1
Llamas on the Web: Memory-Efficient, Performance-Portable, and Multi-Precision LLM Inference with WebGPU
只需29-33%更少内存，LlamaWeb在16种设备上实现高性能多精度LLM推理，彻底改变浏览器端大模型部署。
arXiv:2605.20706v1 Announce Type: cross Abstract: Running language models in the browser presents a unique opportunity to build efficient, private, an…