Flux with LoRA and Controlnet

This service is provided by Qubico using Qubico's specialized inference hardware, designed to optimize LoRA (Low-Rank Adaptation) and ControlNet functionalities for advanced model control in various Flux tasks.

Flux API (Task Creation with LoRA and ControlNet)

Model, Task Type and LoRA Numbers

Model Name	Task Type	LoRA Capacity	Controlnet Capacity
Qubico/flux1-dev-advanced	txt2img-lora	1	0
Qubico/flux1-dev-advanced	img2img-lora	1	0
Qubico/flux1-dev-advanced	controlnet-lora	0 or 1	1
Qubico/flux1-dev-advanced	contact us if you need a complex workflow customization	>1	>1

Note: Qubico/flux1-dev-advanced is the only model that supports LoRA and ControlNet in Qubico's Flux API.
You can explore Available LoRA and ControlNet to find out which LoRA or ControlNet models best fit your needs.

Example: Request Body of Text-to-Image Task with ControlNet and LoRA

{
  "model": "Qubico/flux1-dev-advanced",
  "task_type": "controlnet-lora",
  "input": {
    "prompt": "spiderman",
    "control_net_settings": [
      {
        "control_type": "depth",
        "control_image": "https://example.com/control_image2.png"
      }
    ],
    "lora_settings": [
      {
        "lora_type": "mjv6"
      }
    ]
  },
  "config": {
    "webhook_config": {
      "endpoint": "https://webhook.site/d35ab4aa-95f9-4855-af82-32fbfbb67063",
      "secret": ""
    }
  }
}

Request Example Using cURL

Example: Request Body of Task Creation with LoRA only

Input Example:

{
  "model": "Qubico/flux1-dev-advanced",
  "task_type": "txt2img-lora",
  "input": {
    "prompt": "spiderman",
    "lora_settings": [
      {
        "lora_type": "mjv6"
      }
    ]
  },
  "config": {
    "webhook_config": {
      "endpoint": "https://webhook.site/d35ab4aa-95f9-4855-af82-32fbfbb67063",
      "secret": ""
    }
  }
}

Example: Task Creation with controlnet Canny

{
    "model": "Qubico/flux1-dev-advanced",
    "task_type": "controlnet-lora",
    "input": {
        "steps": 28,
        "prompt": "A girl in city, 25 years old, cool, futuristic",
        "negative_prompt": "low quality, ugly, distorted, artefacts",
        "guidance_scale": 2.5,
        "control_net_settings": [
            {
                "control_type": "canny",
                "control_image": "https://i.ibb.co/yX07dwV/Comfy-UI-controlnet.webp",
                "control_strength": 0.45,
                "return_preprocessed_image": true
            }
        ],
        "lora_settings": [
            {
                "lora_type": "graphic-portrait",
                "lora_strength": 1
            }
        ]
    },
    "config": {
        "webhook_config": {
            "endpoint": "",
            "secret": ""
        }
    }
}

Canny Result 1	Canny Result 2

Example: Task Creation with controlnet Depth

{
    "model": "Qubico/flux1-dev-advanced",
    "task_type": "controlnet-lora",
    "input": {
        "steps": 28,
        "prompt": "A girl in city, 25 years old, cool, futuristic",
        "negative_prompt": "low quality, ugly, distorted, artefacts",
        "guidance_scale": 2.5,
        "control_net_settings": [
            {
                "control_type": "depth",
                "control_image": "https://i.ibb.co/yX07dwV/Comfy-UI-controlnet.webp",
                "control_strength": 0.45,
                "return_preprocessed_image": true
            }
        ],
        "lora_settings": [
            {
                "lora_type": "geometric-woman",
                "lora_strength": 1
            }
        ]
    },
    "config": {
        "webhook_config": {
            "endpoint": "",
            "secret": ""
        }
    }
}

Depth Result 1	Depth Result 2

Example: Task Creation with controlnet SoftEdge

{
    "model": "Qubico/flux1-dev-advanced",
    "task_type": "controlnet-lora",
    "input": {
        "steps": 28,
        "prompt": "A girl in city, 25 years old, cool, futuristic",
        "negative_prompt": "low quality, ugly, distorted, artefacts",
        "guidance_scale": 2.5,
        "control_net_settings": [
            {
                "control_type": "soft_edge",
                "control_image": "https://i.ibb.co/yX07dwV/Comfy-UI-controlnet.webp",
                "control_strength": 0.55,
                "return_preprocessed_image": true
            }
        ],
        "lora_settings": [
            {
                "lora_type": "remes-abstract-poster-style",
                "lora_strength": 1
            }
        ]
    },
    "config": {
        "webhook_config": {
            "endpoint": "https://webhook.site/0edf5555-22ba-4aae-898a-f707440effa4",
            "secret": ""
        }
    }
}

Softedge Result 1	Softedge Result 2

Example: Task Creation with controlnet OpenPose

{
    "model": "Qubico/flux1-dev-advanced",
    "task_type": "controlnet-lora",
    "input": {
        "steps": 28,
        "prompt": "person enjoying a day at the park, full hd, cinematic",
        "negative_prompt": "low quality, ugly, distorted, artefacts",
        "guidance_scale": 4.0,
        "control_net_settings": [
            {
                "control_type": "openpose",
                "control_image": "https://i.ibb.co/vkCbMZY/3-pose-1024.jpg",
                "control_strength": 0.7,
                "return_preprocessed_image": true
            }
        ],
        "lora_settings": [
            {
                "lora_type": "mjv6",
                "lora_strength": 1
            }
        ]
    },
    "config": {
        "webhook_config": {
            "endpoint": "https://webhook.site/0edf5555-22ba-4aae-898a-f707440effa4",
            "secret": ""
        }
    }
}

OpenPose Control Image INPUT	OpenPose Result

Flux with LoRA and Controlnet

Flux API (Task Creation with LoRA and ControlNet)#

Model, Task Type and LoRA Numbers#

Example: Request Body of Text-to-Image Task with ControlNet and LoRA#

Request Example Using cURL#

Example: Request Body of Task Creation with LoRA only#

Example: Task Creation with controlnet Canny#

Example: Task Creation with controlnet Depth#

Example: Task Creation with controlnet SoftEdge#

Example: Task Creation with controlnet OpenPose#