NoobAI / Illustrious - txt2img workflow

Donate here if you want: https://ko-fi.com/lopi999

This is a fairly straightforward workflow that I've made over my entirety of using vpred-based Noob models. By bypassing the ModelSamplingDiscrete node it should be possible to be used with EPS models as well.

Main features:

• An artist switcher to easily switch around different artists or randomize them.

• Hires fix

• Face detailer (as well as a hand detailer)

• Upscaler using the EPS NAI Tile ControlNet

By default, uses Euler CFG++ with 1.7 CFG and Beta scheduler on the first sampler, and UniPC with 5 CFG and Normal scheduler on the hires sampler. Uses Mahiro and Rescale CFG with a factor of 0.7. By default, also uses zeta, which is a custom scheduler provided by my custom nodes pack. It is not really intended to be used on it's own, so use another scheduler like beta or kl_optimal if you don't want to hires fix.

Changelog

V6: I have wiped everything clean and started fresh here. It is functionally similar to earlier versions but seeks to solve a lot of issues, namely with the sampler parameter node and the second FaceDetailer node being picked up as a different node for... whatever reason. This was accomplished using a new, fresh build of Comfy specifically for this purpose of working out a new workflow, then tested fully on my production build of Comfy.

Now I really wanted to publish a version in which I was able to use something else other than the SD Prompt Saver node suite to save images with metadata, but to no avail. If you have issues importing it, please just go to the GitHub repo for it and make an issue, or just use the save image nodes instead.

• Added logic nodes to configure what output to use for the sampler/detailer groups, allowing easier bypassing of hiresfix and handdetailer.

• Added a switch node to configure what group, sampler or detailer, that the upscaler should pull from

• Negative prompt is now inserted into the metadata and has a token counter associated with it

V5: This brings in some new organization to the workflow, and now, enjoy not having to manually set ANYTHING for the SD Prompt Saver nodes (aside from extra info). Note: needs the latest version of comfyui-lopi999-nodes to work.

• Added loader group

• Fixed a few glaring issues with the upscale group

• Added sampler and model parameters that get sent to the SD Prompt Saver node

• Swapped out the latent size picker for a modified version with a resolution multiplier and a resolution swapper

• Added a few more artist mixes I guess

V4: What happens when you decide to push updates before testing them? Well, normally nobody sane would even dare to pose this question, but ComfyUI sure did and it answered for itself. A lot of stuff broke when Comfy pushed out the new frontend update, so, this forced me to have to scramble together a new workflow.

• Replaced Effeciency Nodes LoRA Stacker because it was heavily broken in the new update

• Removed prompt composer because without being able to save space by turning widgets into inputs, it lost it's purpose

• Added a token counter

• Changed the spacing of nodes and groups because now the minimization of nodes has been decreased

V3: I began designing V3 because I wanted to add a bit more features to V2, while also trying to figure out how to make the overall workflow much cleaner.

• Added a prompt saver section, that uses the SD Prompt Reader node family to function. This allows for images to be saved with a lot of metadata intact.

• Efficiency nodes have been replaced with EasyUse nodes, and a global seed has been added. This prevents a lot of seed issues I was having with the Efficiency node family.

• ModelSamplingDiscrete node has been removed, Comfy by default auto-detects the sampling method to use and already knows to use ZSNR for Vpred Noob models.

• TextBox3 and a few other nodes have been gutted out, the node that originally fed into the positive input for the loader is now a Prompt Composer Node. The ugly mess of many individual textboxes with a switch for artist mixes has been gutted out and replaced with a node I made recently, compacting the area of the Artist Mixer group and allowing for a built-in way to index or randomize the mix used.

V2: Added a LoRA stack group.