fermatresearch/sdxl-outpainting-lora 🔢🖼️📝❓✓ → 🖼️

▶️ 112.3K runs 📅 Nov 2023 ⚙️ Cog 0.9.8 🔗 GitHub
image-editing image-outpainting

About

An improved outpainting model that supports LoRA urls. This model uses PatchMatch to improve the mask quality.

Example Output

Prompt:

"beautiful european city with dramatic light"

Output

Example output

Performance Metrics

47.39s Prediction Time
439.49s Total Time
All Input Parameters
{
  "image": "https://replicate.delivery/pbxt/L2nC0t0m7YTGI4vUfOxZUSHgKNilHMJUVO67rqC8qo5lYC9J/fermat_app_a_realistic_image_of_a_street_from_paris_e9db015b-1068-4c6c-9884-0265d939dcb2.png",
  "prompt": "beautiful european city with dramatic light",
  "scheduler": "K_EULER",
  "lora_scale": 0.8,
  "num_outputs": 1,
  "outpaint_up": 0,
  "outpaint_down": 0,
  "outpaint_left": 256,
  "guidance_scale": 7.5,
  "outpaint_right": 256,
  "apply_watermark": false,
  "condition_scale": 0.25,
  "negative_prompt": ""
}
Input Parameters
seed Type: integer
Random seed. Leave blank to randomize the seed
image Type: string
Input image to inpaint
prompt Type: stringDefault: An astronaut riding a rainbow unicorn
Input prompt
scheduler Default: K_EULER
scheduler
lora_scale Type: numberDefault: 0.8Range: 0 - 1
LoRA additive scale. Only applicable on trained models.
num_outputs Type: integerDefault: 1Range: 1 - 4
Number of images to output
outpaint_up Type: integerDefault: 0Range: 0 - 512
How many pixels the mask should grow in the up direction
lora_weights Type: string
Replicate LoRA weights to use. Leave blank to use the default weights.
outpaint_down Type: integerDefault: 0Range: 0 - 512
How many pixels the mask should grow in the down direction
outpaint_left Type: integerDefault: 0Range: 0 - 512
How many pixels the mask should grow in the left direction
guidance_scale Type: numberDefault: 7.5Range: 1 - 50
Scale for classifier-free guidance
outpaint_right Type: integerDefault: 0Range: 0 - 512
How many pixels the mask should grow in the right direction
apply_watermark Type: booleanDefault: true
Applies a watermark to enable determining if an image is generated in downstream applications. If you have other provisions for generating or deploying images safely, you can use this to disable watermarking.
condition_scale Type: numberDefault: 0.15Range: 0 - 1
The bigger this number is, the more ControlNet interferes
negative_prompt Type: stringDefault:
Input Negative Prompt
Output Schema

Output

Type: arrayItems Type: stringItems Format: uri

Example Execution Logs
Using seed: 34499
Applying smart preprocessing...
Running PatchMatch
PatchMatch completed, time taken: 12332.704345703125 ms
Running PatchMatch
PatchMatch completed, time taken: 21385.12060546875 ms
  0%|          | 0/19 [00:00<?, ?it/s]
  5%|▌         | 1/19 [00:00<00:10,  1.77it/s]
 11%|█         | 2/19 [00:00<00:07,  2.42it/s]
 16%|█▌        | 3/19 [00:01<00:07,  2.27it/s]
 21%|██        | 4/19 [00:01<00:06,  2.20it/s]
 26%|██▋       | 5/19 [00:02<00:06,  2.16it/s]
 32%|███▏      | 6/19 [00:02<00:06,  2.14it/s]
 37%|███▋      | 7/19 [00:03<00:05,  2.13it/s]
 42%|████▏     | 8/19 [00:03<00:05,  2.12it/s]
 47%|████▋     | 9/19 [00:04<00:04,  2.11it/s]
 53%|█████▎    | 10/19 [00:04<00:04,  2.11it/s]
 58%|█████▊    | 11/19 [00:05<00:03,  2.10it/s]
 63%|██████▎   | 12/19 [00:05<00:03,  2.10it/s]
 68%|██████▊   | 13/19 [00:06<00:02,  2.10it/s]
 74%|███████▎  | 14/19 [00:06<00:02,  2.10it/s]
 79%|███████▉  | 15/19 [00:07<00:01,  2.10it/s]
 84%|████████▍ | 16/19 [00:07<00:01,  2.09it/s]
 89%|████████▉ | 17/19 [00:08<00:00,  2.09it/s]
 95%|█████████▍| 18/19 [00:08<00:00,  2.09it/s]
100%|██████████| 19/19 [00:08<00:00,  2.09it/s]
100%|██████████| 19/19 [00:08<00:00,  2.12it/s]
Version Details
Version ID
a542ccf352995f3c41f0bcfaef641daa3058bf2b00e08e04feb0295334ab9804
Version Created
June 4, 2024
Run on Replicate →