iceclear/stablesr – Run with an API on Replicate

iceclear / stablesr

Exploiting Diffusion Prior for Real-World Image Super-Resolution

Cold

Public
5.1K runs
A100 (80GB)
GitHub
Paper
License

Iterate in playground

Run with an API

Playground API Examples README Versions

Input

Run this model in Node.js with one line of code:

npx create-replicate --model=iceclear/stablesr

or set up a project from scratch

Install Replicate’s Node.js client library:

npm install replicate

Set the REPLICATE_API_TOKEN environment variable:

export REPLICATE_API_TOKEN=<paste-your-token-here>

Find your API token in your account settings.

Import and set up the client:

import Replicate from "replicate";
import fs from "node:fs";

const replicate = new Replicate({
  auth: process.env.REPLICATE_API_TOKEN,
});

Run iceclear/stablesr using Replicate’s API. Check out the model's schema for an overview of inputs and outputs.

const output = await replicate.run(
  "iceclear/stablesr:01e431cbcf30e2307c61dadca4db988411bc297d35f9bcad8b84ec1462cfe51b",
  {
    input: {
      seed: 42,
      upscale: 4,
      ddpm_steps: 200,
      input_image: "https://replicate.delivery/pbxt/JPdHjT2nuixBTTtKrwMFyy9UE1QqB7Mr0T7baVtgnq7l4ogK/OST_120.png",
      tile_overlap: 32,
      colorfix_type: "adain",
      fidelity_weight: 0.5
    }
  }
);

// To access the file URL:
console.log(output.url()); //=> "http://example.com"

// To write the file to disk:
fs.writeFile("my-image.png", output);

To learn more, take a look at the guide on getting started with Node.js.

Install Replicate’s Python client library:

pip install replicate

Set the REPLICATE_API_TOKEN environment variable:

export REPLICATE_API_TOKEN=<paste-your-token-here>

Find your API token in your account settings.

Import the client:

import replicate

Run iceclear/stablesr using Replicate’s API. Check out the model's schema for an overview of inputs and outputs.

output = replicate.run(
    "iceclear/stablesr:01e431cbcf30e2307c61dadca4db988411bc297d35f9bcad8b84ec1462cfe51b",
    input={
        "seed": 42,
        "upscale": 4,
        "ddpm_steps": 200,
        "input_image": "https://replicate.delivery/pbxt/JPdHjT2nuixBTTtKrwMFyy9UE1QqB7Mr0T7baVtgnq7l4ogK/OST_120.png",
        "tile_overlap": 32,
        "colorfix_type": "adain",
        "fidelity_weight": 0.5
    }
)
print(output)

To learn more, take a look at the guide on getting started with Python.

Set the REPLICATE_API_TOKEN environment variable:

export REPLICATE_API_TOKEN=<paste-your-token-here>

Find your API token in your account settings.

Run iceclear/stablesr using Replicate’s API. Check out the model's schema for an overview of inputs and outputs.

curl -s -X POST \
  -H "Authorization: Bearer $REPLICATE_API_TOKEN" \
  -H "Content-Type: application/json" \
  -H "Prefer: wait" \
  -d $'{
    "version": "iceclear/stablesr:01e431cbcf30e2307c61dadca4db988411bc297d35f9bcad8b84ec1462cfe51b",
    "input": {
      "seed": 42,
      "upscale": 4,
      "ddpm_steps": 200,
      "input_image": "https://replicate.delivery/pbxt/JPdHjT2nuixBTTtKrwMFyy9UE1QqB7Mr0T7baVtgnq7l4ogK/OST_120.png",
      "tile_overlap": 32,
      "colorfix_type": "adain",
      "fidelity_weight": 0.5
    }
  }' \
  https://api.replicate.com/v1/predictions

To learn more, take a look at Replicate’s HTTP API reference docs.

