Using WebAPI img2img w/ multiple ControlNet units

[BobFrederick]

I've searched the discussions trying to find examples of using multiple ControlNet units with img2img and the WebAPI but i'm not seeing anything specific enough to provide clarity with my situation. I am able to get txt2img to work fine but I'm getting a HTTP 500 Error when sending the following payload. Does this seem correct? Many thanks! 🙏

# create payload using img2img method
        payload = {
            "prompt": user_prompt,
            "negative_prompt": negative_prompt,
            "seed": seed,
            "subseed": -1,
            "subseed_strength": 0.0,
            "batch_size": batch_size,
            "steps": steps,
            "cfg_scale": cfg_scale,
            "width" : width,
            "height" : height,
            "init_images": [encoded_inital_img2img],
            "alwayson_scripts": {
                "controlnet": {
                    "args": [
                        {
                            "input_image": encoded_image_a,
                            "module": "invert",
                            "model": "control_v11p_sd15_mlsd [aca30ff0]",
                        },
                        {
                            "input_image": encoded_image_b,
                            "module": "depth",
                            "model": "control_v11f1p_sd15_depth [cfd03158]",
                            "weight": 0.5,
                            "guidance": 1.0,
                            "guidance_start": 0.5,
                            "guidance_end": 1.0
                        },
                        {
                            "input_image": encoded_image_b,
                            "module": "segmentation",
                            "model": "control_v11p_sd15_seg [e1f51eb9]",
                            "weight": 1.0,
                            "guidance": 1.0,
                            "guidance_start": 0.5,
                            "guidance_end": 1.0
                        }
                    ]
                }
            }
        }

[BobFrederick]

I solved this issue and the payload request above was correctly structured however how I was encoding the images to base64 was specific to my application...

I had to use:

• encoded_inital_img2img = base64.b64encode(open(image_filepath, 'rb').read()).decode('utf-8')

instead of:

• encoded_inital_img2img = base64.b64encode(image_filepath).decode('utf-8')