Gigabyte has announced the launch of AI TOP, its in-house software utility designed to bring advanced AI model training capabilities to home users. Making its first appearance at this year’s Computex, AI TOP allows users to locally train and fine-tune AI models with a capacity of up to 236 billion parameters when used with recommended hardware.
AI TOP is essentially a comprehensive solution for local AI model fine-tuning, enhancing privacy and security for sensitive data while providing maximum flexibility and real-time adjustments. According to Gigabyte, the utility comes with a user-friendly interface and has been designed to help beginners and experienced users easily navigate and understand the information and settings. Additionally, the utility includes AI TOP Tutor, which offers various AI TOP solutions, setup guidance, and technical support for all types of AI model operators.
Currently, the software supports over 70 open-source LLM models from Hugging Face, including popular options such as Baichuan 2, Distill-GPT2, GLM4, Llama 2, and Llama 3. Despite Hugging Face hosting over 770,000 models, AI TOP’s selection is limited by the memory capacity constraints inherent in desktop PC hardware.
The free-to-use software is optimized for Gigabyte hardware, including motherboards, SSDs, graphics cards, and power supplies. Supported GPUs include most of Gigabyte’s Nvidia GeForce RTX 40 series, AMD Radeon RX 7900 series, and Radeon Pro W7900 and W7800 series. For optimal performance, Gigabyte has listed AI TOP-branded hardware on its dedicated website, featuring a power supply, a motherboard, two SSDs, and three graphics cards.
Training AI models locally offers enhanced privacy and security by keeping data on your system, cost efficiency by avoiding cloud service fees, and faster processing with reduced latency. It provides customization and control over the training environment, allows offline capabilities, and enables scalable hardware resource management.
Users benefit from immediate feedback and iteration, eliminating data transfer limitations, and making full use of existing hardware. This approach fosters hands-on learning and experimentation, making it ideal for novice and experienced users seeking a secure, flexible, and efficient AI training solution.
You can download the AI TOP utility free of cost from Gigabyte’s official web page. While it is available for all users, it is only supported on Linux systems.
Mozilla is giving Firefox’s Smart Window two new tricks. If you haven’t tried it yet, Smart Window is an optional AI mode in Firefox with a built-in chat that helps you get things done while you browse. With this update, it can fill out online forms using details from your tabs and keep an eye on websites for you.
Tired of typing the same details into every form?
The absolutely absurd idea of putting a touchscreen display on a mouse isn't exactly new. Turtle Beach already did it with its Command Series MC7, which features a 2.25-inch control screen. But apparently, even that wasn't ridiculous enough for Arctic Fox Technologies. Its new Super Mouse looks like someone took the chunky old Atari ST mouse, slapped a touchscreen and scroll wheel on it, and called it an upgrade.
The result is a rather unusual-looking mouse that is now on Kickstarter, starting at $199. And as ridiculous as it looks, some of its features are genuinely intriguing.
Apple is less than a day away from launching its highly anticipated OLED MacBook Pro with a touch-sensitive screen. But it appears that the upgrade is going to cost a pretty penny. According to a report by Bloomberg, Apple is reportedly eyeing a starting price of $2,899 for the 14-inch MacBook Pro, while the 16-inch version could start at $3,399 in the US.
"At $2,899, the 14-inch model would mark the highest starting price for a new MacBook Pro model, surpassing the $2,799 Apple charged for a 17-inch version introduced in 2006," says the report. The report claims that Apple is targeting a minimum price hike of $400 for its upcoming redesigned 14-inch and 16-inch MacBook Pro models.