Output

{
  "completed_at": "2023-08-25T15:19:29.174935Z",
  "created_at": "2023-08-25T15:19:06.422027Z",
  "data_removed": false,
  "error": null,
  "id": "vwqsxtrbx64uwtd2aetfzir55a",
  "input": {
    "seed": 42,
    "upscale": 4,
    "ddpm_steps": 200,
    "input_image": "https://replicate.delivery/pbxt/JPdHjT2nuixBTTtKrwMFyy9UE1QqB7Mr0T7baVtgnq7l4ogK/OST_120.png",
    "tile_overlap": 32,
    "colorfix_type": "adain",
    "fidelity_weight": 0.5
  },
  "logs": "Using seed: 42\nGlobal seed set to 42\nloaded input image of size (128, 128) from /tmp/tmpe7rru6n3OST_120.png\ntorch.Size([1, 3, 512, 512])\nSampling t:   0%|          | 0/200 [00:00<?, ?it/s]\nSampling t:   0%|          | 1/200 [00:00<00:21,  9.20it/s]\nSampling t:   1%|          | 2/200 [00:00<00:21,  9.38it/s]\nSampling t:   2%|▏         | 3/200 [00:00<00:21,  9.02it/s]\nSampling t:   2%|▏         | 4/200 [00:00<00:22,  8.90it/s]\nSampling t:   2%|▎         | 5/200 [00:00<00:21,  8.95it/s]\nSampling t:   3%|▎         | 6/200 [00:00<00:22,  8.77it/s]\nSampling t:   4%|▎         | 7/200 [00:00<00:21,  8.97it/s]\nSampling t:   4%|▍         | 9/200 [00:00<00:20,  9.29it/s]\nSampling t:   5%|▌         | 10/200 [00:01<00:20,  9.16it/s]\nSampling t:   6%|▌         | 11/200 [00:01<00:20,  9.19it/s]\nSampling t:   6%|▌         | 12/200 [00:01<00:21,  8.88it/s]\nSampling t:   6%|▋         | 13/200 [00:01<00:20,  9.02it/s]\nSampling t:   8%|▊         | 15/200 [00:01<00:19,  9.47it/s]\nSampling t:   8%|▊         | 16/200 [00:01<00:19,  9.49it/s]\nSampling t:   9%|▉         | 18/200 [00:01<00:18,  9.90it/s]\nSampling t:  10%|▉         | 19/200 [00:02<00:18,  9.66it/s]\nSampling t:  10%|█         | 20/200 [00:02<00:18,  9.50it/s]\nSampling t:  11%|█         | 22/200 [00:02<00:18,  9.88it/s]\nSampling t:  12%|█▏        | 24/200 [00:02<00:17, 10.17it/s]\nSampling t:  13%|█▎        | 26/200 [00:02<00:17,  9.86it/s]\nSampling t:  14%|█▎        | 27/200 [00:02<00:17,  9.68it/s]\nSampling t:  14%|█▍        | 28/200 [00:02<00:17,  9.59it/s]\nSampling t:  15%|█▌        | 30/200 [00:03<00:17,  9.94it/s]\nSampling t:  16%|█▌        | 32/200 [00:03<00:17,  9.81it/s]\nSampling t:  16%|█▋        | 33/200 [00:03<00:17,  9.48it/s]\nSampling t:  17%|█▋        | 34/200 [00:03<00:17,  9.48it/s]\nSampling t:  18%|█▊        | 35/200 [00:03<00:17,  9.50it/s]\nSampling t:  18%|█▊        | 36/200 [00:03<00:17,  9.32it/s]\nSampling t:  18%|█▊        | 37/200 [00:03<00:17,  9.13it/s]\nSampling t:  20%|█▉        | 39/200 [00:04<00:16,  9.69it/s]\nSampling t:  20%|██        | 41/200 [00:04<00:16,  9.92it/s]\nSampling t:  21%|██        | 42/200 [00:04<00:16,  9.74it/s]\nSampling t:  22%|██▏       | 43/200 [00:04<00:16,  9.67it/s]\nSampling t:  22%|██▏       | 44/200 [00:04<00:16,  