goodguy1963/hidream-l1-full-img2img – Run with an API on Replicate

goodguy1963 / hidream-l1-full-img2img

IMG2IMG for HiDream FULL AND DEV - does creative variations

Cold

Public
455 runs
A100 (80GB)
Weights

Run with an API

Playground API Examples README Versions

Input

image

file

Input image file (required)

use_florence2

boolean

Use Florence2 for automatic caption generation

Default: true

prompt_prefix

string

Shift + Return to add a new line

With Florence2: Prefix to add to generated caption. Without Florence2: Complete prompt to use

Default: "Photo"

florence2_task

string

Florence2 task to perform

Default: "prompt_gen_mixed_caption_plus"

negative_prompt

string

Shift + Return to add a new line

Negative prompt to guide the model away from certain elements

Default: "bad ugly jpeg artifacts"

model_variant

string

HiDream model variant to use

Default: "full"

seed

integer

Random seed

Default: 42

steps

integer

(minimum: 1, maximum: 100)

Number of inference steps

Default: 50

cfg

number

(minimum: 1, maximum: 20)

Guidance scale

Default: 5.54

denoise

number

(minimum: 0, maximum: 1)

Denoising strength

Default: 0.6

shift

number

(minimum: 0, maximum: 20)

ModelSamplingSD3 shift parameter (0-20)

Default: 3

sampler_name

string

Sampling algorithm to use

Default: "uni_pc"

scheduler

string

Scheduler for the sampler

Default: "normal"

Including florence2_text_input

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

npx create-replicate --model=goodguy1963/hidream-l1-full-img2img

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 goodguy1963/hidream-l1-full-img2img using Replicate’s API. Check out the model's schema for an overview of inputs and outputs.

const output = await replicate.run(
  "goodguy1963/hidream-l1-full-img2img:7820b5b85ecffa68a2ad8aeba3dfe526f1046fcaa209863fdcd999480751fc88",
  {
    input: {
      cfg: 1,
      seed: 42,
      image: "https://replicate.delivery/pbxt/MtfhuC8BBnHkfqKiWnIBMmdqO31vTSruHTdWyWwcVS4NTx29/Better%20prompt%20%20%2887%29.jpg",
      shift: 6,
      steps: 8,
      denoise: 0.34,
      scheduler: "linear_quadratic",
      sampler_name: "lcm",
      model_variant: "dev",
      prompt_prefix: "style the full image in ghibli style: ",
      use_florence2: false,
      florence2_task: "prompt_gen_mixed_caption_plus",
      negative_prompt: "bad ugly jpeg artifacts realistic photo",
      florence2_text_input: ""
    }
  }
);

// 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 goodguy1963/hidream-l1-full-img2img using Replicate’s API. Check out the model's schema for an overview of inputs and outputs.

output = replicate.run(
    "goodguy1963/hidream-l1-full-img2img:7820b5b85ecffa68a2ad8aeba3dfe526f1046fcaa209863fdcd999480751fc88",
    input={
        "cfg": 1,
        "seed": 42,
        "image": "https://replicate.delivery/pbxt/MtfhuC8BBnHkfqKiWnIBMmdqO31vTSruHTdWyWwcVS4NTx29/Better%20prompt%20%20%2887%29.jpg",
        "shift": 6,
        "steps": 8,
        "denoise": 0.34,
        "scheduler": "linear_quadratic",
        "sampler_name": "lcm",
        "model_variant": "dev",
        "prompt_prefix": "style the full image in ghibli style: ",
        "use_florence2": False,
        "florence2_task": "prompt_gen_mixed_caption_plus",
        "negative_prompt": "bad ugly jpeg artifacts realistic photo",
        "florence2_text_input": ""
    }
)
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 goodguy1963/hidream-l1-full-img2img 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": "goodguy1963/hidream-l1-full-img2img:7820b5b85ecffa68a2ad8aeba3dfe526f1046fcaa209863fdcd999480751fc88",
    "input": {
      "cfg": 1,
      "seed": 42,
      "image": "https://replicate.delivery/pbxt/MtfhuC8BBnHkfqKiWnIBMmdqO31vTSruHTdWyWwcVS4NTx29/Better%20prompt%20%20%2887%29.jpg",
      "shift": 6,
      "steps": 8,
      "denoise": 0.34,
      "scheduler": "linear_quadratic",
      "sampler_name": "lcm",
      "model_variant": "dev",
      "prompt_prefix": "style the full image in ghibli style: ",
      "use_florence2": false,
      "florence2_task": "prompt_gen_mixed_caption_plus",
      "negative_prompt": "bad ugly jpeg artifacts realistic photo",
      "florence2_text_input": ""
    }
  }' \
  https://api.replicate.com/v1/predictions

