cinemasetfree.com
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
Cost
Community model (estimated from hardware time)
Input Parameters
| Name | Type | Description | Default | Constraints |
|---|---|---|---|---|
creativity | number | Denoising strength. 1 means total destruction of the original image | 0.35 | min: 0.1, max: 1 |
format | string | Format of the output. | "jpg" | jpg png |
guess_mode | boolean | In this mode, the ControlNet encoder will try best to recognize the content of the input image even if you remove all prompts. | false | — |
guidance_scale | number | Scale for classifier-free guidance, should be 0. | 0 | min: 0, max: 30 |
hdr | number | HDR improvement over the original image | 0 | min: 0, max: 1 |
image | string (uri) | Control image for scribble controlnet | — | — |
lora_details_strength | number | Strength of the image's details | 1 | min: -5, max: 3 |
lora_sharpness_strength | number | Strength of the image's sharpness. We don't recommend values above 2. | 1.25 | min: -5, max: 10 |
negative_prompt | string | Negative prompt | "teeth, tooth, open mouth, longbody, lowres, bad anatomy, bad hands, missing fingers, extra digit, fewer digits, cropped, worst quality, low quality, mutant" | — |
prompt | string | Prompt for the model | — | — |
resemblance | number | Conditioning scale for controlnet | 0.85 | min: 0, max: 1 |
resolution | integer | Image resolution | 2560 | 2048 2560 4096 |
scheduler | string | Choose a scheduler. | "DDIM" | DDIM DPMSolverMultistep K_EULER_ANCESTRAL K_EULER |
seed | integer | Seed | — | — |
steps | integer | Steps | 8 | — |
creativity number Denoising strength. 1 means total destruction of the original image
0.35 min: 0.1, max: 1 format string Format of the output.
"jpg" guess_mode boolean In this mode, the ControlNet encoder will try best to recognize the content of the input image even if you remove all prompts.
false guidance_scale number Scale for classifier-free guidance, should be 0.
0 min: 0, max: 30 hdr number HDR improvement over the original image
0 min: 0, max: 1 image string Control image for scribble controlnet
lora_details_strength number Strength of the image's details
1 min: -5, max: 3 lora_sharpness_strength number Strength of the image's sharpness. We don't recommend values above 2.
1.25 min: -5, max: 10 negative_prompt string Negative prompt
"teeth, tooth, open mouth, longbody, lowres, bad anatomy, bad hands, missing fingers, extra digit, fewer digits, cropped, worst quality, low quality, mutant" prompt string Prompt for the model
resemblance number Conditioning scale for controlnet
0.85 min: 0, max: 1 resolution integer Image resolution
2560 scheduler string Choose a scheduler.
"DDIM" seed integer Seed
steps integer Steps
8 8e6a54d7b284 Updated: 6/8/2026 651.7K runs