9.69it/s]\nSampling t:  23%|██▎       | 46/200 [00:04<00:15,  9.76it/s]\nSampling t:  24%|██▎       | 47/200 [00:04<00:16,  9.48it/s]\nSampling t:  24%|██▍       | 48/200 [00:05<00:16,  9.27it/s]\nSampling t:  25%|██▌       | 50/200 [00:05<00:15,  9.62it/s]\nSampling t:  26%|██▌       | 52/200 [00:05<00:15,  9.56it/s]\nSampling t:  26%|██▋       | 53/200 [00:05<00:15,  9.44it/s]\nSampling t:  27%|██▋       | 54/200 [00:05<00:15,  9.47it/s]\nSampling t:  28%|██▊       | 56/200 [00:05<00:14,  9.85it/s]\nSampling t:  28%|██▊       | 57/200 [00:05<00:14,  9.73it/s]\nSampling t:  29%|██▉       | 58/200 [00:06<00:14,  9.59it/s]\nSampling t:  30%|██▉       | 59/200 [00:06<00:14,  9.68it/s]\nSampling t:  30%|███       | 61/200 [00:06<00:13, 10.13it/s]\nSampling t:  32%|███▏      | 63/200 [00:06<00:13, 10.22it/s]\nSampling t:  32%|███▎      | 65/200 [00:06<00:13,  9.88it/s]\nSampling t:  34%|███▎      | 67/200 [00:06<00:13,  9.86it/s]\nSampling t:  34%|███▍      | 69/200 [00:07<00:13, 10.06it/s]\nSampling t:  36%|███▌      | 71/200 [00:07<00:13,  9.90it/s]\nSampling t:  36%|███▌      | 72/200 [00:07<00:13,  9.51it/s]\nSampling t:  37%|███▋      | 74/200 [00:07<00:12,  9.76it/s]\nSampling t:  38%|███▊      | 75/200 [00:07<00:12,  9.66it/s]\nSampling t:  38%|███▊      | 77/200 [00:08<00:12,  9.52it/s]\nSampling t:  39%|███▉      | 78/200 [00:08<00:12,  9.48it/s]\nSampling t:  40%|███▉      | 79/200 [00:08<00:12,  9.47it/s]\nSampling t:  40%|████      | 80/200 [00:08<00:12,  9.29it/s]\nSampling t:  40%|████      | 81/200 [00:08<00:13,  9.01it/s]\nSampling t:  42%|████▏     | 83/200 [00:08<00:12,  9.47it/s]\nSampling t:  42%|████▎     | 85/200 [00:08<00:11,  9.77it/s]\nSampling t:  43%|████▎     | 86/200 [00:08<00:11,  9.69it/s]\nSampling t:  44%|████▍     | 88/200 [00:09<00:11,  9.97it/s]\nSampling t:  45%|████▌     | 90/200 [00:09<00:10, 10.02it/s]\nSampling t:  46%|████▌     | 91/200 [00:09<00:11,  9.72it/s]\nSampling t:  46%|████▌     | 92/200 [00:09<00:11,  9.37it/s]\nSampling t:  46%|████▋     | 93/200 [00:09<00:11,  9.25it/s]\nSampling t:  48%|████▊     | 95/200 [00:09<00:10,  9.74it/s]\nSampling t:  48%|████▊     | 96/200 [00:10<00:10,  9.61it/s]\nSampling t:  48%|████▊     | 97/200 [00:10<00:10,  9.53it/s]\nSampling t:  49%|████▉     | 98/200 [00:10<00:10,  9.41it/s]\nSampling t:  50%|████▉     | 99/200 [00:10<00:10,  9.34it/s]\nSampling t:  50%|█████     | 100/200 [00:10<00:10,  9.45it/s]\nSampling t:  50%|█████     | 101/200 [00:10<00:10,  9.17it/s]\nSampling t:  51%|█████     | 102/200 [00:10<00:11,  8.86it/s]\nSampling t:  52%|█████▏    | 103/200 [00:10<00:11,  8.77it/s]\nSampling t:  52%|█████▏    | 104/200 [00:10<00:10,  8.97it/s]\nSampling t:  52%|█████▎    | 105/200 [00:11<00:10,  8.95it/s]\nSampling