Generating Hannibal Character Videos with FL2VA Model

Nimblecloud13 · reddit · 2026-08-24

A user shared experiments generating videos using the FL2VA (Text-to-Video) model, featuring odd pairings of known characters like Hannibal. The workflow follows a Hugging Face dataset example, set at 25 steps, 960x544 resolution, and 12 seconds duration.

Original post →

More from Multimodal

Multimodal channel →