This lora is used to make videos of targets turning into giants. It does a pretty good job of continuing the original image. Simply upload any picture and have the subject in the picture transform into a giant walking down the street!

These objects are recommended for video generation: children, pets, dolls

Use the I2V-14B-420P as a base. Used 10 videos, each about 5 seconds long and cropped out the face. I describe what happens, for example:

Trigger word: DXJDH

Here is the recommended PROMPT writeup:

DXJDH, At the beginning, the protagonist begins to step forward in a gesture with smooth movements. Immediately after, the protagonist steps out of the original background and is in the middle of a wide, clean street; the original background collapses, and the protagonist maintains his forward-stepping stance as he walks through the buildings. The viewpoint gradually pulls away, the protagonist firmly strides forward, each step seems heavy and powerful, the protagonist's huge body stands out in the city, weaving in the middle of the street, each step taken by the protagonist brings up a huge amount of dust, bringing a strong visual impact.