t:  53%|█████▎    | 106/200 [00:11<00:10,  8.74it/s]\nSampling t:  54%|█████▎    | 107/200 [00:11<00:10,  8.80it/s]\nSampling t:  54%|█████▍    | 108/200 [00:11<00:10,  9.05it/s]\nSampling t:  55%|█████▍    | 109/200 [00:11<00:10,  9.05it/s]\nSampling t:  55%|█████▌    | 110/200 [00:11<00:10,  8.84it/s]\nSampling t:  56%|█████▌    | 112/200 [00:11<00:09,  9.69it/s]\nSampling t:  57%|█████▋    | 114/200 [00:11<00:08, 10.14it/s]\nSampling t:  58%|█████▊    | 116/200 [00:12<00:08, 10.29it/s]\nSampling t:  59%|█████▉    | 118/200 [00:12<00:08, 10.08it/s]\nSampling t:  60%|█████▉    | 119/200 [00:12<00:08,  9.91it/s]\nSampling t:  60%|██████    | 120/200 [00:12<00:08,  9.67it/s]\nSampling t:  60%|██████    | 121/200 [00:12<00:08,  9.69it/s]\nSampling t:  61%|██████    | 122/200 [00:12<00:08,  9.69it/s]\nSampling t:  62%|██████▏   | 123/200 [00:12<00:07,  9.64it/s]\nSampling t:  62%|██████▏   | 124/200 [00:12<00:08,  9.34it/s]\nSampling t:  62%|██████▎   | 125/200 [00:13<00:08,  9.29it/s]\nSampling t:  63%|██████▎   | 126/200 [00:13<00:08,  9.20it/s]\nSampling t:  64%|██████▎   | 127/200 [00:13<00:07,  9.18it/s]\nSampling t:  64%|██████▍   | 128/200 [00:13<00:07,  9.28it/s]\nSampling t:  64%|██████▍   | 129/200 [00:13<00:07,  9.08it/s]\nSampling t:  65%|██████▌   | 130/200 [00:13<00:07,  8.91it/s]\nSampling t:  66%|██████▌   | 131/200 [00:13<00:08,  8.60it/s]\nSampling t:  66%|██████▌   | 132/200 [00:13<00:08,  8.28it/s]\nSampling t:  66%|██████▋   | 133/200 [00:14<00:08,  8.22it/s]\nSampling t:  67%|██████▋   | 134/200 [00:14<00:08,  8.13it/s]\nSampling t:  68%|██████▊   | 136/200 [00:14<00:07,  9.09it/s]\nSampling t:  69%|██████▉   | 138/200 [00:14<00:06,  9.26it/s]\nSampling t:  70%|██████▉   | 139/200 [00:14<00:06,  8.85it/s]\nSampling t:  70%|███████   | 140/200 [00:14<00:06,  8.65it/s]\nSampling t:  70%|███████   | 141/200 [00:14<00:06,  8.46it/s]\nSampling t:  71%|███████   | 142/200 [00:15<00:06,  8.30it/s]\nSampling t:  72%|███████▏  | 143/200 [00:15<00:06,  8.38it/s]\nSampling t:  72%|███████▏  | 144/200 [00:15<00:06,  8.34it/s]\nSampling t:  72%|███████▎  | 145/200 [00:15<00:06,  8.57it/s]\nSampling t:  73%|███████▎  | 146/200 [00:15<00:06,  8.72it/s]\nSampling t:  74%|███████▍  | 148/200 [00:15<00:05,  9.03it/s]\nSampling t:  75%|███████▌  | 150/200 [00:15<00:05,  9.21it/s]\nSampling t:  76%|███████▌  | 151/200 [00:16<00:05,  9.16it/s]\nSampling t:  76%|███████▌  | 152/200 [00:16<00:05,  9.04it/s]\nSampling t:  77%|███████▋  | 154/200 [00:16<00:04,  9.48it/s]\nSampling t:  78%|███████▊  | 155/200 [00:16<00:04,  9.58it/s]\nSampling t:  78%|███████▊  | 156/200 [00:16<00:04,  9.39it/s]\nSampling t:  78%|███████▊  | 157/200 [00:16<00:04,  9.19it/s]\nSampling t:  79%|███████▉  | 158/200 [00:16<00:04,  