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

Output

{
  "completed_at": "2025-04-25T06:34:32.916764Z",
  "created_at": "2025-04-25T06:34:22.046000Z",
  "data_removed": false,
  "error": null,
  "id": "5hv2p1c1vsrj60cpdeksxc84wr",
  "input": {
    "cfg": 1,
    "seed": 42,
    "image": "https://replicate.delivery/pbxt/MtfhuC8BBnHkfqKiWnIBMmdqO31vTSruHTdWyWwcVS4NTx29/Better%20prompt%20%20%2887%29.jpg",
    "shift": 6,
    "steps": 8,
    "denoise": 0.34,
    "scheduler": "linear_quadratic",
    "sampler_name": "lcm",
    "model_variant": "dev",
    "prompt_prefix": "style the full image in ghibli style: ",
    "use_florence2": false,
    "negative_prompt": "bad ugly jpeg artifacts realistic photo"
  },
  "logs": "📦 Downloading required HiDream models...\nUsing model variant: dev (hidream_i1_dev_bf16.safetensors)\nTotal models to check: 6\n🔄 Processing model 1/6: hidream_i1_dev_bf16.safetensors\n✅ Model hidream_i1_dev_bf16.safetensors already exists (31.86 GB), skipping download\n🔄 Processing model 2/6: clip_g_hidream.safetensors\n✅ Model clip_g_hidream.safetensors already exists (1.29 GB), skipping download\n🔄 Processing model 3/6: clip_l_hidream.safetensors\n✅ Model clip_l_hidream.safetensors already exists (236.1 MB), skipping download\n🔄 Processing model 4/6: llama_3.1_8b_instruct_fp8_scaled.safetensors\n✅ Model llama_3.1_8b_instruct_fp8_scaled.safetensors already exists (8.46 GB), skipping download\n🔄 Processing model 5/6: t5xxl_fp8_e4m3fn_scaled.safetensors\n✅ Model t5xxl_fp8_e4m3fn_scaled.safetensors already exists (4.80 GB), skipping download\n🔄 Processing model 6/6: ae.safetensors\n✅ Model ae.safetensors already exists (319.8 MB), skipping download\n📊 Download Summary:\n✅ Successfully downloaded/verified: 6/6\nModels ready:\n- hidream_i1_dev_bf16.safetensors (31.86 GB)\n- clip_g_hidream.safetensors (1.29 GB)\n- clip_l_hidream.safetensors (236.1 MB)\n- llama_3.1_8b_instruct_fp8_scaled.safetensors (8.46 GB)\n- t5xxl_fp8_e4m3fn_scaled.safetensors (4.80 GB)\n- ae.safetensors (319.8 MB)\nSeed set to: 42\nRunning workflow\n[ComfyUI] got prompt\nExecuting node 82, title: Load Image, class type: LoadImage\nExecuting node 109, title: Florence2Run, class type: Florence2Run\nExecuting node 108, title: String Function 🐍, class type: StringFunction|pysssss\nExecuting node 83, title: VAE Encode, class type: VAEEncode\nExecuting node 3, title: KSampler, class type: KSampler\n[ComfyUI] </s><s><s>A photo-realistic shoot from a close-up camera angle about a young woman with long blonde hair leaning over a marble table in a modern indoor setting, wearing a gray tank top and blue jeans. the image also shows a vibrant, colorful abstract painting in the background. on the middle of the image, a 20-year-old woman with light skin, blonde hair, and blue eyes appears to be leaning over the table, looking directly at the camera with a neutral expression. she has a slender physique and is wearing a grey tank top that accentuates her medium breasts. her hair is styled in a long hair style and falls down her back. she is positioned in the center of the frame, with her hands resting on the marble surface in front of her. her facial expression is neutral, and her eyes are looking straight at the viewer. her body is slim, and she is wearing blue jeans that are loose-fitting, accentuating her bare shoulders. her fingernails are painted with a light pink nail polish. on either side of her, two other 18-25 years old women with light brown hair are standing, one wearing a colorful jacket and the other wearing an orange top. the background is a large, abstract painting on the wall, and the lighting is bright and natural, coming from large windows that let in natural light.