Tether releases cross-platform BitNet LoRA framework, enabling fine-tuning of billion-parameter models on consumer-grade devices

Techub News reports that Tether has announced the launch of a cross-platform BitNet LoRA fine-tuning framework within QVAC Fabric, enabling training and inference optimization for Microsoft BitNet (1-bit LLM). This framework significantly reduces computational and memory requirements, allowing billion-parameter-scale models to be trained and fine-tuned on laptops, consumer-grade GPUs, and even smartphones. This solution marks the first implementation of BitNet model fine-tuning on mobile GPUs (including Adreno, Mali, and Apple Bionic). Tests show that a 125M parameter model can be fine-tuned in approximately 10 minutes, a 1B model within about an hour, and it can even scale to a 13B parameter model on mobile devices. Furthermore, the framework supports heterogeneous hardware such as Intel, AMD, and Apple Silicon, achieving the first-ever 1-bit LLM LoRA fine-tuning on non-NVIDIA devices. In terms of performance, BitNet models achieve 2 to 11 times faster inference speeds on mobile GPUs compared to CPUs, while reducing memory usage by up to approximately 77.8% compared to traditional 16-bit models. Tether states that this technology has the potential to break the dependence on high-end computing power and cloud infrastructure, promote the decentralization and localization of AI training, and provide a foundation for new application scenarios such as federated learning.