martintmv-git/sdxl-cinematic – Run with an API on Replicate

martintmv-git / sdxl-cinematic

SDXL Fine-tune on cinematic shots

Warm

Public
348 runs
L40S
SDXL fine-tune

Iterate in playground

Run with an API

Playground API Examples README Versions

Input

prompt

string

Shift + Return to add a new line

A medieval knight standing in a misty forest, with a castle visible in the distance, in the style of TOK, crisp, sharp, photorealistic, movie sceneA medieval knight standing in a misty forest, with a castle visible in the distance, in the style of TOK, crisp, sharp, photorealistic, movie scene

Input prompt

Default: "An astronaut riding a rainbow unicorn"

negative_prompt

string

Shift + Return to add a new line

cropped, worst quality, low quality, glitch, deformed, mutated, disfiguredcropped, worst quality, low quality, glitch, deformed, mutated, disfigured

Input Negative Prompt

Default: ""

image

file

Input image for img2img or inpaint mode

mask

file

Input mask for inpaint mode. Black areas will be preserved, white areas will be inpainted.

width

integer

Width of output image

Default: 1024

height

integer

Height of output image

Default: 1024

num_outputs

integer

(minimum: 1, maximum: 4)

Number of images to output.

Default: 1

scheduler

string

scheduler

Default: "K_EULER"

num_inference_steps

integer

(minimum: 1, maximum: 500)

Number of denoising steps

Default: 50

guidance_scale

number

(minimum: 1, maximum: 50)

Scale for classifier-free guidance

Default: 7.5

prompt_strength

number

(minimum: 0, maximum: 1)

Prompt strength when using img2img / inpaint. 1.0 corresponds to full destruction of information in image

Default: 0.8

seed

integer

Random seed. Leave blank to randomize the seed

refine

string

Which refine style to use

Default: "no_refiner"

high_noise_frac

number

(minimum: 0, maximum: 1)

For expert_ensemble_refiner, the fraction of noise to use

Default: 0.8

refine_steps

integer

For base_image_refiner, the number of steps to refine, defaults to num_inference_steps

apply_watermark

boolean

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.

Default: true

lora_scale

number

(minimum: 0, maximum: 1)

LoRA additive scale. Only applicable on trained models.

Default: 0.6

disable_safety_checker

boolean

Disable safety checker for generated images. This feature is only available through the API. See https://replicate.com/docs/how-does-replicate-work#safety

Default: false

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

npx create-replicate --model=martintmv-git/sdxl-cinematic

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";

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

Run martintmv-git/sdxl-cinematic using Replicate’s API. Check out the model's schema for an overview of inputs and outputs.

const output = await replicate.run(
  "martintmv-git/sdxl-cinematic:40224311d882f9cd120ea648c0bd11836b606b70b4dc215a7c6661235eefbaf2",
  {
    input: {
      width: 1024,
      height: 512,
      prompt: "A medieval knight standing in a misty forest, with a castle visible in the distance, in the style of TOK, crisp, sharp, photorealistic, movie scene",
      refine: "expert_ensemble_refiner",
      scheduler: "K_EULER",
      lora_scale: 0.8,
      num_outputs: 1,
      guidance_scale: 7.5,
      apply_watermark: true,
      high_noise_frac: 0.8,
      negative_prompt: "cropped, worst quality, low quality, glitch, deformed, mutated, disfigured",
      prompt_strength: 0.8,
      num_inference_steps: 70
    }
  }
);

// To access the file URL:
console.log(output[0].url()); //=> "http://example.com"

// To write the file to disk:
fs.writeFile("my-image.png", output[0]);

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 martintmv-git/sdxl-cinematic using Replicate’s API. Check out the model's schema for an overview of inputs and outputs.