\n[ComfyUI]\n[ComfyUI] camera_angle: close up, art_style: photo realistic, location: indoor, background: A brightly lit indoor space with large windows allowing natural light to filter in, creating a vibrant and lively atmosphere., text: NA, distance_to_camera: upper body, clothing: tank top, image_composition: middle, pants: blue jeans, accessory: NA;NA;NA, shoes: none;no, action: leaning on the table;standing, facing_direction: facing viewer;facing viewer, eye_direction : looking at viewer;looking at viewer, facial_expression: neutral;neutral, ear: no eye_color: blue;blue, gender: 1woman;1man;1woman, hair_color;blonde;brown, hair style: long hair;short hair, race: light skin;light skin, body: slim;slim</s>\n[ComfyUI] Offloading model...\n[ComfyUI]\n[ComfyUI] 0%|          | 0/8 [00:00<?, ?it/s]\n[ComfyUI] 12%|█▎        | 1/8 [00:00<00:03,  1.87it/s]\n[ComfyUI] 25%|██▌       | 2/8 [00:01<00:03,  1.87it/s]\n[ComfyUI] 38%|███▊      | 3/8 [00:01<00:02,  1.86it/s]\n[ComfyUI] 50%|█████     | 4/8 [00:02<00:02,  1.86it/s]\n[ComfyUI] 62%|██████▎   | 5/8 [00:02<00:01,  1.86it/s]\n[ComfyUI] 75%|███████▌  | 6/8 [00:03<00:01,  1.86it/s]\n[ComfyUI] 88%|████████▊ | 7/8 [00:03<00:00,  1.86it/s]\n[ComfyUI] 100%|██████████| 8/8 [00:04<00:00,  1.86it/s]\nExecuting node 8, title: VAE Decode, class type: VAEDecode\nExecuting node 9, title: Save Image, class type: SaveImage\n[ComfyUI] 100%|██████████| 8/8 [00:04<00:00,  1.86it/s]\n[ComfyUI] Prompt executed in 10.19 seconds\noutputs:  {'9': {'images': [{'filename': 'ComfyUI-HiDream_00001_.png', 'subfolder': '', 'type': 'output'}]}, '108': {'text': ['photorealistic, A photo-realistic shoot from a close-up camera angle about a young woman with long blonde hair leaning over a marble table in a modern indoor setting, wearing a gray tank top and blue jeans. the image also shows a vibrant, colorful abstract painting in the background. on the middle of the image, a 20-year-old woman with light skin, blonde hair, and blue eyes appears to be leaning over the table, looking directly at the camera with a neutral expression. she has a slender physique and is wearing a grey tank top that accentuates her medium breasts. her hair is styled in a long hair style and falls down her back. she is positioned in the center of the frame, with her hands resting on the marble surface in front of her. her facial expression is neutral, and her eyes are looking straight at the viewer. her body is slim, and she is wearing blue jeans that are loose-fitting, accentuating her bare shoulders. her fingernails are painted with a light pink nail polish. on either side of her, two other 18-25 years old women with light brown hair are standing, one wearing a colorful jacket and the other wearing an orange top. the background is a large, abstract painting on the wall, and the lighting is bright and natural, coming from large windows that let in natural light. camera_angle: close up, art_style: photo realistic, location: indoor, background: A brightly lit indoor space with large windows allowing natural light to filter in, creating a vibrant and lively atmosphere., text: NA, distance_to_camera: upper body, clothing: tank top, image_composition: middle, pants: blue jeans, accessory: NA;NA;NA, shoes: none;no, action: leaning on the table;standing, facing_direction: facing viewer;facing viewer, eye_direction : looking at viewer;looking at viewer, facial_expression: neutral;neutral, ear: no eye_color: blue;blue, gender: 1woman;1man;1woman, hair_color;blonde;brown, hair style: long hair;short hair, race: light skin;light skin, body: slim;slim']}}\n====================================\nComfyUI-HiDream_00001_.png",
  "metrics": {
    "predict_time": 10.837739739,
    "total_time": 10.870764
  },
  "output": [
    "https://replicate.delivery/yhqm/ApsGsCEF5A4uJpYeMkCsgedkNAcoa7vQtJfBtZcUeukiv3YSB/ComfyUI-HiDream_00001_.webp"
  ],
  "started_at": "2025-04-25T06:34:22.079024Z",
  "status": "succeeded",
  "urls": {
    "stream": "https://stream.replicate.com/v1/files/yswh-so3bljzrke4gfwpuecxsecgajq2laryavznu7mcxq3j4ffuxdb6q",
    "get": "https://api.replicate.com/v1/predictions/5hv2p1c1vsrj60cpdeksxc84wr",
    "cancel": "https://api.replicate.com/v1/predictions/5hv2p1c1vsrj60cpdeksxc84wr/cancel"
  },
  "version": "bba8d8531fb0899600f5cf7b7e25094708fdb19ca45b48b74ef6b892e5c1a626"
}

