MediaTek

MediaTek and Kuaishou jointly announced yesterday that they have launched a highly efficient end-to-end video generation technology. This technology is a continuation and improvement of the video generation technology that debuted at the 2024 Mobile World Congress. It combines Kuaishou's AI model I2V (image to video) Adapter with the AI computing power of MediaTek Dimensity 9300 and Dimensity 8300 mobile platforms to generate dynamic videos from static images on the end-to-end. With this technology, users only need to select a picture on the device, and the application can recognize the people and scenes in the photo and generate a natural and smooth video, providing...

At the MediaTek Dimensity Developer Conference MDDC2024 this morning, the highly anticipated Dimensity 9300++ flagship processor finally unveiled its mysterious veil. The CPU architecture of this processor is carefully laid out, integrating 1 Cortex-X4 core up to 3.40GHz, 3 cores at 2.85GHz, and 4 cores at 2.00GHz, ensuring that it can handle various complex tasks with ease. It is worth mentioning that the Dimensity 9300+ is the first in the industry to achieve a higher-speed Llama27B end-side large model operation, with an amazing speed…

According to MediaTek's official news, during the Generative AI Forum held today, MediaTek's generative AI service platform, MediaTek DaVinci, also known as "MediaTek DaVinci," was officially launched, and more than 40 vendors have joined its ecosystem. It is reported that the tool is initially an internal generative AI tool for the Group, and is widely used in requirements analysis and specification design for software development, automatic matchmaking for human resources, reimbursement process for finance, patent translation for legal affairs, and contract litigation, etc., with the Group's penetration rate reaching 96%. At the same time, MediaTek's Innovation Base also released the platform's latest...

MediaTek announced that it has successfully deployed the Tongyi Qianwen large model on flagship chips such as Dimensity 9300, realizing the deep adaptation of large models on mobile phone chips for the first time. Tongyi Qianwen runs multiple rounds of AI dialogues offline. Alibaba Cloud said that it will work closely with MediaTek to provide end-side large model solutions to global mobile phone manufacturers. Tongyi Qianwen has currently open-sourced large language models with 1.8 billion, 7 billion, 14 billion, and 72 billion parameters, as well as multimodal large models for visual understanding and audio understanding. Alibaba Cloud also released Tongyi Qianwen 2.0 in October last year, with model parameters reaching thousands of...

MediaTek Research, a research institute under MediaTek, recently announced the launch of a new open source large language model (LLM) called MR Breeze-7B. This open source model is good at processing traditional Chinese and English, with a total of 7 billion parameters, and is designed based on the acclaimed Mistral model. Compared with the previous generation product BLOOM-3B, MR Breeze-7B has absorbed 20 times more knowledge, allowing it to handle traditional Chinese with higher accuracy…