This is created in/for ComfyUi Version 0.5.1 portable version.

I am running this workflow on a system spec. with a 4070 12Gb VRAM and with 32Gb of system RAM. It takes about 12-15 minutes to run an entire sequence for me - which is a 15 second clip.

It does work in new versions as well - results may vary - or break due to memory management changes. (or out of date nodes)

This is a Wan 2.2 workflow (NSFW) for GGUF for Image 2 Video with notes in the workflow telling you where files go and links to download those files. I tried to make it as easy as possible to use. There is some custom nodes that may be required - like the "Power Lora Loader" - you could go without it - but you will be in pain to change this to support normal Lora loading. You can however choose to not download the "PainterI2V" node and swap it for the regular "WanImagetoVideo" node.

The workflow most likely can also be retrofitted to work for other non-quantized models and a Text to Video.

What is This - What does it do?

This could be your One-All-Be-All NSFW Generation (or SFW I guess).

It allows you to generate over 20 different NSFW scenes that can be changed with change of a single number from a 0 to a 1, all the way up to 19 (starts with 0-19 = 20) and image sometimes to help guide it - notes in the "Instructions" on what numbers work better with those guided images.

OPERATION:

1) Go to the "Main Base"

2) You will only need to specify character traits, such as eye color, or dress items in the 'Character Descriptions" - See Notepad box above it for more information.

3) Load your Image for the Image-2-Video on the right.

4) Set the 'Switch for Action' number based on the 'action list' on the left and hit run. The workflow will run the desired selected result. (For Female Characters Typically - If you want it to work for something else, some flavor text may need modified in the groups of each action)

So, what exactly does this do - how does it work? It will generate 5 second clips (default setting - can change) in sequence following a index number value, it will save your First Clip sequence, take the last frame and use it in a loop as your starting frame for the Second Clip, and again for the Third - after, it will save out a new video combined with all the video into a 15 second clip.

Each subgroup is linked to a "Any Index Switch" which is monitored by the Index Value for the Action, and inside is the three fields to run that are monitored on another index switch with the value from the loop. While the "Fields Used (cycle steps)" is the loop steps count - which the workflow is configured for this to happen 3 times.

Each clip generated is running its own prompt - take a look at the Subgroup on "19" with instructions inside of it to create your own sequence - and even extend it beyond 3 clips to even 4+ (not recommended beyond 5 - due to artifacts and color shifting)

There is also a audio clip sound played during each generation completion - and a double back to back when finished.

I have named all the Lora download files to the workflow to match the current download default names of the files. you can however set them manually (or name your lora to match in your lora folder) - keep in mind - this can take a while to set for this workflow, as the workflow consist of several Subgraphs for each action with power lora loaders inside them, with a total of 6 power lora loaders per grouping, then there is also a total of 20 Actions preconfigured.

If you want to take a test run to see how this thing operates - or if it works for you - navigate to "Group 00" and inside of it will have a notepad with only the Lora's for that grouping to run (Guided image needed for best result).

I have left notes inside the workflow on alternative methods to set your Lora's using notepad text editor and using the find and replace method.

I highly recommend taking a look at the workflow before downloading things in the notes inside the workflow. I have also set shortcut keys to be able to jump around the workflow - Start with navigating to the "Main Base" which is Hotkey #1 on your keyboard - or bookmark #1 (Bottom left of the workboard space).

At the "Main Base" take a look at the note to the left, under the "Fast Group Bypasser (rgthree)"

Frame Interpolation model used:

https://github.com/Fannovel16/ComfyUi-Frame-Interpolation