Generated in

10.8 seconds

Tweak it ShareReport

📦 Downloading required HiDream models...
Using model variant: dev (hidream_i1_dev_bf16.safetensors)
Total models to check: 6
🔄 Processing model 1/6: hidream_i1_dev_bf16.safetensors
✅ Model hidream_i1_dev_bf16.safetensors already exists (31.86 GB), skipping download
🔄 Processing model 2/6: clip_g_hidream.safetensors
✅ Model clip_g_hidream.safetensors already exists (1.29 GB), skipping download
🔄 Processing model 3/6: clip_l_hidream.safetensors
✅ Model clip_l_hidream.safetensors already exists (236.1 MB), skipping download
🔄 Processing model 4/6: llama_3.1_8b_instruct_fp8_scaled.safetensors
✅ Model llama_3.1_8b_instruct_fp8_scaled.safetensors already exists (8.46 GB), skipping download
🔄 Processing model 5/6: t5xxl_fp8_e4m3fn_scaled.safetensors
✅ Model t5xxl_fp8_e4m3fn_scaled.safetensors already exists (4.80 GB), skipping download
🔄 Processing model 6/6: ae.safetensors
✅ Model ae.safetensors already exists (319.8 MB), skipping download
📊 Download Summary:
✅ Successfully downloaded/verified: 6/6
Models ready:
- hidream_i1_dev_bf16.safetensors (31.86 GB)
- clip_g_hidream.safetensors (1.29 GB)
- clip_l_hidream.safetensors (236.1 MB)
- llama_3.1_8b_instruct_fp8_scaled.safetensors (8.46 GB)
- t5xxl_fp8_e4m3fn_scaled.safetensors (4.80 GB)
- ae.safetensors (319.8 MB)
Seed set to: 42
Running workflow
[ComfyUI] got prompt
Executing node 82, title: Load Image, class type: LoadImage
Executing node 109, title: Florence2Run, class type: Florence2Run
Executing node 108, title: String Function 🐍, class type: StringFunction|pysssss
Executing node 83, title: VAE Encode, class type: VAEEncode
Executing node 3, title: KSampler, class type: KSampler
[ComfyUI] </s><s><s>A photo-realistic shoot from a close-up camera angle about a young woman with long blonde hair leaning over a marble table in a modern indoor setting, wearing a gray tank top and blue jeans. the image also shows a vibrant, colorful abstract painting in the background. on the middle of the image, a 20-year-old woman with light skin, blonde hair, and blue eyes appears to be leaning over the table, looking directly at the camera with a neutral expression. she has a slender physique and is wearing a grey tank top that accentuates her medium breasts. her hair is styled in a long hair style and falls down her back. she is positioned in the center of the frame, with her hands resting on the marble surface in front of her. her facial expression is neutral, and her eyes are looking straight at the viewer. her body is slim, and she is wearing blue jeans that are loose-fitting, accentuating her bare shoulders. her fingernails are painted with a light pink nail polish. on either side of her, two other 18-25 years old women with light brown hair are standing, one wearing a colorful jacket and the other wearing an orange top. the background is a large, abstract painting on the wall, and the lighting is bright and natural, coming from large windows that let in natural light.
[ComfyUI]
