bright stylized beautiful __poses__ 
 


TeaCache For Img Gen

Apply First Block Cache

RandomNoise

BasicScheduler

Save Image

KSamplerSelect

FLUX GGUF with LORA - FASTER

POSITIVE Prompt

MODEL

model

SIGMAS

images

NOISE

GUIDER

SAMPLER

noise

guider

sampler

sigmas

latent_image

output

denoised_output

conditioning

anything

samples

IMAGE

NOTES

CLIP

value

STRING

LATENT

clip

text

CONDITIONING

Jeff Thomann

Flux GGUF w/WAVESPEED OR TEACACHE - switch between them... &nbsp;WAVESPEED is a bit faster. &nbsp;generating in about 42 seconds per image on 8 gb vram, although, results are a bit blurry at times. - simple workflow for low vram machines.Teacahe node in theory speeds up the model a bit. &nbsp;It does help some. &nbsp;I'm getting around 51 seconds to generate an image on 8 gb vram.

Flux GGUF w/(WAVESPEED OR TEACACHE ) - simple workflow for low vram machines.

OpenArt

Workflows

Active Sessions

Flux GGUF w/(WAVESPEED OR TEACACHE ) - simple workflow for low vram machines.

Description

Discussion

Loading...