FonaTech/Project_Chronos

⚡ Zero-Stall MoE Inference via Lookahead Prediction & Async DMA Prefetching. Optimized for SSD I/O with Hybrid MLA+Sliding Window Attention.

⭐ 203 ⑂ 21 Python Apache-2.0 · 2026-04-26推送

artificial-intelligence generative-ai large-language-model llm lora mixture-of-experts open-models open-source high-throughput async-dma dual-layer-moe io-latency-hiding lookahead-routing mla-attention predictive-inference sliding-window-attention ssd-offloading streaming-llm

203

Watchers

0

贡献者

0

Commits

0

Releases

0

Open Issues

2026-04-26

最近推送

原文中文

暂无 README