LTX 2.3 I2V | Ollama LLM Enhancer | Dual Pass
Note: The v2 version has the basic LTX processing based on RuneXXX's full steps model and with optimized prompt enhancer for the INT8/ConvRot models.
===
Features:
Fairly simple Image 2 Video. This exact workflow is my own daily driver used for uploaded clips to Civitai.
LTX 2.3 I2V:
Image and (short context) pro,mpt is sent to Ollama for enhanced prompt.
Upscale sampler
Ollama part can be skipped by flipping the true/false switches.
Runs without issues on a 12GB GPU with a Q8 GGUF model. Larger models would require at least 16GB VRAM.
Note: For the I2V Ollama enhancement, I'm using the qwen3-vl-abliterated-8b model myself.
Description
FAQ
Comments (2)
Where did you get the diffusion model in this workflow?
I like to try different models. The workflow itself can handle almost any (I alternate between https://civitai.red/models/2732210/sexgod-pinkcherry-ltx-23-uncensored-nsfw-checkpoint?modelVersionId=3113845, https://civitai.red/models/2601098/sulphur-2-base?modelVersionId=2921800 (and the distilled version), https://civitai.red/models/2484503/ltx-23-distilled-gguf-unsloth?modelVersionId=2800693 (Q8 GGUF) and the Vantage Q8 at https://huggingface.co/vantagewithai/LTX-2.3-GGUF/tree/main/distilled.
Both 'NSFW' models (Sulpur / PinkCherry) are a bit tricky for regular (SFW) movement and don't always have proper voices (needs an often retry with a diff seed).