🎧 Trigger Words

asmr ear lick — main trigger, always include

Also recommended in your prompt:

• binaural

• 3dio microphone

⚙️ Recommended Settings

Strength: 1.0

About

This LoRA produces surprisingly consistent, good-quality audio.

The dataset primarily included clips a single actor performing alone but also included a few of two actors performing simultaneously, so dual/two-sided prompts will sometimes work. Be aware it's unreliable at syncing both sides at once at the moment - expect to try a few seeds if that's what you're going for.

Heads up: part of the dataset contained NSFW themes, so outputs may lean that direction even on tamer prompts.

🚧 Roadmap

Training is already underway on a larger dataset, possibly with a higher step count. Updated versions to follow. (Training ain't cheap, lol.)

💬 Feedback

This is my first LoRA, so any feedback, sample gens, or prompt tips are appreciated.

🛠️ Bonus: LTX Training UI

I built a custom web UI specifically for this workflow and I'm considering releasing it. What it does:

• Browse and select videos straight from your local drive

• Automatic clipping / segmenting

• Automatic captioning via a model

• Train directly from the interface

• Optional RunPod integration — drop in your API key and it spins up a pod for training and serverless for captioning on its own

If that sounds useful to you, let me know and I'll prioritize getting it out.