florestefano1975 / comfyui-portrait-master Goto Github PK

The project & model is amazing, but can we add more attributes/labels for more detailed controls?
e.g. beard, eyebrows, wrinkle .etc

I wanted to use just your Portrait Master node, but I wonder where and how I would connect the text out?

I tried your Workflow, but somewhat kept getting error on macOS. So I decided just to use the "Portrait Master" node, which seems to be the most useful. The rest, I am not sure necessary to set it up that way, unless there might affect the final result?

Edit: Was trying to load the linked json workflow, but had mistakenly saved the link as json rather than the json that is included in the link, sorry.

1st picture

Then ...

Do these need to be manually implimented? As well as manually being placed?

Thank you so much for making this, very useful. After using it, I found a few improvements that would make this node better:

1. Add a custom list. By adding a custom list, add more options that are adapted to individual needs.
2. Change the position of the random switch. Set the random switch next to the options, or make random an option. This will reduce the length of the node and make it easier to operate.

May I ask if all the functions described in the readme are controlled via prompt?

I check the code ,but clothes_list is not shown in the plugin.

i've just updated comfyui-portrait-master to v 2.6
loaded the icluded SDXL workflow, but when i try to generate an image i've this error on PromptControlSimple box:

Error occurred when executing PromptControlSimple:

control_to_clip_common() takes from 2 to 4 positional arguments but 5 were given

File "D:\ComfyUI\execution.py", line 152, in recursive_execute
output_data, output_ui = get_output_data(obj, input_data_all)
File "D:\ComfyUI\execution.py", line 82, in get_output_data
return_values = map_node_over_list(obj, input_data_all, obj.FUNCTION, allow_interrupt=True)
File "D:\ComfyUI\execution.py", line 75, in map_node_over_list
results.append(getattr(obj, func)(**slice_dict(input_data_all, i)))
File "D:\ComfyUI\custom_nodes\comfyui-prompt-control\prompt_control\node_aio.py", line 32, in apply
pos_cond = pos_filtered = control_to_clip_common(self, clip, pos_sched, lora_cache, cond_cache)

Got the following error message:

ERROR:root:Failed to validate prompt for output 13:
ERROR:root:* PortraitMaster 10:
ERROR:root:  - Failed to convert an input value to a FLOAT value: circular_pupil, raw photo, (realistic:1.5), could not convert string to float: 'raw photo, (realistic:1.5)'
ERROR:root:Output will be ignored
invalid prompt: {'type': 'prompt_outputs_failed_validation', 'message': 'Prompt outputs failed validation', 'details': '', 'extra_info': {}}

Thanks for the Node and Workflows.

When I add provided workflow in Comfyui, everything except KSampler (Inspire) works fine. It is giving error attached in screenshot.

I was able to generate output by replacing KSampler (Inspire) by simple KSampler Advanced, but I am not sure if output is best as some parameters are different in KSampler Inspire and KSampler Advanced. Can you please let me know if they both give same output.

Also, is there a way to get consistent output i.e. if we generate a portrait, then we generate several other portraits of same person/ image in different poses. Not able to acieve that even if Seed is fixed.

Thanks

I am using the excellent PM. It does very well in TurboXL mode at 512x512; but unless I'm much mistaken, the XL in TurboXL means that images should be at 1024x1024?!

It never works for me at least, as I get weirdly contorted faces, almost 'stuffed' into the frame, 4 eyes, 4 eyebrows. As I say, anything 512x512 is excellent!!!

I'd love to see more tutorials of the node.
Initially, while running the node, I ran into snags as to which settings should be on or off, and video tutorials could help!

Thanks so much!

Hello,

I noticed that without selecting a specific "face shape", all male characters have the same "top model" square jawline.
Addind "face shape" to the list of randomize parameters might fix that.

How to install and use

Hello, to achieve consistent results with portrait-master-pose-library-2.2b.jpg, we should use a specific workflow. Could you please provide more details about the specific workflow you are referring to?

I have unlocked the panels in PM2.3, but they do not shrink enough (or move) for me to see where I can change the VAE?

It still works but, is it normal that that node output is not connected to anything ?

I change the nationality, for example from german to british, and the model is the same.

This is a feat of nature but if you want extreame success I think this needs to be almost in reverse. Take a face and then spit out the settings for that face. Then the ability to save that setting. That way you have a base to start from. All the bases you set are (no offense) the same group of people. No Asian, East Indian or Native American.

I have no idea how much work that would be but... that would make the need for face swap obsolete of sorts. Hope it can be implemented somehow.

Hmm any advice what im doing wrong ?

Hai,i try to follow the installation instruction, and after install it all with comfy manager, i encounter this error when running it, any tips to fix it ? thanks

hi
Thanks for this wonderful work
But every time I open it, I find that the settings are a problem and I have to adjust them manually
Is there a solution for that?
Thank you all

Have a fresh install of comfyUI, and installed all the custom nodes.

Getting this:

Prompt outputs failed validation
PortraitMaster:

• Failed to convert an input value to a FLOAT value: iris_details, -, could not convert string to float: '-'
• Failed to convert an input value to a FLOAT value: circular_iris, -, could not convert string to float: '-'
• Value not in list: light_type: 'disable' not in (list of length 33)
• Value not in list: light_direction: '(realistic:1.5)' not in ['-', 'from bottom', 'from bottom-left', 'from bottom-right', 'from front', 'from left', 'from rear', 'from right', 'from top', 'from top-left', 'from top-right']
• Failed to convert an input value to a FLOAT value: light_weight, (symmetric face:1.5), could not convert string to float: '(symmetric face:1.5)'
• Value not in list: photorealism_improvement: '' not in ['enable', 'disable']
• Value not in list: style_1: '1.73' not in (list of length 66)

Also get this:

Error occurred when executing PromptControlSimple:

control_to_clip_common() takes from 2 to 4 positional arguments but 5 were given

File "F:\AI\ComfyUI_2024\ComfyUI\execution.py", line 152, in recursive_execute
output_data, output_ui = get_output_data(obj, input_data_all)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "F:\AI\ComfyUI_2024\ComfyUI\execution.py", line 82, in get_output_data
return_values = map_node_over_list(obj, input_data_all, obj.FUNCTION, allow_interrupt=True)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "F:\AI\ComfyUI_2024\ComfyUI\execution.py", line 75, in map_node_over_list
results.append(getattr(obj, func)(**slice_dict(input_data_all, i)))
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "F:\AI\ComfyUI_2024\ComfyUI\custom_nodes\comfyui-prompt-control\prompt_control\node_aio.py", line 32, in apply
pos_cond = pos_filtered = control_to_clip_common(self, clip, pos_sched, lora_cache, cond_cache)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

Can I upload photos as samples and have the portal master set parameters based on my photo portrait to create a new image? Thank you!