output = replicate.run(
    "martintmv-git/sdxl-cinematic:40224311d882f9cd120ea648c0bd11836b606b70b4dc215a7c6661235eefbaf2",
    input={
        "width": 1024,
        "height": 512,
        "prompt": "A medieval knight standing in a misty forest, with a castle visible in the distance, in the style of TOK, crisp, sharp, photorealistic, movie scene",
        "refine": "expert_ensemble_refiner",
        "scheduler": "K_EULER",
        "lora_scale": 0.8,
        "num_outputs": 1,
        "guidance_scale": 7.5,
        "apply_watermark": True,
        "high_noise_frac": 0.8,
        "negative_prompt": "cropped, worst quality, low quality, glitch, deformed, mutated, disfigured",
        "prompt_strength": 0.8,
        "num_inference_steps": 70
    }
)
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 martintmv-git/sdxl-cinematic 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": "martintmv-git/sdxl-cinematic:40224311d882f9cd120ea648c0bd11836b606b70b4dc215a7c6661235eefbaf2",
    "input": {
      "width": 1024,
      "height": 512,
      "prompt": "A medieval knight standing in a misty forest, with a castle visible in the distance, in the style of TOK, crisp, sharp, photorealistic, movie scene",
      "refine": "expert_ensemble_refiner",
      "scheduler": "K_EULER",
      "lora_scale": 0.8,
      "num_outputs": 1,
      "guidance_scale": 7.5,
      "apply_watermark": true,
      "high_noise_frac": 0.8,
      "negative_prompt": "cropped, worst quality, low quality, glitch, deformed, mutated, disfigured",
      "prompt_strength": 0.8,
      "num_inference_steps": 70
    }
  }' \
  https://api.replicate.com/v1/predictions

