dmitru/sdxl-darky 🖼️🔢📝❓✓ → 🖼️

▶️ 163 runs 📅 Dec 2023 ⚙️ Cog 0.8.6

About

SDXL fine-tuned using my own sketches and photos in a dark, moody style

Example Output

Prompt:

"an abstract psychedelic collage of photos and sketches made in the style of TOK. An imagined magic animal with human face"

Output

Performance Metrics

18.97s Prediction Time

32.61s Total Time

All Input Parameters

{
  "width": 1024,
  "height": 1024,
  "prompt": "an abstract psychedelic collage of photos and sketches made in the style of TOK. An imagined magic animal with human face",
  "refine": "expert_ensemble_refiner",
  "scheduler": "K_EULER",
  "lora_scale": 0.8,
  "num_outputs": 1,
  "guidance_scale": 7.5,
  "apply_watermark": true,
  "high_noise_frac": 0.94,
  "negative_prompt": "",
  "prompt_strength": 0.8,
  "num_inference_steps": 50
}

Input Parameters

mask Type: string: Input mask for inpaint mode. Black areas will be preserved, white areas will be inpainted.
seed Type: integer: Random seed. Leave blank to randomize the seed
image Type: string: Input image for img2img or inpaint mode
width Type: integerDefault: 1024: Width of output image
height Type: integerDefault: 1024: Height of output image
prompt Type: stringDefault: An astronaut riding a rainbow unicorn: Input prompt
refine Default: no_refiner: Which refine style to use
scheduler Default: K_EULER: scheduler
lora_scale Type: numberDefault: 0.6Range: 0 - 1: LoRA additive scale. Only applicable on trained models.
num_outputs Type: integerDefault: 1Range: 1 - 4: Number of images to output.
refine_steps Type: integer: For base_image_refiner, the number of steps to refine, defaults to num_inference_steps
guidance_scale Type: numberDefault: 7.5Range: 1 - 50: Scale for classifier-free guidance
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.
high_noise_frac Type: numberDefault: 0.8Range: 0 - 1: For expert_ensemble_refiner, the fraction of noise to use
negative_prompt Type: stringDefault:: Input Negative Prompt
prompt_strength Type: numberDefault: 0.8Range: 0 - 1: Prompt strength when using img2img / inpaint. 1.0 corresponds to full destruction of information in image
replicate_weights Type: string: Replicate LoRA weights to use. Leave blank to use the default weights.
num_inference_steps Type: integerDefault: 50Range: 1 - 500: Number of denoising steps
disable_safety_checker Type: booleanDefault: false: Disable safety checker for generated images. This feature is only available through the API. See https://replicate.com/docs/how-does-replicate-work#safety

