ByteDanceIts intelligent AI assistantBean curdNow upgradedVincent FigureCapacity to supportGenerate specified text with one click. Users can add text requirements to the raw image prompt words to generate an image with the specified text. This feature is currently available in theBeanbag App Open TestThe instant dream has also been tested on a small scale.
According to a report by "Tiger Sniff" today, the person in charge of Doubao's big model team said that Doubao's text-generation model has a better ability to learn native Chinese data by connecting the LLM and DIT architectures, and on this basis, it has strengthened the ability to generate Chinese characters, which greatly improves the generating effect.Beanbag's website and computer version will also be online soon..
Also at the beginning of this month, Doubao also launched its Picture Understanding feature, adding photo and camera buttons to the Doubao App and Doubao PC, which allow you to recognize the content of a picture after uploading it. Doubao's "Picture Understanding" feature recognizes the elements and basic features contained in a picture. Users can also ask where an attraction is located or which movie or TV show a character is from.