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

Output

{
  "completed_at": "2023-12-18T22:24:04.685958Z",
  "created_at": "2023-12-18T22:23:49.147331Z",
  "data_removed": false,
  "error": null,
  "id": "zh3rek3b464uvbicmr3jajgrnq",
  "input": {
    "width": 1024,
    "height": 512,
    "prompt": "A medieval knight standing in a misty forest, with a castle visible in the distance, in the style of TOK, crisp, sharp, photorealistic, movie scene",
    "refine": "expert_ensemble_refiner",
    "scheduler": "K_EULER",
    "lora_scale": 0.8,
    "num_outputs": 1,
    "guidance_scale": 7.5,
    "apply_watermark": true,
    "high_noise_frac": 0.8,
    "negative_prompt": "cropped, worst quality, low quality, glitch, deformed, mutated, disfigured",
    "prompt_strength": 0.8,
    "num_inference_steps": 70
  },
  "logs": "Using seed: 47266\nEnsuring enough disk space...\nFree disk space: 3213668261888\nDownloading weights: https://replicate.delivery/pbxt/HfFr4HaeymgDTEUFvcGuiUsLXUYRmRQdQYe9RZnduvh2hNHkA/trained_model.tar\n2023-12-18T22:23:53Z | INFO  | [ Initiating ] dest=/src/weights-cache/ff016a5a915406b1 minimum_chunk_size=150M url=https://replicate.delivery/pbxt/HfFr4HaeymgDTEUFvcGuiUsLXUYRmRQdQYe9RZnduvh2hNHkA/trained_model.tar\n2023-12-18T22:23:53Z | INFO  | [ Complete ] dest=/src/weights-cache/ff016a5a915406b1 size=\"186 MB\" total_elapsed=0.500s url=https://replicate.delivery/pbxt/HfFr4HaeymgDTEUFvcGuiUsLXUYRmRQdQYe9RZnduvh2hNHkA/trained_model.tar\nb''\nDownloaded weights in 0.6219019889831543 seconds\nLoading fine-tuned model\nDoes not have Unet. assume we are using LoRA\nLoading Unet LoRA\nPrompt: A medieval knight standing in a misty forest, with a castle visible in the distance, in the style of <s0><s1>, crisp, sharp, photorealistic, movie scene\ntxt2img mode\n  0%|          | 0/55 [00:00<?, ?it/s]\n  2%|▏         | 1/55 [00:00<00:10,  5.39it/s]\n  4%|▎         | 2/55 [00:00<00:08,  6.39it/s]\n  5%|▌         | 3/55 [00:00<00:07,  6.78it/s]\n  7%|▋         | 4/55 [00:00<00:07,  6.99it/s]\n  9%|▉         | 5/55 [00:00<00:07,  7.11it/s]\n 11%|█         | 6/55 [00:00<00:06,  7.18it/s]\n 13%|█▎        | 7/55 [00:01<00:06,  7.22it/s]\n 15%|█▍        | 8/55 [00:01<00:06,  7.24it/s]\n 16%|█▋        | 9/55 [00:01<00:06,  7.27it/s]\n 18%|█▊        | 10/55 [00:01<00:06,  7.28it/s]\n 20%|██        | 11/55 [00:01<00:06,  7.29it/s]\n 22%|██▏       | 12/55 [00:01<00:05,  7.30it/s]\n 24%|██▎       | 13/55 [00:01<00:05,  7.29it/s]\n 25%|██▌       | 14/55 [00:01<00:05,  7.30it/s]\n 27%|██▋       | 15/55 [00:02<00:05,  7.29it/s]\n 29%|██▉       | 16/55 [00:02<00:05,  7.30it/s]\n 31%|███       | 17/55 [00:02<00:05,  7.30it/s]\n 33%|███▎      | 18/55 [00:02<00:05,  7.30it/s]\n 35%|███▍      | 19/55 [00:02<00:04,  7.29it/s]\n 36%|███▋      | 20/55 [00:02<00:04,  7.29it/s]\n 38%|███▊      | 21/55 [00:02<00:04,  7.28it/s]\n 40%|████      | 22/55 [00:03<00:04,  7.28it/s]\n 42%|████▏     | 23/55 [00:03<00:04,  7.27it/s]\n 44%|████▎     | 24/55 [00:03<00:04,  7.28it/s]\n 45%|████▌     | 25/55 [00:03<00:04,  7.28it/s]\n 47%|████▋     | 26/55 [00:03<00:03,  7.28it/s]\n 49%|████▉     | 27/55 [00:03<00:03,  7.29it/s]\n 51%|█████     | 28/55 [00:03<00:03,  7.28it/s]\n 53%|█████▎    | 29/55 [00:04<00:03,  7.27it/s]\n 55%|█████▍    | 30/55 [00:04<00:03,  7.27it/s]\n 56%|█████▋    | 31/55 [00:04<00:03,  7.27it/s]\n 58%|█████▊    | 32/55 [00:04<00:03,  7.27it/s]\n 60%|██████    | 33/55 [00:04<00:03,  7.27it/s]\n 62%|██████▏   | 34/55 [00:04<00:02,  7.28it/s]\n 64%|██████▎   | 35/55 [00:04<00:02,  7.28it/s]\n 65%|██████▌   | 36/55 [00:04<00:02,  7.28it/s]\n 67%|██████▋   | 37/55 [00:05<00:02,  7.27it/s]\n 69%|██████▉   | 38/55 [00:05<00:02,  7.26it/s]\n 71%|███████   | 39/55 [00:05<00:02,  7.27it/s]\n 73%|███████▎  | 40/55 [00:05<00:02,  7.27it/s]\n 75%|███████▍  | 41/55 [00:05<00:01,  7.27it/s]\n 76%|███████▋  | 42/55 [00:05<00:01,  7.28it/s]\n 78%|███████▊  | 43/55 [00:05<00:01,  7.27it/s]\n 80%|████████  | 44/55 [00:06<00:01,  7.27it/s]\n 82%|████████▏ | 45/55 [00:06<00:01,  7.26it/s]\n 84%|████████▎ | 46/55 [00:06<00:01,  7.26it/s]\n 85%|████████▌ | 47/55 [00:06<00:01,  7.26it/s]\n 87%|████████▋ | 48/55 [00:06<00:00,  7.26it/s]\n 89%|████████▉ | 49/55 [00:06<00:00,  7.27it/s]\n 91%|█████████ | 50/55 [00:06<00:00,  7.27it/s]\n 93%|█████████▎| 51/55 [00:07<00:00,  7.27it/s]\n 95%|█████████▍| 52/55 [00:07<00:00,  7.27it/s]\n 96%|█████████▋| 53/55 [00:07<00:00,  7.26it/s]\n 98%|█████████▊| 54/55 [00:07<00:00,  7.26it/s]\n100%|██████████| 55/55 [00:07<00:00,  7.26it/s]\n100%|██████████| 55/55 [00:07<00:00,  7.23it/s]\n  0%|          | 0/15 [00:00<?, ?it/s]\n  7%|▋         | 1/15 [00:00<00:01,  7.41it/s]\n 13%|█▎        | 2/15 [00:00<00:01,  8.07it/s]\n 20%|██        | 3/15 [00:00<00:01,  8.30it/s]\n 27%|██▋       | 4/15 [00:00<00:01,  8.41it/s]\n 33%|███▎      | 5/15 [00:00<00:01,  8.48it/s]\n 40%|████      | 6/15 [00:00<00:01,  8.52it/s]\n 47%|████▋     | 7/15 [00:00<00:00,  8.55it/s]\n 53%|█████▎    | 8/15 [00:00<00:00,  8.55it/s]\n 60%|██████    | 9/15 [00:01<00:00,  8.54it/s]\n 67%|██████▋   | 10/15 [00:01<00:00,  8.55it/s]\n 73%|███████▎  | 11/15 [00:01<00:00,  8.55it/s]\n 80%|████████  | 12/15 [00:01<00:00,  8.55it/s]\n 87%|████████▋ | 13/15 [00:01<00:00,  8.56it/s]\n 93%|█████████▎| 14/15 [00:01<00:00,  8.58it/s]\n100%|██████████| 15/15 [00:01<00:00,  8.58it/s]\n100%|██████████| 15/15 [00:01<00:00,  8.49it/s]",
  "metrics": {
    "predict_time": 11.489998,
    "total_time": 15.538627
  },
  "output": [
    "https://replicate.delivery/pbxt/jVFMVKiPFSKeOafqhfsbQudz0UKQEKHewPtNE1sTwm8TIecQC/out-0.png"
  ],
  "started_at": "2023-12-18T22:23:53.195960Z",
  "status": "succeeded",
  "urls": {
    "get": "https://api.replicate.com/v1/predictions/zh3rek3b464uvbicmr3jajgrnq",
    "cancel": "https://api.replicate.com/v1/predictions/zh3rek3b464uvbicmr3jajgrnq/cancel"
  },
  "version": "40224311d882f9cd120ea648c0bd11836b606b70b4dc215a7c6661235eefbaf2"
}

