happyhorse-1.0-r2v

HappyHorse-1.0-R2V is a reference-generated video variant that supports up to 9 reference pictures as visual anchors, combined with text prompts to generate videos. Its core advantage lies in "more stable subject and scene reference", which can accurately maintain creative intentions and maintain the consistency of character facial features, clothing details and design language throughout the entire video, rather than relying solely on the first frame image. The input modes are also images and text, and the output is video. Services are provided through the Alibaba Cloud Bailian platform. The deployment area and current limiting configuration are the same as the previous two variants. This model is suitable for narrative scenarios that require character consistency, such as short dramas, advertising, and e-commerce content production.

Context

0 tokens

Max output

0 tokens

Input

llm image video

Output

video