This is an advanced workflow designed for those familiar with ComfyUI who want pro grade results. It is node heavy and has numerous options. I encourage you to take the time to learn it. If you feel overwhelmed, you can start with my simple model HERE
也有中文说明
Instagram: https://www.instagram.com/synth.studio.models/
Buy me a☕ https://ko-fi.com/lonecatone
This represents hundreds of hours of work. If you enjoy it, please 👍like, 💬 comment , and feel free to ⚡tip 😉
Note: If you have ANY issues with nodes not downloading, read the notes or reach out. There's nothing that special about any of them that aren't core modules. Most issues can be resolved by updating everything
⛔⚠️🛑✋ Read the notes completly before using. Most common install and node problems are listed in the directions
A complete user and troubleshooting guide inclusing a list of common error messages can be found here https://civarchive.com/articles/24678/how-to-use-my-workflows-a-comprehensive-breakdown
V14.0
Because it is not in my nature to leave things alone....
Gutted the stock face detailer nodes and replaced them with a SEGS suite for infinately better results
Note: Yes, I will be going through all my workflows and replacing them, including teh NSFW ones
Controlnet Update.
Added local bypass nodes so you don't have to zoom and scroll
Major update to notes, especially for detailers
V13.0
Overhauled the Controlnet:
Now you can select the aspect ratio of the original image without having to "guess" it with the aspect ratio node
Left room for the resize node so you can choose which side to crop the image by
Redid the dependencies so that every tie you shut it on and off, you don;t need to reset the switches so it will not give you an error 😡
Massively improved the notes throughout
Minor layout tweaks
V11.3
Detailers are now able to handle float32 models, so I rewmoved the SDXL Based checkpoint.
Please give me feedback with any issues on this, especailly RAM related.
Added a crop section so you can...well... crop images easily
Updated several areas
Added the ability to use a simple date with the metadata folder
added a "remove banned tags" node
V8.0
Removed Florence Prompt assist and replaced it with Qwen Advanced.
Reorganized the controlnet section because it bugged me. Now it's easier to use.
Added a LoRa controller for the detailer for better character facial and eye consistency as well as a localized prompt
Gutted the backend of the post production suite and added 📷optical realism, giving a significantly better way to fine tune your images.
Fixed some backend mathAdded a 📂 subfolder generator that seperates the metadata files according to the name you give the file for easier use.
V7.0
Added KSampler Configure Nodes so you can leave the setting in place for both ZIT and base without having to change them
Removed some problematic nodes
Removed Qwen Prompt enhancer. I refuse to put an API one in there.
global scheduler and samplers for consistency
Better math
Improved notes throughout.
All in 1 workflow:
Controlnet
Checkpoint option for merged models
I2I
Prompt generation
Prompt enhancement
Lora Loader
Advanced KSampler
Draft mode for making corrections without having to rerun the workflow.
Seed variance enhancer
Ultimate Upscaler for prescaling and hi res fix
Detailing Suite for face, hands, eyes, and expressions
Seed VR2 Upscaler to get 4k on low VRAM
Massive Post Production Suite
JPEG Scrubber
Smart Noise scrubber
Full tutorial available.
也有中文说明
Check out my other models:
My NSFW version with Anantomy Correctors:https://civarchive.com/models/2270894/zit-nsfw-anatomy-corrector
For ultimate Anime detail and control: https://civarchive.com/models/2220766/zimage-ultra-anime-detail-workflow-get-the-most-out-of-your-generations
For Realism and variety over the base model: https://civarchive.com/models/2231181/z-image-ultra-real-workflow
Description
Pulled out & replaced nodes that broke the last upgrade
Added two options for prompt assist (Florence and Qwen)
Modified and did a major overhaul of teh post production suite
Upgraded Optical Realism
Ver 9.1 fixed broken Medadata save node (Thanks Comfyui)
FAQ
Comments (31)
Love all your workflows!
Thanks!!!
@lonecatone23 For what its worth Qwen is working perfectly fine, I did have to do a little rebuilding and GPT help but its really, really good. Do you perhaps take requests for new workflows? like I love yours and I love wild card things, is that something you would ever consider adding into your flow?
@Landover62 I LOOOOOVE a challenge. Suggestions Please!!!
Wait until you see my LTX 2.3 one I'm shooting out today. 🔥🔥
@Landover62 You know I have wildcard workflows, right?