8.98it/s]\nSampling t:  80%|███████▉  | 159/200 [00:16<00:04,  9.21it/s]\nSampling t:  80%|████████  | 160/200 [00:17<00:04,  9.32it/s]\nSampling t:  80%|████████  | 161/200 [00:17<00:04,  9.02it/s]\nSampling t:  82%|████████▏ | 163/200 [00:17<00:03,  9.72it/s]\nSampling t:  82%|████████▎ | 165/200 [00:17<00:03,  9.76it/s]\nSampling t:  83%|████████▎ | 166/200 [00:17<00:03,  9.47it/s]\nSampling t:  84%|████████▎ | 167/200 [00:17<00:03,  9.10it/s]\nSampling t:  84%|████████▍ | 169/200 [00:17<00:03,  9.29it/s]\nSampling t:  85%|████████▌ | 170/200 [00:18<00:03,  9.31it/s]\nSampling t:  86%|████████▌ | 171/200 [00:18<00:03,  9.30it/s]\nSampling t:  86%|████████▌ | 172/200 [00:18<00:02,  9.38it/s]\nSampling t:  86%|████████▋ | 173/200 [00:18<00:02,  9.51it/s]\nSampling t:  87%|████████▋ | 174/200 [00:18<00:02,  9.58it/s]\nSampling t:  88%|████████▊ | 175/200 [00:18<00:02,  8.88it/s]\nSampling t:  88%|████████▊ | 176/200 [00:18<00:02,  8.76it/s]\nSampling t:  88%|████████▊ | 177/200 [00:18<00:02,  8.92it/s]\nSampling t:  89%|████████▉ | 178/200 [00:18<00:02,  8.82it/s]\nSampling t:  90%|████████▉ | 179/200 [00:19<00:02,  8.97it/s]\nSampling t:  90%|█████████ | 180/200 [00:19<00:02,  9.14it/s]\nSampling t:  91%|█████████ | 182/200 [00:19<00:01,  9.73it/s]\nSampling t:  92%|█████████▏| 183/200 [00:19<00:01,  9.72it/s]\nSampling t:  92%|█████████▏| 184/200 [00:19<00:01,  9.75it/s]\nSampling t:  92%|█████████▎| 185/200 [00:19<00:01,  9.75it/s]\nSampling t:  93%|█████████▎| 186/200 [00:19<00:01,  9.35it/s]\nSampling t:  94%|█████████▎| 187/200 [00:19<00:01,  9.23it/s]\nSampling t:  94%|█████████▍| 188/200 [00:20<00:01,  8.89it/s]\nSampling t:  94%|█████████▍| 189/200 [00:20<00:01,  9.03it/s]\nSampling t:  96%|█████████▌| 191/200 [00:20<00:00,  9.63it/s]\nSampling t:  96%|█████████▋| 193/200 [00:20<00:00, 10.09it/s]\nSampling t:  97%|█████████▋| 194/200 [00:20<00:00,  9.77it/s]\nSampling t:  98%|█████████▊| 196/200 [00:20<00:00,  9.63it/s]\nSampling t:  98%|█████████▊| 197/200 [00:20<00:00,  9.64it/s]\nSampling t: 100%|█████████▉| 199/200 [00:21<00:00,  9.94it/s]\nSampling t: 100%|██████████| 200/200 [00:21<00:00,  9.92it/s]\nSampling t: 100%|██████████| 200/200 [00:21<00:00,  9.42it/s]",
  "metrics": {
    "predict_time": 22.786526,
    "total_time": 22.752908
  },
  "output": "https://replicate.delivery/pbxt/2j7WjFe95d2lPiElGAl1HMSABEqYkVGT6jM5DFRXsUWAxzuIA/out.png",
  "started_at": "2023-08-25T15:19:06.388409Z",
  "status": "succeeded",
  "urls": {
    "get": "https://api.replicate.com/v1/predictions/vwqsxtrbx64uwtd2aetfzir55a",
    "cancel": "https://api.replicate.com/v1/predictions/vwqsxtrbx64uwtd2aetfzir55a/cancel"
  },
  "version": "01e431cbcf30e2307c61dadca4db988411bc297d35f9bcad8b84ec1462cfe51b"
}

