A Google natív képgenerálást tett elérhetővé a Gemini 2.0 Flash fejlesztőknek
Google has released native image generation capabilities specifically for developers. Unlike traditional tools that call a separate model like DALL-E or Imagen, native generation allows the same model to generate both text and images. This integration enables features like natural and selective editing, where specific parts of an image can be modified without altering the entire composition.
- Access it via Google AI Studio by selecting the 'Gemini 2.0 Flash Experimental' model
- Enables natural and selective editing of images without changing the whole picture
- Google AI Studio now also supports asking questions about YouTube videos directly rather than just the transcript
Miért fontos?
Native multimodal generation makes models feel less 'dumb' by allowing them to handle text and imagery within the same context, leading to more precise editing and creative control.
Kapcsolódó hírek
A Mistral új modellt adott ki a hagyományos OCR rendszerek kiváltására
2025. március 11.