Tether, the issuer of the leading stablecoin USDT, has announced the launch of an innovative AI training framework designed to facilitate the fine-tuning of large language models on consumer-grade hardware. This framework, integrated within the QVAC platform, allows for the utilization of smartphones and non-Nvidia GPUs, broadening the accessibility of AI model development.
According to Tether’s announcement, the framework leverages Microsoft’s BitNet architecture and LoRA techniques to significantly reduce memory and computational demands. This advancement aims to lower both the cost and hardware barriers typically associated with AI model training.
Cross-Platform Capabilities
The framework supports cross-platform training and inference across various chipsets, including those from AMD, Intel, and Apple Silicon, as well as mobile GPUs from Qualcomm and Apple. Tether’s engineers have demonstrated the capability to fine-tune models with up to 1 billion parameters on smartphones in under two hours, while smaller models can be adjusted in mere minutes. The system can also accommodate models as large as 13 billion parameters on mobile devices.
Efficiency and Performance
Built on the 1-bit model architecture of BitNet, the framework is reported to reduce VRAM requirements by up to 77.8% compared to traditional 16-bit models. This efficiency allows larger models to operate on hardware with limited resources. Additionally, the performance improvements extend to inference, with mobile GPUs executing BitNet models several times faster than CPUs.
Potential Use Cases
Tether highlighted several potential applications for this framework, including on-device training and federated learning. These methods enable models to be updated across distributed devices without the need to transmit data to centralized servers, thereby reducing dependence on cloud infrastructure.
This initiative aligns with a broader trend of crypto companies venturing into AI and machine learning, as evidenced by recent investments and developments in the sector. As Tether forges ahead with its AI infrastructure, it joins a growing movement that seeks to integrate advanced computational capabilities into everyday devices.
This article was produced by NeonPulse.today using human and AI-assisted editorial processes, based on publicly available information. Content may be edited for clarity and style.
