AI Research & Engineering: RecSys, Search, NLP, Generative AI and Beyond

Tag ASR

Streaming ASR 实战:Chunked Attention、KV Cache、Look-ahead 全解析(流式语音识别架构与源码详解)

本文从工程视角彻底拆透流式 ASR:算法延迟 vs 计算延迟、流式三大天敌、Chunked Attention 与 Dynamic Chunk Training、KV Cache、Causal Conv、Whisper 流式化、RNN-T 天然流式、VAD + Endpoint 工业架构、Moshi/GPT-4o Realtime 端到端语音 LLM。CTC、Whisper、RNN-T、Conformer、SSL 系列的姊妹篇。

Loading

© 2026 Yudong‘s Blog — Powered by WordPress

Theme by Anders NorenUp ↑