DeepSeek has unveiled a new AI model called Janus-Pro, which supports multimodal media blending. This model comes in two sizes: 1B parameters and 7B parameters, with the ability to analyze and generate new images as specified.
Testing results using GenEval and DPG-Bench datasets have shown that Janus-Pro-7B outperforms OpenAI’s DALL-E 3 model, as well as other image-generating models such as PixArt-alpha, Emu3-Gen, and Stable Diffusion XL by Stability AI.
The model is available for download on GitHub under the MIT license, allowing for commercial use.
Source: TechCrunch
TLDR: DeepSeek’s Janus-Pro AI model excels in generating images and outperforms other models, available for commercial use on GitHub under the MIT license.
Leave a Comment