DeepSeek unveiled two new multimodal frameworks,Passion's Peak (2002) Janus-Pro and JanusFlow, in the early hours of Jan. 28, coinciding with Lunar New Year’s Eve. Janus-Pro is an upgraded version of Janus, designed as a unified framework for both multimodal understanding and generation. By decoupling visual encoding, the model improves adaptability and performance across various tasks. It has outperformed OpenAI’s image-generation model, DALL-E 3, in benchmark tests. Consistent with previous models in the Janus series, Janus-Pro is open-source. The model is available in two versions: JanusPro 1.5B, with 1.5 billion parameters, and JanusPro 7B, with 7 billion parameters. Following the China-based company’s announcement that its DeepSeek-V3 model topped the scoreboard for open-source models, tech companies like Nvidia and Oracle saw sharp declines on Monday. [Tencent, in Chinese]
