My (old) go-to workflows for use in comfy UI. Get the new one HERE
You will need some additional nodes, in the comfy manager go to 'manager'>install missing custom nodes and that should do the trick.
Try my Flux prompt craft GPT
Get the GGUF CLIPs here
My SDXL/PONY workflow uses Clip Vector Sculpting to improve image quality
The FLUX AIO and FLUX + Upscale AIO V1 is as simple as possible for users using one of the AIO (All In One) safetensors checkpoints here on Civit and not the Official DiT only version from Hugging Face.
The FLUX + UPSCALE + Lora v3 is the one I use most, and is compatible with NepotismFUX V2, it's as simple as possible while still allowing you to change whatever you need on the fly. perfect for intermediate and expert users alike.
Description
improved workflow with image normalization, so now you can use high-resolution images without it slowing down
changed unet loader to gguf
changed clip loader to gguf
default settings updated to best practices
FAQ
Comments (39)
I figured it out, I just had to ungroup the nodes so I could see all the individual parts :)
I am using Flux Flow V2 (Adv+Lora) and wondering if there is a way to add wildcards for prompting to this workflow. I have tried a few different approaches and cannot seem to figure it out, I am a comfyUI newbie.
Any suggestions greatly appreciated, this is the best workflow I have found and I have tried many!
Hi sweet, firstly I love your workflows, they are a work of art💜But using the FluxFlow Advanced Img2Img+LoRA+Upscale, I can't add any LoRA to the LoRA Stack, am I missing something here????
@BobsBlazed Yeah that's the one I'm using.....clicking the 'Add LoRA' does nothing! Can't add anything matey🤷🏻♀️
@Apolonia make sure your nodes are up to date
@BobsBlazed Tried it on TensorArt & SeaArt.... doesn't work on either platform's ComfyUI I'm afraid 🤷🏻♀️
Shame, because I really like it😔
@Apolonia i run it locally, not sure about those platforms sorry
check out the updated version! https://civitai.com/models/1644427/bobs-flux-workflow-v2
@BobsBlazed Hey, yes I downloaded it yesterday, thanks.
I built my own PC and have been local since January, so any issues with things on cloud sites are just an ancient memory now😊
Appreciate the heads up, thanks sugar💋
Flux Flow (gguf img2img) Does not find the missing custom nodes
comfy manager installs the missing nodes, if its not working check for comfy manager updates and then update comfy ui as well
@BobsBlazed Do you know why it takes way longer to generate img2img then text2img?
@wildernessVikingQ381 input images are scaled to 1MP, so it shouldn't be slow even if your image is very large, if its slower for you it's likely due to a hardware limitation on your end.
also using lora will slow your generations
Is the clip_l in your workflow any different to the normal clip_l? I can't find that for download if it is.
Yes it is significantly better than the default (in my opinion) it adheres better to the prompt and creates more imaginative and detailed outputs
My most up to date clip L for flux is V5 but feel free to experiment with the other versions too!
I'm desperate and willing to give Buzz to whoever edit the "FLUX (Advanced+LORA+GGUF-UNET+GGUF-CLIP)" workflow to add a setting to change Flux Guidance setting + add batch image generation. Thank you
I have added it to the workflow archive, please re-download, both the original and the version with guidance and batch images are in the same zip file
@BobsBlazed Amazing. Thanks a million.
Where's CFG in your workflow(s)? I want to try using Dedestilled Flux but your workflow has no CFG option that I can see?
I use the basic guider so there is no CFG to modify, but I am working on a more advanced workflow currently that will allow CFG modification as well as allow any aspect ratio and calculate the optimal latent size. if you're interested in that node you can read more about it here, I developed it myself https://github.com/BobsBlazed/Bobs_Latent_Optimizer
Hi! I updated this workflow based on feedback and my own optimizations https://civitai.com/models/1644427/bobs-flux-workflow-v2 check it out :D
Been getting a out of memory error on image2image
I have RTX 4090 24GIG
double check that your nodes are up to date, then check the resolution of your input image and make sure its not larger than about 2 megapixels
@BobsBlazed it was my fault using to big of an image
works very well thank you!
Hi! Help me, please! When i using FLUX (Advanced+lora+img2img) workflow i get an error: mat1 and mat2 shapes cannot be multiplied. What could be wrong?
sounds like youre using a non flux lora
Kind of a new issue. The advanced 2 lora workflow worked wonders for a long time. now suddenly won't upscale any longer, and just stops at the generated image before the upscale (again nothing changed anywhere else). I wonder what happened or if I am doing something wrong. No missing items and no changes that could have caused this. Anything you may be aware of?
could be due to a comfyui update, try dragging the json into comfy, that should fix the group node. sometimes group nodes are finicky with updates
Strange that even after a clean install still won't work. Odd that the Flux_seed, Seed, Upscale_denoise fields have no selectable values and just NaN on them. Any ideas?
@GhostlyGaze right click and "reload node"
my new workflow should work better for you https://civitai.com/models/1644427/bobs-flux-workflow-v2
this was my absolute fav workflow ever, can you update it to work with Kontext GGUF? Please?