← Back to all generators

fermatresearch/high-resolution-controlnet-tile

UPDATE: new upscaling algorithm for a much improved image quality. Fermat.app open-source implementation of an efficient ControlNet 1.1 tile for high-quality upscales. Increase the creativity to encourage hallucination.

Capabilities

Reference ImagesNegative PromptSeed

Cost

Community model (estimated from hardware time)

Input Parameters

creativitynumber

Denoising strength. 1 means total destruction of the original image

Default: 0.35min: 0.1, max: 1
formatstring

Format of the output.

Default: "jpg"
jpgpng
guess_modeboolean

In this mode, the ControlNet encoder will try best to recognize the content of the input image even if you remove all prompts.

Default: false
guidance_scalenumber

Scale for classifier-free guidance, should be 0.

Default: 0min: 0, max: 30
hdrnumber

HDR improvement over the original image

Default: 0min: 0, max: 1
imagestring

Control image for scribble controlnet

lora_details_strengthnumber

Strength of the image's details

Default: 1min: -5, max: 3
lora_sharpness_strengthnumber

Strength of the image's sharpness. We don't recommend values above 2.

Default: 1.25min: -5, max: 10
negative_promptstring

Negative prompt

Default: "teeth, tooth, open mouth, longbody, lowres, bad anatomy, bad hands, missing fingers, extra digit, fewer digits, cropped, worst quality, low quality, mutant"
promptstring

Prompt for the model

resemblancenumber

Conditioning scale for controlnet

Default: 0.85min: 0, max: 1
resolutioninteger

Image resolution

Default: 2560
204825604096
schedulerstring

Choose a scheduler.

Default: "DDIM"
DDIMDPMSolverMultistepK_EULER_ANCESTRALK_EULER
seedinteger

Seed

stepsinteger

Steps

Default: 8
Version: 8e6a54d7b284Updated: 7/25/2026651.7K runs