Generated in

22.8 seconds

Tweak itReport

Using seed: 42
Global seed set to 42
loaded input image of size (128, 128) from /tmp/tmpe7rru6n3OST_120.png
torch.Size([1, 3, 512, 512])
Sampling t:   0%|          | 0/200 [00:00<?, ?it/s]
Sampling t:   0%|          | 1/200 [00:00<00:21,  9.20it/s]
Sampling t:   1%|          | 2/200 [00:00<00:21,  9.38it/s]
Sampling t:   2%|▏         | 3/200 [00:00<00:21,  9.02it/s]
Sampling t:   2%|▏         | 4/200 [00:00<00:22,  8.90it/s]
Sampling t:   2%|▎         | 5/200 [00:00<00:21,  8.95it/s]
Sampling t:   3%|▎         | 6/200 [00:00<00:22,  8.77it/s]
Sampling t:   4%|▎         | 7/200 [00:00<00:21,  8.97it/s]
Sampling t:   4%|▍         | 9/200 [00:00<00:20,  9.29it/s]
Sampling t:   5%|▌         | 10/200 [00:01<00:20,  9.16it/s]
Sampling t:   6%|▌         | 11/200 [00:01<00:20,  9.19it/s]
Sampling t:   6%|▌         | 12/200 [00:01<00:21,  8.88it/s]
Sampling t:   6%|▋         | 13/200 [00:01<00:20,  9.02it/s]
Sampling t:   8%|▊         | 15/200 [00:01<00:19,  9.47it/s]
Sampling t:   8%|▊         | 16/200 [00:01<00:19,  9.49it/s]
Sampling t:   9%|▉         | 18/200 [00:01<00:18,  9.90it/s]
Sampling t:  10%|▉         | 19/200 [00:02<00:18,  9.66it/s]
Sampling t:  10%|█         | 20/200 [00:02<00:18,  9.50it/s]
Sampling t:  11%|█         | 22/200 [00:02<00:18,  9.88it/s]
Sampling t:  12%|█▏        | 24/200 [00:02<00:17, 10.17it/s]
Sampling t:  13%|█▎        | 26/200 [00:02<00:17,  9.86it/s]
Sampling t:  14%|█▎        | 27/200 [00:02<00:17,  9.68it/s]
Sampling t:  14%|█▍        | 28/200 [00:02<00:17,  9.59it/s]
Sampling t:  15%|█▌        | 30/200 [00:03<00:17,  9.94it/s]
Sampling t:  16%|█▌        | 32/200 [00:03<00:17,  9.81it/s]
Sampling t:  16%|█▋        | 33/200 [00:03<00:17,  9.48it/s]
Sampling t:  17%|█▋        | 34/200 [00:03<00:17,  9.48it/s]
Sampling t:  18%|█▊        | 35/200 [00:03<00:17,  9.50it/s]
Sampling t:  18%|█▊        | 36/200 [00:03<00:17,  9.32it/s]
Sampling t:  18%|█▊        | 37/200 [00:03<00:17,  9.13it/s]
Sampling t:  20%|█▉        | 39/200 [00:04<00:16,  9.69it/s]
Sampling t:  20%|██        | 41/200 [00:04<00:16,  9.92it/s]
Sampling t:  21%|██        | 42/200 [00:04<00:16,  9.74it/s]
Sampling t:  22%|██▏       | 43/200 [00:04<00:16,  9.67it/s]
Sampling t:  22%|██▏       | 44/200 [00:04<00:16,  9.69it/s]
Sampling t:  23%|██▎       | 46/200 [00:04<00:15,  9.76it/s]
Sampling t:  24%|██▎       | 47/200 [00:04<00:16,  9.48it/s]
Sampling t:  24%|██▍       | 48/200 [00:05<00:16,  9.27it/s]
Sampling t:  25%|██▌       | 50/200 [00:05<00:15,  9.62it/s]
Sampling t:  26%|██▌       | 52/200 [00:05<00:15,  9.56it/s]
Sampling t:  26%|██▋       | 53/200 [00:05<00:15,  9.44it/s]
Sampling t:  27%|██▋       | 54/200 [00:05<00:15,  9.47it/s]
Sampling t:  28%|██▊       | 56/200 [00:05<00:14,  9.85it/s]
Sampling t:  28%|██▊       | 57/200 [00:05<00:14,  9.73it/s]
Sampling t:  29%|██▉       | 58/200 [00:06<00:14,  9.59it/s]
Sampling t:  30%|██▉       | 59/200 [00:06<00:14,  9.68it/s]
Sampling t:  30%|███       | 61/200 [00:06<00:13, 10.13it/s]
Sampling t:  32%|███▏      | 63/200 [00:06<00:13, 10.22it/s]
Sampling t:  32%|███▎      | 65/200 [00:06<00:13,  9.88it/s]
Sampling t:  34%|███▎      | 67/200 [00:06<00:13,  9.86it/s]
Sampling t:  34%|███▍      | 69/200 [00:07<00:13, 10.06it/s]
Sampling t:  36%|███▌      | 71/200 [00:07<00:13,  9.90it/s]
Sampling t:  36%|███▌      | 72/200 [00:07<00:13,  9.51it/s]
Sampling t:  37%|███▋      | 74/200 [00:07<00:12,  9.76it/s]
Sampling t:  38%|███▊      | 75/200 [00:07<00:12,  9.66it/s]
Sampling t:  38%|███▊      | 77/200 [00:08<00:12,  9.52it/s]
Sampling t:  39%|███▉      | 78/200 [00:08<00:12,  9.48it/s]
Sampling t:  40%|███▉      | 79/200 [00:08<00:12,  9.47it/s]
