The Essential Calvin & Hobbes - FLUX.2 Klein 9b Base -> 4x upscaler
https://redd.it/1tqhdxm
@rStableDiffusion

What style is this?

https://redd.it/1tqiufn
@rStableDiffusion

Film Auteur (LTXV) version 2.0.5 update

It's been about a while since I first posted about this node I've been working on for LTX 2.3, triXope Film Auteur (LTXV). Since then, I've been working hard to implement and perfect numerous features, iron out bugs, and clean up the UI for readability. It's gone through several phases/iterations since my previous post, but I feel that I'm finally ready to release the latest edition that is version 2.0.5.

If you missed the original post, basically Film Auteur (LTXV) is a custom node for ComfyUI that simplifies working with LTX while simultaneously bringing all features (and then some) into one single node - a complete production-ready suite - one node to rule them all (so they say).

With this node there is no need to run any video extenders or multiple runs for separate clips. Enter as little or as many prompts as you want, separated by "|" (eg. prompt 1 | prompt 2 | prompt 3 | etc.), or just a single prompt for a long clip, and the node will handle it all. No need to worry about OOM errors.

Here is the list of features (so far):

Text-to-Video
Image-to-Video
Image Reference-to-Video (experimental work-in-progress)
Audio-to-Video
Audio Reference (with ID-LoRA)
Ollama integration for prompt enhancement
Normalized Attention Guidance (NAG) integration
Integrated "Director Mode" with multi-shot inferencing
Image input accepts image batch for storyboard processing or reference images
LTXV Add Guide & LTX Add Video IC-LoRA Guide fully implemented under the hood for added control & consistency over reference images
Inifinite length by use of autoregressive chunking and built-in sliding context windows
1 or 2 spatial upscale passes
Temporal upscaling option (doubles the framerate and improves motion, lip sync, and visual fidelity)
Face restoration to help with cleaning up faces and removing artifacts (work-in-progress)
Integrated Audio Mastering Pass (Soft Limiter & Normalization)
Built-in sageattention and fp16 accumulation
Built in chunk feed forward (to assist in computational efficiency)
Unload models & clear cache (optional switch)
Built in stage 1 preview
Internal Real-Time ETA counter (with assist node) (work-in-progress)

Upcoming/Planned features:

Prompt Relay
Keyframes (first, middle, last frame, etc.)
RTX Super Resolution upscaler
and many more

Please look over the list of features, and over all settings in the node, before asking whether something is or isn't included.

There is currently one workflow included for text-to-video. I will work on placing more.

Search triXope in the ComfyUI manager or check it out here:
https://github.com/triXope/ComfyUI-triXope

Disclaimer: I am NOT a coder or developer by trade... I am simply a hobbyist with a passion for innovation and happen to be extremely resourceful when it comes to learning new crafts/skills.

P.S. Feel free to toss out any thoughts, recommendations, or suggestions - I'm always working to improve/enhance the note. And by all means, if you find this node to be the least bit useful or interesting, please pass this post along to any family, friends, or colleagues that may be interested.

https://redd.it/1tqgzi7
@rStableDiffusion

Cracked the case on high res + quality Qwen Edit 2511 outputs, here are minimalistic workflows & lots of info on how/why

https://redd.it/1tqm8ic
@rStableDiffusion