DeepSeek launched the Janus-Pro-7B vision AI model on January 27, 2025, which has outperformed OpenAI’s DALL-E 3 and Stability AI’s Stable Diffusion in image generation benchmarks. The model is available on Hugging Face and is open-source under the MIT license.
Janus-Pro-7B Overview: This multimodal AI model is designed for efficiency and versatility, capable of generating photorealistic images and performing complex visual reasoning. It is engineered to deliver high-level performance without demanding vast computational resources, making it accessible for enterprise customers.
大型言語模型：R1 BY DEEPSEEK
圖片生成模型：Janus-pro-7b BY DEEPSEEK
視頻生成模型：海螺AI
3D建模模型：HUNYUAN3D BY 騰訊
Market Impact: The release of Janus-Pro-7B triggered a notable market response, causing a 17% drop in Nvidia’s stock price and a sell-off of U.S. artificial intelligence stocks. This development highlights the intensifying competition between Chinese tech companies and Silicon Valley giants in the AI sector.
講緊deepseek , 佢係R1 相等於chatGPT01, chatGPT01 solves complex and critical problems. It is designed for researchers, engineers, and other individuals who use research-grade intelligence . 要比錢才用到，唔洗錢(外國）的是ChatGPT 4-o ( limited use per day though). chatGPT01 is research level . 而deepseek R1是similar to chatgpt 01 level but it is free . 所以用佢好過請大學生做tutor , 用來小學、中學、大學都卓卓有餘，佢最好係中文勁過chatGPT 好多，所以文言文、古詩、修詞手法etc etc 佢全部明過我同伱。
我之前是用chatGPT 的，因為佢英文強，但ChatGPT 香港大陸理論上是用不到的，現在有deepseek ， 香港可用，又免費。而且中文強勁過ChatGPT 多多聲，英文都很好。但我知大陸有其他對小學生學習英文的功能eg 豆包，我未用過，某程度d spoken English 有Chinese accent . 但deepseek 未有語音功能yet , but good enough just for Chinese , English , math, etc
發表於 
唔該晒
返回列表