Sampling t:  40%|████      | 80/200 [00:08<00:12,  9.29it/s]
Sampling t:  40%|████      | 81/200 [00:08<00:13,  9.01it/s]
Sampling t:  42%|████▏     | 83/200 [00:08<00:12,  9.47it/s]
Sampling t:  42%|████▎     | 85/200 [00:08<00:11,  9.77it/s]
Sampling t:  43%|████▎     | 86/200 [00:08<00:11,  9.69it/s]
Sampling t:  44%|████▍     | 88/200 [00:09<00:11,  9.97it/s]
Sampling t:  45%|████▌     | 90/200 [00:09<00:10, 10.02it/s]
Sampling t:  46%|████▌     | 91/200 [00:09<00:11,  9.72it/s]
Sampling t:  46%|████▌     | 92/200 [00:09<00:11,  9.37it/s]
Sampling t:  46%|████▋     | 93/200 [00:09<00:11,  9.25it/s]
Sampling t:  48%|████▊     | 95/200 [00:09<00:10,  9.74it/s]
Sampling t:  48%|████▊     | 96/200 [00:10<00:10,  9.61it/s]
Sampling t:  48%|████▊     | 97/200 [00:10<00:10,  9.53it/s]
Sampling t:  49%|████▉     | 98/200 [00:10<00:10,  9.41it/s]
Sampling t:  50%|████▉     | 99/200 [00:10<00:10,  9.34it/s]
Sampling t:  50%|█████     | 100/200 [00:10<00:10,  9.45it/s]
Sampling t:  50%|█████     | 101/200 [00:10<00:10,  9.17it/s]
Sampling t:  51%|█████     | 102/200 [00:10<00:11,  8.86it/s]
Sampling t:  52%|█████▏    | 103/200 [00:10<00:11,  8.77it/s]
Sampling t:  52%|█████▏    | 104/200 [00:10<00:10,  8.97it/s]
Sampling t:  52%|█████▎    | 105/200 [00:11<00:10,  8.95it/s]
Sampling t:  53%|█████▎    | 106/200 [00:11<00:10,  8.74it/s]
Sampling t:  54%|█████▎    | 107/200 [00:11<00:10,  8.80it/s]
Sampling t:  54%|█████▍    | 108/200 [00:11<00:10,  9.05it/s]
Sampling t:  55%|█████▍    | 109/200 [00:11<00:10,  9.05it/s]
Sampling t:  55%|█████▌    | 110/200 [00:11<00:10,  8.84it/s]
Sampling t:  56%|█████▌    | 112/200 [00:11<00:09,  9.69it/s]
Sampling t:  57%|█████▋    | 114/200 [00:11<00:08, 10.14it/s]
Sampling t:  58%|█████▊    | 116/200 [00:12<00:08, 10.29it/s]
Sampling t:  59%|█████▉    | 118/200 [00:12<00:08, 10.08it/s]
Sampling t:  60%|█████▉    | 119/200 [00:12<00:08,  9.91it/s]
Sampling t:  60%|██████    | 120/200 [00:12<00:08,  9.67it/s]
Sampling t:  60%|██████    | 121/200 [00:12<00:08,  9.69it/s]
Sampling t:  61%|██████    | 122/200 [00:12<00:08,  9.69it/s]
Sampling t:  62%|██████▏   | 123/200 [00:12<00:07,  9.64it/s]
Sampling t:  62%|██████▏   | 124/200 [00:12<00:08,  9.34it/s]
Sampling t:  62%|██████▎   | 125/200 [00:13<00:08,  9.29it/s]
Sampling t:  63%|██████▎   | 126/200 [00:13<00:08,  9.20it/s]
Sampling t:  64%|██████▎   | 127/200 [00:13<00:07,  9.18it/s]
Sampling t:  64%|██████▍   | 128/200 [00:13<00:07,  9.28it/s]
Sampling t:  64%|██████▍   | 129/200 [00:13<00:07,  9.08it/s]
Sampling t:  65%|██████▌   | 130/200 [00:13<00:07,  8.91it/s]
Sampling t:  66%|██████▌   | 131/200 [00:13<00:08,  8.60it/s]
Sampling t:  66%|██████▌   | 132/200 [00:13<00:08,  8.28it/s]
Sampling t:  66%|██████▋   | 133/200 [00:14<00:08,  8.22it/s]
Sampling t:  67%|██████▋   | 134/200 [00:14<00:08,  8.13it/s]
Sampling t:  68%|██████▊   | 136/200 [00:14<00:07,  9.09it/s]
Sampling t:  69%|██████▉   | 138/200 [00:14<00:06,  9.26it/s]
Sampling t:  70%|██████▉   | 139/200 [00:14<00:06,  8.85it/s]
Sampling t:  70%|███████   | 140/200 [00:14<00:06,  8.65it/s]
Sampling t:  70%|███████   | 141/200 [00:14<00:06,  8.46it/s]
Sampling t:  71%|███████   | 142/200 [00:15<00:06,  8.30it/s]
Sampling t:  72%|███████▏  | 143/200 [00:15<00:06,  8.38it/s]
Sampling t:  72%|███████▏  | 144/200 [00:15<00:06,  8.34it/s]
Sampling t:  72%|███████▎  | 145/200 [00:15<00:06,  8.57it/s]
Sampling t:  73%|███████▎  | 146/200 [00:15<00:06,  8.72it/s]
Sampling t:  74%|███████▍  | 148/200 [00:15<00:05,  9.03it/s]
Sampling t:  75%|███████▌  | 150/200 [00:15<00:05,  9.21it/s]
Sampling t:  76%|███████▌  | 151/200 [00:16<00:05,  9.16it/s]
Sampling t:  76%|███████▌  | 152/200 [00:16<00:05,  9.04it/s]
Sampling t:  77%|███████▋  | 154/200 [00:16<00:04,  9.48it/s]