@lonecatone23 I didnt know? I have alot of your workflows, Ive even spent like$35 on a singular workflo and its not even as good as yours, it doesnt have the optical or filter pass ect
@Landover62 Lol, I know who's workflow that was. He puts out garbage.
Also getting stuck with QwenVL, I already make sure everything is up to date but still failed to load, heres log:
Native code method attempted to call: lib.cquantize_blockwise_bf16_nf4()
🚨 CUDA VERSION MISMATCH 🚨
Requested CUDA version: 13.0
Detected PyTorch CUDA version: 13.0
Available pre-compiled versions:
- 11.8
- 12.0
- 12.1
- 12.2
- 12.3
- 12.4
- 12.5
- 12.6
- 12.8
- 12.9
This means:
The version you're trying to use is NOT distributed with this package
Attempted to use bitsandbytes native library functionality but it's not available.
This typically happens when:
1. bitsandbytes doesn't ship with a pre-compiled binary for your CUDA version
2. The library wasn't compiled properly during installation from source
To make bitsandbytes work, the compiled library version MUST exactly match the linked CUDA version.
If your CUDA version doesn't have a pre-compiled binary, you MUST compile from source.
You have two options:
1. COMPILE FROM SOURCE (required if no binary exists):
https://huggingface.co/docs/bitsandbytes/main/en/installation#cuda-compile
2. Use BNB_CUDA_VERSION to specify a DIFFERENT CUDA version from the detected one, which is installed on your machine and matching an available pre-compiled version listed above
Original error: Configured CUDA binary not found at C:\Users\80440\Documents\ComfyUI\.venv\Lib\site-packages\bitsandbytes\libbitsandbytes_cuda130.dll
🔍 Run this command for detailed diagnostics:
python -m bitsandbytes
If you've tried everything and still have issues:
1. Include ALL version info (operating system, bitsandbytes, pytorch, cuda, python)
2. Describe what you've tried in detail
3. Open an issue with this information:
https://github.com/bitsandbytes-foundation/bitsandbytes/issues
Native code method attempted to call: lib.cquantize_blockwise_bf16_nf4()
Error: Bnb4bitQuantize on tensors destined for model.visual.merger.mlp.2.weight. Ckpt contains: 1
Notes:
- MISSING :those params were newly initialized because missing from the checkpoint. Consider training on your downstream task.
- CONVERSION :originate from the conversion scheme
!!! Exception during processing !!! We encountered some issues during automatic conversion of the weights. For details look at the CONVERSION entries of the above report!
Traceback (most recent call last):
File "C:\Users\80440\AppData\Local\Programs\ComfyUI\resources\ComfyUI\execution.py", line 524, 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 "C:\Users\80440\AppData\Local\Programs\ComfyUI\resources\ComfyUI\execution.py", line 333, in get_output_data
return_values = await asyncmap_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 "C:\Users\80440\AppData\Local\Programs\ComfyUI\resources\ComfyUI\execution.py", line 307, in asyncmap_node_over_list
await process_inputs(input_dict, i)
File "C:\Users\80440\AppData\Local\Programs\ComfyUI\resources\ComfyUI\execution.py", line 295, in process_inputs
result = f(**inputs)
^^^^^^^^^^^
File "C:\Users\80440\Documents\ComfyUI\custom_nodes\ComfyUI-QwenVL\AILab_QwenVL_PromptEnhancer.py", line 136, in process
enhanced = self._invoke_qwen(
^^^^^^^^^^^^^^^^^^
File "C:\Users\80440\Documents\ComfyUI\custom_nodes\ComfyUI-QwenVL\AILab_QwenVL_PromptEnhancer.py", line 167, in invokeqwen
output = self.run(
^^^^^^^^^
File "C:\Users\80440\Documents\ComfyUI\custom_nodes\ComfyUI-QwenVL\AILab_QwenVL.py", line 828, in run
self.load_model(
File "C:\Users\80440\Documents\ComfyUI\custom_nodes\ComfyUI-QwenVL\AILab_QwenVL.py", line 730, in load_model
self.model = AutoModelForVision2Seq.from_pretrained(model_path, **load_kwargs).eval()
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Users\80440\Documents\ComfyUI\.venv\Lib\site-packages\transformers\models\auto\auto_factory.py", line 374, in from_pretrained
return model_class.from_pretrained(
^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Users\80440\Documents\ComfyUI\.venv\Lib\site-packages\transformers\modeling_utils.py", line 4073, in from_pretrained
loading_info = cls._finalize_model_loading(model, load_config, loading_info)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Users\80440\Documents\ComfyUI\.venv\Lib\site-packages\transformers\modeling_utils.py", line 4237, in finalizemodel_loading
log_state_dict_report(
File "C:\Users\80440\Documents\ComfyUI\.venv\Lib\site-packages\transformers\utils\loading_report.py", line 273, in log_state_dict_report
raise RuntimeError(
RuntimeError: We encountered some issues during automatic conversion of the weights. For details look at the CONVERSION entries of the above report!