Output Schema

Output

Type: array • Items Type: string • Items Format: uri

Example Execution Logs

Using seed: 4087
Ensuring enough disk space...
Free disk space: 2069769101312
Downloading weights: https://replicate.delivery/pbxt/5eL80vr8miy4D69h1xYYCzy1YsDoiVMkw50B0VTjPSuJxGfRA/trained_model.tar
b'Downloaded 186 MB bytes in 2.319s (80 MB/s)\nExtracted 186 MB in 0.056s (3.3 GB/s)\n'
Downloaded weights in 2.7585837841033936 seconds
Loading fine-tuned model
Does not have Unet. assume we are using LoRA
Loading Unet LoRA
Prompt: an abstract psychedelic collage of photos and sketches made in the style of <s0><s1>. An imagined magic animal with human face
txt2img mode
  0%|          | 0/47 [00:00<?, ?it/s]
  2%|▏         | 1/47 [00:00<00:12,  3.64it/s]
  4%|▍         | 2/47 [00:00<00:12,  3.65it/s]
  6%|▋         | 3/47 [00:00<00:12,  3.65it/s]
  9%|▊         | 4/47 [00:01<00:11,  3.64it/s]
 11%|█         | 5/47 [00:01<00:11,  3.65it/s]
 13%|█▎        | 6/47 [00:01<00:11,  3.64it/s]
 15%|█▍        | 7/47 [00:01<00:10,  3.64it/s]
 17%|█▋        | 8/47 [00:02<00:10,  3.64it/s]
 19%|█▉        | 9/47 [00:02<00:10,  3.64it/s]
 21%|██▏       | 10/47 [00:02<00:10,  3.64it/s]
 23%|██▎       | 11/47 [00:03<00:09,  3.64it/s]
 26%|██▌       | 12/47 [00:03<00:09,  3.64it/s]
 28%|██▊       | 13/47 [00:03<00:09,  3.64it/s]
 30%|██▉       | 14/47 [00:03<00:09,  3.64it/s]
 32%|███▏      | 15/47 [00:04<00:08,  3.63it/s]
 34%|███▍      | 16/47 [00:04<00:08,  3.63it/s]
 36%|███▌      | 17/47 [00:04<00:08,  3.63it/s]
 38%|███▊      | 18/47 [00:04<00:07,  3.63it/s]
 40%|████      | 19/47 [00:05<00:07,  3.63it/s]
 43%|████▎     | 20/47 [00:05<00:07,  3.63it/s]
 45%|████▍     | 21/47 [00:05<00:07,  3.63it/s]
 47%|████▋     | 22/47 [00:06<00:06,  3.63it/s]
 49%|████▉     | 23/47 [00:06<00:06,  3.63it/s]
 51%|█████     | 24/47 [00:06<00:06,  3.63it/s]
 53%|█████▎    | 25/47 [00:06<00:06,  3.63it/s]
 55%|█████▌    | 26/47 [00:07<00:05,  3.62it/s]
 57%|█████▋    | 27/47 [00:07<00:05,  3.62it/s]
 60%|█████▉    | 28/47 [00:07<00:05,  3.62it/s]
 62%|██████▏   | 29/47 [00:07<00:04,  3.62it/s]
 64%|██████▍   | 30/47 [00:08<00:04,  3.62it/s]
 66%|██████▌   | 31/47 [00:08<00:04,  3.62it/s]
 68%|██████▊   | 32/47 [00:08<00:04,  3.62it/s]
 70%|███████   | 33/47 [00:09<00:03,  3.62it/s]
 72%|███████▏  | 34/47 [00:09<00:03,  3.62it/s]
 74%|███████▍  | 35/47 [00:09<00:03,  3.62it/s]
 77%|███████▋  | 36/47 [00:09<00:03,  3.62it/s]
 79%|███████▊  | 37/47 [00:10<00:02,  3.62it/s]
 81%|████████  | 38/47 [00:10<00:02,  3.62it/s]
 83%|████████▎ | 39/47 [00:10<00:02,  3.62it/s]
 85%|████████▌ | 40/47 [00:11<00:01,  3.62it/s]
 87%|████████▋ | 41/47 [00:11<00:01,  3.62it/s]
 89%|████████▉ | 42/47 [00:11<00:01,  3.62it/s]
 91%|█████████▏| 43/47 [00:11<00:01,  3.62it/s]
 94%|█████████▎| 44/47 [00:12<00:00,  3.62it/s]
 96%|█████████▌| 45/47 [00:12<00:00,  3.62it/s]
 98%|█████████▊| 46/47 [00:12<00:00,  3.62it/s]
100%|██████████| 47/47 [00:12<00:00,  3.62it/s]
100%|██████████| 47/47 [00:12<00:00,  3.63it/s]
  0%|          | 0/3 [00:00<?, ?it/s]
 33%|███▎      | 1/3 [00:00<00:00,  4.23it/s]
 67%|██████▋   | 2/3 [00:00<00:00,  4.20it/s]
100%|██████████| 3/3 [00:00<00:00,  4.19it/s]
100%|██████████| 3/3 [00:00<00:00,  4.19it/s]

Version Details

Version ID: c330f29c7249e5bad6b07144c6a329d4a2ac925b6687392f3a2aa724cc1491ed
Version Created: December 2, 2023

Run on Replicate →