[ComfyUI] camera_angle: close up, art_style: photo realistic, location: indoor, background: A brightly lit indoor space with large windows allowing natural light to filter in, creating a vibrant and lively atmosphere., text: NA, distance_to_camera: upper body, clothing: tank top, image_composition: middle, pants: blue jeans, accessory: NA;NA;NA, shoes: none;no, action: leaning on the table;standing, facing_direction: facing viewer;facing viewer, eye_direction : looking at viewer;looking at viewer, facial_expression: neutral;neutral, ear: no eye_color: blue;blue, gender: 1woman;1man;1woman, hair_color;blonde;brown, hair style: long hair;short hair, race: light skin;light skin, body: slim;slim</s>
[ComfyUI] Offloading model...
[ComfyUI]
[ComfyUI] 0%|          | 0/8 [00:00<?, ?it/s]
[ComfyUI] 12%|█▎        | 1/8 [00:00<00:03,  1.87it/s]
[ComfyUI] 25%|██▌       | 2/8 [00:01<00:03,  1.87it/s]
[ComfyUI] 38%|███▊      | 3/8 [00:01<00:02,  1.86it/s]
[ComfyUI] 50%|█████     | 4/8 [00:02<00:02,  1.86it/s]
[ComfyUI] 62%|██████▎   | 5/8 [00:02<00:01,  1.86it/s]
[ComfyUI] 75%|███████▌  | 6/8 [00:03<00:01,  1.86it/s]
[ComfyUI] 88%|████████▊ | 7/8 [00:03<00:00,  1.86it/s]
[ComfyUI] 100%|██████████| 8/8 [00:04<00:00,  1.86it/s]
Executing node 8, title: VAE Decode, class type: VAEDecode
Executing node 9, title: Save Image, class type: SaveImage
[ComfyUI] 100%|██████████| 8/8 [00:04<00:00,  1.86it/s]
[ComfyUI] Prompt executed in 10.19 seconds
outputs:  {'9': {'images': [{'filename': 'ComfyUI-HiDream_00001_.png', 'subfolder': '', 'type': 'output'}]}, '108': {'text': ['photorealistic, A photo-realistic shoot from a close-up camera angle about a young woman with long blonde hair leaning over a marble table in a modern indoor setting, wearing a gray tank top and blue jeans. the image also shows a vibrant, colorful abstract painting in the background. on the middle of the image, a 20-year-old woman with light skin, blonde hair, and blue eyes appears to be leaning over the table, looking directly at the camera with a neutral expression. she has a slender physique and is wearing a grey tank top that accentuates her medium breasts. her hair is styled in a long hair style and falls down her back. she is positioned in the center of the frame, with her hands resting on the marble surface in front of her. her facial expression is neutral, and her eyes are looking straight at the viewer. her body is slim, and she is wearing blue jeans that are loose-fitting, accentuating her bare shoulders. her fingernails are painted with a light pink nail polish. on either side of her, two other 18-25 years old women with light brown hair are standing, one wearing a colorful jacket and the other wearing an orange top. the background is a large, abstract painting on the wall, and the lighting is bright and natural, coming from large windows that let in natural light. camera_angle: close up, art_style: photo realistic, location: indoor, background: A brightly lit indoor space with large windows allowing natural light to filter in, creating a vibrant and lively atmosphere., text: NA, distance_to_camera: upper body, clothing: tank top, image_composition: middle, pants: blue jeans, accessory: NA;NA;NA, shoes: none;no, action: leaning on the table;standing, facing_direction: facing viewer;facing viewer, eye_direction : looking at viewer;looking at viewer, facial_expression: neutral;neutral, ear: no eye_color: blue;blue, gender: 1woman;1man;1woman, hair_color;blonde;brown, hair style: long hair;short hair, race: light skin;light skin, body: slim;slim']}}
====================================
ComfyUI-HiDream_00001_.png

