DeepSeek’s AI Momentum Continues with Release of Janus-Pro, a Multimodal Text-to-Image Model
DeepSeek isn’t slowing down
Just moments after knocking ChatGPT out of the top spot in the App Store for most downloaded free apps, the Chinese AI startup DeepSeek released its new multimodal text-to-image AI model, Janus-Pro, on Monday. This model is open-source under an MIT license, making it commercially viable, and is available for download via HuggingFace and GitHub.
Janus-Pro: A Competitor to R1
Similar to DeepSeek’s flagship model, R1, Janus-Pro is a suite of models with varying sizes, ranging from 1B to 7B parameters. According to DeepSeek’s testing, Janus-Pro-7B outperforms established image generators like Stable Diffusion and DALL-E on the GenEval and DPG-Bench benchmarks. The model uses an "autoregressive framework" and "surpasses" unified models.
Key Features and Limitations
Janus-Pro builds on Janus, its original version released last year, and can create and analyze images. However, smaller-parameter models in the family are limited to analyzing images of 384 x 384 resolution, which is a drawback. Nevertheless, Janus-Pro’s performance is still competitive, given DeepSeek’s reportedly lower training costs compared to those of US-based AI companies.
Reactions and Implications
Nvidia has praised Janus-Pro as "an excellent AI advancement." The model’s release has sparked mixed reactions, with some expressing concern about the accuracy of DeepSeek’s reported training costs. However, the true impact of Janus-Pro remains to be seen as more users test it against other image models.
FAQs
Q: What is Janus-Pro?
A: Janus-Pro is a multimodal text-to-image AI model developed by DeepSeek.
Q: What are the key features of Janus-Pro?
A: Janus-Pro can create and analyze images, has an autoregressive framework, and surpasses unified models.
Q: What are the limitations of Janus-Pro?
A: Smaller-parameter models are limited to analyzing images of 384 x 384 resolution.
Q: What are the implications of Janus-Pro’s release?
A: The model’s release has significant implications for the AI industry, including potential changes to plans like Stargate, a $500 billion initiative between several AI giants.
