Friends who are engaged in MG animation design and editing, especially beginners, often spend a lot of time looking for MG image materials, or buying other people's materials. Time-consuming and energy-consuming, today I would like to recommend Baidu's MG-style character generator and animation material generator based on the ERNIE-ViL large model.
ERNIE-ViL is the industry's first multimodal pre-training model that incorporates scene graph knowledge. ERNIE-ViL integrates scene graph knowledge into the pre-training process of the visual-language model, learns the joint representation of scene semantics, and significantly enhances the cross-modal semantic understanding ability. ERNIE-ViL also refreshed the world's best results in five typical multimodal tasks including visual commonsense reasoning, visual question answering, reference expression understanding, cross-modal image retrieval, and cross-modal text retrieval. And it topped the list on the authoritative list of visual commonsense reasoning tasks (VCR) in the multimodal field.
Baidu Wenxin Yanggu creative application space provides a lot of creative applications, including MG style character generator and animation material generation applications. You only need to enter the prompt words of the creative prompt, and the basic creative materials can be generated according to your text content. Of course, it may not be 100% ideal, but it is very good for beginners. The generated colors are quite beautiful and very artistic.
Experience address: https://wenxin.baidu.com/younger/creativeSpace
animation material generation https://wenxin.baidu.com/younger/template?id=10070
MG style character generation https://wenxin.baidu.com/ younger/template?id=10061