Generated in

11.5 seconds

Tweak itReport

Using seed: 47266
Ensuring enough disk space...
Free disk space: 3213668261888
Downloading weights: https://replicate.delivery/pbxt/HfFr4HaeymgDTEUFvcGuiUsLXUYRmRQdQYe9RZnduvh2hNHkA/trained_model.tar
2023-12-18T22:23:53Z | INFO  | [ Initiating ] dest=/src/weights-cache/ff016a5a915406b1 minimum_chunk_size=150M url=https://replicate.delivery/pbxt/HfFr4HaeymgDTEUFvcGuiUsLXUYRmRQdQYe9RZnduvh2hNHkA/trained_model.tar
2023-12-18T22:23:53Z | INFO  | [ Complete ] dest=/src/weights-cache/ff016a5a915406b1 size="186 MB" total_elapsed=0.500s url=https://replicate.delivery/pbxt/HfFr4HaeymgDTEUFvcGuiUsLXUYRmRQdQYe9RZnduvh2hNHkA/trained_model.tar
b''
Downloaded weights in 0.6219019889831543 seconds
Loading fine-tuned model
Does not have Unet. assume we are using LoRA
Loading Unet LoRA
Prompt: A medieval knight standing in a misty forest, with a castle visible in the distance, in the style of <s0><s1>, crisp, sharp, photorealistic, movie scene
txt2img mode
  0%|          | 0/55 [00:00<?, ?it/s]
  2%|▏         | 1/55 [00:00<00:10,  5.39it/s]
  4%|▎         | 2/55 [00:00<00:08,  6.39it/s]
  5%|▌         | 3/55 [00:00<00:07,  6.78it/s]
  7%|▋         | 4/55 [00:00<00:07,  6.99it/s]
  9%|▉         | 5/55 [00:00<00:07,  7.11it/s]
 11%|█         | 6/55 [00:00<00:06,  7.18it/s]
 13%|█▎        | 7/55 [00:01<00:06,  7.22it/s]
 15%|█▍        | 8/55 [00:01<00:06,  7.24it/s]
 16%|█▋        | 9/55 [00:01<00:06,  7.27it/s]
 18%|█▊        | 10/55 [00:01<00:06,  7.28it/s]
 20%|██        | 11/55 [00:01<00:06,  7.29it/s]
 22%|██▏       | 12/55 [00:01<00:05,  7.30it/s]
 24%|██▎       | 13/55 [00:01<00:05,  7.29it/s]
 25%|██▌       | 14/55 [00:01<00:05,  7.30it/s]
 27%|██▋       | 15/55 [00:02<00:05,  7.29it/s]
 29%|██▉       | 16/55 [00:02<00:05,  7.30it/s]
 31%|███       | 17/55 [00:02<00:05,  7.30it/s]
 33%|███▎      | 18/55 [00:02<00:05,  7.30it/s]
 35%|███▍      | 19/55 [00:02<00:04,  7.29it/s]
 36%|███▋      | 20/55 [00:02<00:04,  7.29it/s]
 38%|███▊      | 21/55 [00:02<00:04,  7.28it/s]
 40%|████      | 22/55 [00:03<00:04,  7.28it/s]
 42%|████▏     | 23/55 [00:03<00:04,  7.27it/s]
 44%|████▎     | 24/55 [00:03<00:04,  7.28it/s]
 45%|████▌     | 25/55 [00:03<00:04,  7.28it/s]
 47%|████▋     | 26/55 [00:03<00:03,  7.28it/s]
 49%|████▉     | 27/55 [00:03<00:03,  7.29it/s]
 51%|█████     | 28/55 [00:03<00:03,  7.28it/s]
 53%|█████▎    | 29/55 [00:04<00:03,  7.27it/s]
 55%|█████▍    | 30/55 [00:04<00:03,  7.27it/s]
 56%|█████▋    | 31/55 [00:04<00:03,  7.27it/s]
 58%|█████▊    | 32/55 [00:04<00:03,  7.27it/s]
 60%|██████    | 33/55 [00:04<00:03,  7.27it/s]
 62%|██████▏   | 34/55 [00:04<00:02,  7.28it/s]
 64%|██████▎   | 35/55 [00:04<00:02,  7.28it/s]
 65%|██████▌   | 36/55 [00:04<00:02,  7.28it/s]
 67%|██████▋   | 37/55 [00:05<00:02,  7.27it/s]
 69%|██████▉   | 38/55 [00:05<00:02,  7.26it/s]
 71%|███████   | 39/55 [00:05<00:02,  7.27it/s]
 73%|███████▎  | 40/55 [00:05<00:02,  7.27it/s]
 75%|███████▍  | 41/55 [00:05<00:01,  7.27it/s]
 76%|███████▋  | 42/55 [00:05<00:01,  7.28it/s]
 78%|███████▊  | 43/55 [00:05<00:01,  7.27it/s]
 80%|████████  | 44/55 [00:06<00:01,  7.27it/s]
 82%|████████▏ | 45/55 [00:06<00:01,  7.26it/s]
 84%|████████▎ | 46/55 [00:06<00:01,  7.26it/s]
 85%|████████▌ | 47/55 [00:06<00:01,  7.26it/s]
 87%|████████▋ | 48/55 [00:06<00:00,  7.26it/s]
 89%|████████▉ | 49/55 [00:06<00:00,  7.27it/s]
 91%|█████████ | 50/55 [00:06<00:00,  7.27it/s]
 93%|█████████▎| 51/55 [00:07<00:00,  7.27it/s]
 95%|█████████▍| 52/55 [00:07<00:00,  7.27it/s]
 96%|█████████▋| 53/55 [00:07<00:00,  7.26it/s]
 98%|█████████▊| 54/55 [00:07<00:00,  7.26it/s]
100%|██████████| 55/55 [00:07<00:00,  7.26it/s]
100%|██████████| 55/55 [00:07<00:00,  7.23it/s]
  0%|          | 0/15 [00:00<?, ?it/s]
  7%|▋         | 1/15 [00:00<00:01,  7.41it/s]
 13%|█▎        | 2/15 [00:00<00:01,  8.07it/s]
 20%|██        | 3/15 [00:00<00:01,  8.30it/s]
 27%|██▋       | 4/15 [00:00<00:01,  8.41it/s]
 33%|███▎      | 5/15 [00:00<00:01,  8.48it/s]
 40%|████      | 6/15 [00:00<00:01,  8.52it/s]
 47%|████▋     | 7/15 [00:00<00:00,  8.55it/s]
 53%|█████▎    | 8/15 [00:00<00:00,  8.55it/s]
 60%|██████    | 9/15 [00:01<00:00,  8.54it/s]
 67%|██████▋   | 10/15 [00:01<00:00,  8.55it/s]
 73%|███████▎  | 11/15 [00:01<00:00,  8.55it/s]
 80%|████████  | 12/15 [00:01<00:00,  8.55it/s]
 87%|████████▋ | 13/15 [00:01<00:00,  8.56it/s]
 93%|█████████▎| 14/15 [00:01<00:00,  8.58it/s]
100%|██████████| 15/15 [00:01<00:00,  8.58it/s]
100%|██████████| 15/15 [00:01<00:00,  8.49it/s]

Examples

View more examples

Run time and cost

This model costs approximately $0.020 to run on Replicate, or 50 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 L40S GPU hardware. Predictions typically complete within 21 seconds.

Readme

This model doesn't have a readme.