Nice little mod to have for beginners but depending on models, the results of such a prompt may vary a lot. For example, polluting the prompt with that many tags can probably blur the overall result of illustrous based models, which is the one I use. Another idea to complete the mod would be to separate prompts for bodies and portraits too, which usually don't show the same thing and don't have the same format. I very much appreciate the initiative though!

This is great and fun to play with. Only problems I have are that I cant get it to generate prompts for SFCAric races and I cant seem to copy the nude negative prompt as the window closes whenever I the copy button or try to manually select the prompt. Is there a fix for this?

The new character UI changed substantially in 0.13.3. I'll need to move this button somewhere else when I get the chance.

Hi,

I managed to install ComfyUI and the mod and to generate images for my characters, but I can't manage to generate the portrait from body. It gives me this:

Do you have an idea what I missed in the installation?

Running into the same error with the ComfyUI console logs giving 

"invalid prompt: {'type': 'missing_node_type', 'message': "Node 'Face Bounding Box' not found. The custom node may not be installed.", 'details': "Node ID '#2'", 'extra_info': {'node_id': '2', 'class_type': 'FaceBoundingBox', 'node_title': 'Face Bounding Box'}}".

Possible cause I found checking I noticed that my import of "ComfyUI_FaceAnalysis" through ComfyUI mentioned in the ReadMe failed. I'd assume it's due to it requiring DLib or InsightFace. Trying to install those and get it working with Comfy have been a pain I'm working through.

Not 100% if this would be the cause of the issue, just wanted to chime in with my own findings. Hope this helps!

Got it fixed and it's working now, I was correct and it was due to not having InsightFace installed which I had to do manually. What got it to work for me:

1. Using the terminal (I used VSCode Terminal) find the folder path where ComfyUI stores the virtual environment (.venv) folder, it will be the same folder that has the models. output, custom_nodes folder. 

2. Start the virtual environment. Should be like ./venv/Scripts/Activate

3. run the command "python -m pip install * " where the * is "insightface" then "onnxruntime" "opencv-python" and "numpy"

4. verify it installed correctly with the command "python -c "import insightface; print('InsightFace installed')""

Next time you open ComfyUI and check it should show no errors with the ComfyUI_FaceAnalysis node and it should work making portraits in game. This is what worked for me and I could of forgotten a step or two, ChatGPT was great in troubleshooting some stuff.

This was on a Windows 11 desktop for the platform.

Fantastic mod though it's very well implemented!

Hi Pyro, 
I'm getting an error "ComfyUI Error: No images in ComfyUI output". I can generate images manually using ComfyUI but nothing seems to happen in game. 

I've released 1.1.4, which contains an alternate portrait workflow for those having trouble with Face Analysis.

To use the new workflow, you will need to download new dependencies using the manager:

Node packs: ComfyUI Impact Pack, ComfyUI Impact Subpack
Models: face_yolov8m (bbox)

im curious, would it not be better to generate the nudes first before the clothed, or is there some reason you are generating the clothed first?

This is an amazing mod!  However, I seem to be running into an issue where the image generation just randomly gets stuck without saying there are any errors.  The status will say "Generating... (step 10/20)" and the step it gets stuck on is always random.  I can see looking at my ComfyUI console that all the steps are completing and the image shows up in my outputs folder, but it seems like the mod is getting interrupted somehow.  Pulling up the debug menu with F2 doesn't show any info.

Any idea what could be going on?  I can provide additional information if you need it.

I'm tracking an issue where sometimes the prompts generate for the wrong character. If you encounter it, the workaround is simple: just switch to another character's info screen and switch back.

I can't find face_yolov8m (bbox) :(

Unsure why, but I can't generate portraits using the ultralytics method. I've followed the installation process correctly. txt2img and img2img work perfectly fine, but I'm presented with this error for portraits specifically:

ComfyUI Error: ComfyUI execution error: {"prompt_id":"2d7ba987-8840-42c4-564c-6ed56ed59556","node_id":"22","node_type":"ImpactFrom_SEG_ELT","executed":[],"exception_message":"From_SEG_ELT.doit() missing 1 required positional argument: 'seg_elt'
","exception_type":"TypeError","traceback":["  File \"F:\\...\\ComfyUI\\resources\\ComfyUI\\execution.py\", line 534, in execute
output_data, output_ui, has_subgraph, has_pending_tasks = await get_output_data(prompt_id, unique_id, obj, input_data_all, execution_block_cb=execution_block_cb, pre_execute_cb=pre_execute_cb, v3_data=v3_data)
File \"F:\\...\\ComfyUI\\resources\\ComfyUI\\execution.py\", line 334, in get_output_data
return_values = await _async_map_node_over_list(prompt_id, unique_id, obj, input_data_all, obj.FUNCTION, allow_interrupt=True, execution_block_cb=execution_block_cb, pre_execute_cb=pre_execute_cb, v3_data=v3_data)
File \"F:\\...\\ComfyUI\\resources\\ComfyUI\\execution.py\", line 304, in _async_map_node_over_list
await process_inputs({})
File \"F:\\...\\ComfyUI\\resources\\ComfyUI\\execution.py\", line 296, in process_inputs\n    result = f(**inputs)
"current_inputs":{"seg_elt":[]},"current_outputs":["19","10","17","23","18","12","13","24","8","1","21","22","20","25","4","9"],"timestamp":1777412457081}

Is this a bug, or am I doing something wrong?

EDIT: I've been playing through a new save for a bit, and I think I can say that for whatever reason, the portraits are working with the exception of my own character, which is just a male Beast Wolf. I'm not really sure why, as I've made portraits for several humanoid characters and a few other beastkin, and it seems to work. My only guess is that the character itself has a pronounced muzzle compared to other characters, which may be making it difficult for him to have his face recognized.