These are experimental FP8 merges of the LTX2 diffusion model. "All in ones" do not seem to be in the cards, as LTX2 has lots of (large) moving parts, including 2 VAEs for video and audio. So, for the meantime, I'm just going to focus on the diffusion model. For all other models needed to run LTX2, please refer here: https://huggingface.co/Kijai/LTXV2_comfy

Workflows for LTX2 are also rather messy, unfortunately, so I don't have a clean one to provide at the moment. I'm monitoring the development of NSFW LORAs, which are slowly rolling in.

I2V and First to Last Frame Tip: Use really low strengths for LTXVAddGuide nodes, like 0.3. Higher values seem to be more likely to get ignored or cause cuts. With such low strength values, LTXPreprocess nodes that add compression artifacts can be reduced (or even skipped to preserve more detail). Also prompting things like "Special instructions: continuous shot with no cuts or transitions" might help.

----- VERSION HISTORY -----

Phr00tMerge V1: I feel like the default distilled "overcooks" a bit too much, so I made a merge with 0.875 distilled + 0.5 of the "detailer" LORA to hopefully improve details. euler_a/beta recommended.

Phr00tMergeNSFW V1: This adds in the breast LORA, panties LORA and a "prone face cam" LORA. I don't expect it to work very good (if at all) and I haven't really tested it, but hey... v1, right?

Downloads last month

-

Downloads are not tracked for this model. How to track
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Model tree for Phr00t/LTX2-Rapid-Merges

Base model

Lightricks/LTX-2
Finetuned
(12)
this model