Sampling t:  78%|███████▊  | 155/200 [00:16<00:04,  9.58it/s]
Sampling t:  78%|███████▊  | 156/200 [00:16<00:04,  9.39it/s]
Sampling t:  78%|███████▊  | 157/200 [00:16<00:04,  9.19it/s]
Sampling t:  79%|███████▉  | 158/200 [00:16<00:04,  8.98it/s]
Sampling t:  80%|███████▉  | 159/200 [00:16<00:04,  9.21it/s]
Sampling t:  80%|████████  | 160/200 [00:17<00:04,  9.32it/s]
Sampling t:  80%|████████  | 161/200 [00:17<00:04,  9.02it/s]
Sampling t:  82%|████████▏ | 163/200 [00:17<00:03,  9.72it/s]
Sampling t:  82%|████████▎ | 165/200 [00:17<00:03,  9.76it/s]
Sampling t:  83%|████████▎ | 166/200 [00:17<00:03,  9.47it/s]
Sampling t:  84%|████████▎ | 167/200 [00:17<00:03,  9.10it/s]
Sampling t:  84%|████████▍ | 169/200 [00:17<00:03,  9.29it/s]
Sampling t:  85%|████████▌ | 170/200 [00:18<00:03,  9.31it/s]
Sampling t:  86%|████████▌ | 171/200 [00:18<00:03,  9.30it/s]
Sampling t:  86%|████████▌ | 172/200 [00:18<00:02,  9.38it/s]
Sampling t:  86%|████████▋ | 173/200 [00:18<00:02,  9.51it/s]
Sampling t:  87%|████████▋ | 174/200 [00:18<00:02,  9.58it/s]
Sampling t:  88%|████████▊ | 175/200 [00:18<00:02,  8.88it/s]
Sampling t:  88%|████████▊ | 176/200 [00:18<00:02,  8.76it/s]
Sampling t:  88%|████████▊ | 177/200 [00:18<00:02,  8.92it/s]
Sampling t:  89%|████████▉ | 178/200 [00:18<00:02,  8.82it/s]
Sampling t:  90%|████████▉ | 179/200 [00:19<00:02,  8.97it/s]
Sampling t:  90%|█████████ | 180/200 [00:19<00:02,  9.14it/s]
Sampling t:  91%|█████████ | 182/200 [00:19<00:01,  9.73it/s]
Sampling t:  92%|█████████▏| 183/200 [00:19<00:01,  9.72it/s]
Sampling t:  92%|█████████▏| 184/200 [00:19<00:01,  9.75it/s]
Sampling t:  92%|█████████▎| 185/200 [00:19<00:01,  9.75it/s]
Sampling t:  93%|█████████▎| 186/200 [00:19<00:01,  9.35it/s]
Sampling t:  94%|█████████▎| 187/200 [00:19<00:01,  9.23it/s]
Sampling t:  94%|█████████▍| 188/200 [00:20<00:01,  8.89it/s]
Sampling t:  94%|█████████▍| 189/200 [00:20<00:01,  9.03it/s]
Sampling t:  96%|█████████▌| 191/200 [00:20<00:00,  9.63it/s]
Sampling t:  96%|█████████▋| 193/200 [00:20<00:00, 10.09it/s]
Sampling t:  97%|█████████▋| 194/200 [00:20<00:00,  9.77it/s]
Sampling t:  98%|█████████▊| 196/200 [00:20<00:00,  9.63it/s]
Sampling t:  98%|█████████▊| 197/200 [00:20<00:00,  9.64it/s]
Sampling t: 100%|█████████▉| 199/200 [00:21<00:00,  9.94it/s]
Sampling t: 100%|██████████| 200/200 [00:21<00:00,  9.92it/s]
Sampling t: 100%|██████████| 200/200 [00:21<00:00,  9.42it/s]

Examples

View more examples

Run time and cost

This model costs approximately $0.10 to run on Replicate, or 10 runs per $1, but this varies depending on your inputs. It is also open source and you can run it on your own computer with Docker.

This model runs on Nvidia A100 (80GB) GPU hardware. Predictions typically complete within 75 seconds. The predict time for this model varies significantly based on the inputs.

Readme

Exploiting Diffusion Prior for Real-World Image Super-Resolution

Demo on 4K Results

StableSR is capable of achieving arbitrary upscaling in theory, below is a 8x example with a result beyond 4K (5120x3680).

Citation

If our work is useful for your research, please consider citing:

@article{wang2024exploiting,
  author = {Wang, Jianyi and Yue, Zongsheng and Zhou, Shangchen and Chan, Kelvin C.K. and Loy, Chen Change},
  title = {Exploiting Diffusion Prior for Real-World Image Super-Resolution},
  article = {International Journal of Computer Vision},
  year = {2024}
}

License

This project is licensed under NTU S-Lab License 1.0. Redistribution and use should follow this license.

Acknowledgement

This project is based on stablediffusion, latent-diffusion, SPADE, mixture-of-diffusers and BasicSR. Thanks for their awesome work.