Tencent Unveils HunyuanVideo: A New Text-to-Video Tool
Tencent Holdings launches HunyuanVideo, a text-to-video tool based on its AI model. The tool aims to create high-quality, realistic videos with advanced features like camera angles and reflections. HunyuanVideo can generate up to five-second videos in 720p resolution, responding to prompts in English or Chinese.
With over 13 billion parameters, making it the world's largest open-source model for video generation, HunyuanVideo aims to produce high-quality, "hyperrealistic" videos with convincing camera angles and reflections.
Tencent's HunyuanVideo is part of the company's efforts to compete in the AI video generation space, following the footsteps of other Chinese tech companies like OpenAI and Kuaishou Technology. The tool, based on the Diffusion Transformer architecture, can create up to five-second videos in 720p resolution, responding to prompts in English or Chinese.
Users can experience HunyuanVideo through Tencent's AI chatbot app, Yuanbao, which accepts text prompts and offers various video style, quality, and size options. The tool also allows users to select different camera movements and shots for a more personalised video creation experience.
Since the introduction of its Hunyuan foundation model last year, Tencent has been continuously enhancing its AI capabilities. In collaboration with academic institutions, Tencent has launched tools like Follow-Your-Click for image-to-video conversion and text-to-image generation, improving performance by 20% compared to previous versions.
Tencent Holdings launches HunyuanVideo, a text-to-video tool based on its AI model.
The tool aims to create high-quality, realistic videos with advanced features like camera angles and reflections.
HunyuanVideo can generate up to five-second videos in 720p resolution, responding to prompts in English or Chinese.
Source: SCMP