Prompt executed in 146.83 seconds
This is kinda weird, 13.0 CUDA VERSION MISMATCH ??? I tried to reinstall my Comfy but it still happened, is that something wrong with 5070TI ? Also Thanks for workflow, that help me save lot of time.
@155T fuck Qwen. I pull it in and out of my workflows. It doesn't like to do anything with quants.
Just yank it. It's literally the reason why I have noth optons for Florence and that. You can literally just delete the whole node and the swu=itch and the workflow will heal itself
Getting this issue:
TypeError: can only concatenate str (not "NoneType") to str
Well, that's a new one. It is either an issue with how the metadata is being saved (see if the metadata node is purple around it) or a prompt issue (something is oddly bypassed). You can feed iot to Grok and it wil tell you. I'll check it out when I get home
Got it — now I can see the exact problem. Here's node 1569 (OpticalRealism) and its widgets_values:
widgets_values: [ true, // atmosphere_enabled 0.05, // haze_strength -0.05, // lift_blacks ← PROBLEM #2: min is 0.0 true, // (boolean) 0.6, // light_wrap_strength 0.05, // chromatic_aberration 0, // vignette_intensity "Manual", // grain_power ← PROBLEM #1: string, not float 0.3, // (monochrome grain or highlight rolloff) true // (boolean) ]Two fixes needed on that node:
1. grain_power (8th value) — change "Manual" to a number like 0.3
This looks like the OpticalRealism node was updated and the widget definition changed. The workflow was saved with a version that offered "Manual" as an option for grain, but the installed version's process_image() function on line 127 just does if grain_power > 0: expecting a float. The string "Manual" can't be compared to 0.
2. lift_blacks (3rd value) — change -0.05 to 0.0
The node's input definition sets a minimum of 0.0 for this parameter, and -0.05 fails validation.
@singularity42 It was. frigging annoying. I've swept all my workflows but thios one apparantly.
Thanks.
Update: I added ver 9.2 (with a skin upscaler)
to complicated setup for what it does. could be simplefied and better streamlined.
Oh yay! some random account who didn't even download it has an opinion. 🙄
Fake profile says what?
@lonecatone23 tbf, i can imagine your workflows can be quite overwhelming for those used to more simple click-and-go WF's. that being said you provide a lot of helpful tutorial notes for the unique nodes you use; I actually discovered a lot of cool exotic nodes from studying your WF's
@Anime_Art_Alchemist That's why I also made an essential one. This is my passion
Я не буду сдерживаться: этот воркфлоу - кусок заплесневевшего собачьего дерьма. Тонны бесполезных и при том весьма тяжёлых, требующих кучу зависимостей, нод, напиханных исключительно для создания ложного впечатления монументальной работы. Тогда как все лучшие воркфлоу всегда легки и просты. Я даже не говорю о том какой каловый эта воркфлоу выдаёт результат: напрочь ужаренные изображения, лоснящиеся от слоп-"масла".
@Chertilo Погоди-ка. На секунду мне показалось, что ты говоришь о своей маме. 🤔
Вот когда ты сможешь реально генерировать достойные изображения — используя ЛЮБОЙ рабочий процесс, а не те «какашки», что у тебя на странице, — тогда и дай знать. А до тех пор всем глубоко плевать на твоё мнение, Принцесса.
Thanks for sharing this, all your workflows are great for those of us who require autistic levels of precision :)
I'm trying to figure out if you meant Autistic or Artistic as both fit 🤔
@lonecatone23 You're right, both are valid xD
I can't find the I2I toggle in version 9.2. Was it deleted?
Yeah, I removed it because it didn't really do anything. Do you want it back?
@aiibel. I just added it back in for you. download (.3 above.
This is due for an update this weekend anyway
@lonecatone23 Thank you very much for adding it back. I appreciate your quick response!
@aiibel 🙏