bytedance/res-adapter 🔢📝❓✓ → ❓

▶️ 1.4K runs 📅 Mar 2024 ⚙️ Cog 0.9.4 🔗 GitHub 📄 Paper ⚖️ License

resolution-adapter stable-diffusion text-to-image

About

Domain Consistent Resolution Adapter for Diffusion Models: generating consistent images with resolutions outside of their trained domain

Example Output

Prompt:

"Female in a fantasy world, Fluid brushwork, Bright colors, Emphasis on light and atmosphere, realism, impressionism, Post-impressionism, light and color, attention to detail, by master of portraiture, by master of realism,,
(realism:1.2) and (naturalism:1.2), (impressionistic:1.2) brushwork, (portraiture:1.2) and (figurative:1.2) focus, (play of light and shadow:1.2), (elegant compositions:1.2) and (balanced:1.2) arrangements, (rich color palette:1.2) and (bold:1.2) use of color, and (masterful technique:1.2) with (attention to detail:1.2),,"

Output

{"with_res_adapter":"https://replicate.delivery/pbxt/nJe82vh2Pfp3a0291qaXM2WYLYufP09tfAeDfjhe6a0MyoMUJA/output.png","without_res_adapter":null}

Performance Metrics

29.22s Prediction Time

29.24s Total Time

All Input Parameters

{
  "seed": 42,
  "width": 1024,
  "height": 1024,
  "prompt": "Female in a fantasy world, Fluid brushwork, Bright colors, Emphasis on light and atmosphere, realism, impressionism, Post-impressionism, light and color, attention to detail, by master of portraiture, by master of realism,,\n(realism:1.2) and (naturalism:1.2), (impressionistic:1.2) brushwork, (portraiture:1.2) and (figurative:1.2) focus, (play of light and shadow:1.2), (elegant compositions:1.2) and (balanced:1.2) arrangements, (rich color palette:1.2) and (bold:1.2) use of color, and (masterful technique:1.2) with (attention to detail:1.2),,",
  "model_name": "dreamlike-art/dreamlike-diffusion-1.0",
  "show_baseline": false,
  "guidance_scale": 7.5,
  "negative_prompt": "NSFW, poor bad amateur assignment cut out ugly",
  "resadapter_alpha": 0.7,
  "num_inference_steps": 25
}

Input Parameters

seed Type: integer: Random seed. Leave blank to randomize the seed
width Type: integerDefault: 1024: Width of output image
height Type: integerDefault: 1024: Height of output image
prompt Type: stringDefault: Award-winning photo of a mystical fox girl fox in a serene forest clearing, sunlight filtering through the trees,ethereal,enchanting,vibrant orange fur,piercing amber eyes,delicate floral crown, flowing gown,surrounded by a gentle breeze, whispering leaves,magical atmosphere,captured by renowned photographer Emily Thompson using a Nikon D850,creating a dreamlike and captivating image: Input prompt
model_name Default: dreamlike-art/dreamlike-diffusion-1.0: Choose a stable diffusion model.
show_baseline Type: booleanDefault: true: Show baseline without res-adapter for comparison.
guidance_scale Type: numberDefault: 7.5Range: 0 - 20: Scale for classifier-free guidance
negative_prompt Type: stringDefault: NSFW, poor bad amateur assignment cut out ugly: Specify things to not see in the output
resadapter_alpha Type: numberDefault: 0.7Range: 0 - 1: Alpha for resadapter
num_inference_steps Type: integerDefault: 25: Number of denoising steps

Output Schema

Example Execution Logs

Using seed: 42
Generating images with res_adapter...
Token indices sequence length is longer than the specified maximum sequence length for this model (167 > 77). Running this sequence through the model will result in indexing errors
The following part of your input was truncated because CLIP can only handle sequences up to 77 tokens: ['( portraiture : 1. 2 ) and ( figurative : 1. 2 ) focus, ( play of light and shadow : 1. 2 ), ( elegant compositions : 1. 2 ) and ( balanced : 1. 2 ) arrangements, ( rich color palette : 1. 2 ) and ( bold : 1. 2 ) use of color, and ( masterful technique : 1. 2 ) with ( attention to detail : 1. 2 ),,']
  0%|          | 0/25 [00:00<?, ?it/s]
  4%|▍         | 1/25 [00:01<00:26,  1.12s/it]
  8%|▊         | 2/25 [00:01<00:15,  1.45it/s]
 12%|█▏        | 3/25 [00:02<00:19,  1.13it/s]
 16%|█▌        | 4/25 [00:03<00:20,  1.03it/s]
 20%|██        | 5/25 [00:04<00:20,  1.03s/it]
 24%|██▍       | 6/25 [00:05<00:20,  1.06s/it]
 28%|██▊       | 7/25 [00:07<00:19,  1.08s/it]
 32%|███▏      | 8/25 [00:08<00:18,  1.09s/it]
 36%|███▌      | 9/25 [00:09<00:17,  1.10s/it]
 40%|████      | 10/25 [00:10<00:16,  1.10s/it]
 44%|████▍     | 11/25 [00:11<00:15,  1.11s/it]
 48%|████▊     | 12/25 [00:12<00:14,  1.11s/it]
 52%|█████▏    | 13/25 [00:13<00:13,  1.11s/it]
 56%|█████▌    | 14/25 [00:14<00:12,  1.11s/it]
 60%|██████    | 15/25 [00:16<00:11,  1.12s/it]
 64%|██████▍   | 16/25 [00:17<00:10,  1.12s/it]
 68%|██████▊   | 17/25 [00:18<00:08,  1.12s/it]
 72%|███████▏  | 18/25 [00:19<00:07,  1.12s/it]
 76%|███████▌  | 19/25 [00:20<00:06,  1.12s/it]
 80%|████████  | 20/25 [00:21<00:05,  1.12s/it]
 84%|████████▍ | 21/25 [00:22<00:04,  1.12s/it]
 88%|████████▊ | 22/25 [00:23<00:03,  1.12s/it]
 92%|█████████▏| 23/25 [00:24<00:02,  1.12s/it]
 96%|█████████▌| 24/25 [00:26<00:01,  1.12s/it]
100%|██████████| 25/25 [00:27<00:00,  1.12s/it]
100%|██████████| 25/25 [00:27<00:00,  1.09s/it]

Version Details

Version ID: ae94f01b3f844c42112078a7b5346d5142bd87a95d8c50bc27cbece8ed267f71
Version Created: April 7, 2024

Run on Replicate →