This output was created using a different version of the model, goodguy1963/hidream-l1-full-img2img:bba8d853.

Examples

View more examples

Run time and cost

This model costs approximately $0.12 to run on Replicate, or 8 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 85 seconds. The predict time for this model varies significantly based on the inputs.

Readme

HiDream Img2Img ComfyUI Workflow with settings that worked for me

This workflow enables advanced image-to-image generation using the HiDream model suite and Florence-2 prompt generator, designed for use with ComfyUI and Replicate.

Overview

Image-to-image generation with HiDream diffusion model
Florence-2 for prompt generation and captioning
VAE encoding/decoding and advanced CLIP-based text encoding
Negative prompt support for artifact reduction
LOW VRAM MODE

Required Models & Credits

Diffusion Model

hidream_i1_full_fp16.safetensors
Place in: ComfyUI/models/diffusion_models
Download
Thanks to HiDream.ai for the model!

For low VRAM user - GPU with less than 24GB VRAM:

You can replace the standard Diffusion Model Loader with the custom node Unet LOADER from city96/ComfyUI-GGUF by city96.
Download the HiDream-I1 Full GGUF model from: https://huggingface.co/city96/HiDream-I1-Full-gguf/tree/main Place all in: ComfyUI/models/unet

Text Encoders

Place all in: ComfyUI/models/text_encoders - clip_g_hidream.safetensors
Download - clip_l_hidream.safetensors
Download - llama_3.1_8b_instruct_fp8_scaled.safetensors
Download - t5xxl_fp8_e4m3fn_scaled.safetensors
Download

VAE

ae.safetensors
Place in: ComfyUI/models/vae
Download

Florence-2 Prompt Generator (NO need to download - will be downloaded automatacally at runtime)

Florence-2-large
Microsoft Florence-2
Thanks to MiaoshouAI for the correct implementation!

Usage

Download all required models and place them in the correct directories as listed above.
Drag the workflow image in ComfyUI
Use the workflow to generate new images from your input images and prompts.

For low VRAM user - GPU with less than 24GB VRAM:

You can replace the standard Diffusion Model Loader with the custom node Unet LOADER from city96/ComfyUI-GGUF by city96.
Download the HiDream-I1 Full GGUF model from: https://huggingface.co/city96/HiDream-I1-Full-gguf/tree/main
Follow the instructions in the custom node repository to set up and use the GGUF model with the Unet LOADER node.

Workflow Diagram

See the full workflow structure here:
WORKFLOW-HIDREAM-IMG2IMG.png

Acknowledgements

HiDream.ai for the diffusion model and encoders
Microsoft for Florence-2
MiaoshouAI for the Florence-2 prompt generator implementation
ComfyUI team for the UI and workflow engine

(wait for HiDream-E1 for even better results)