Unverified50% confidenceBenchmarkExact time
Conformer 系列由 Google 于 2020 年提出，将自注意力的全局建模能力与深度可分离卷积的局部特征提取能力相结合，在 LibriSpeech 基准上曾创下 SOTA
1
Sources
50%
Confidence
Long-term
Relevance
7/22/2026
First Seen
Sources
Related Claims
UnverifiedThe Conformer (Convolution-augmented Transformer) architecture was proposed by Google in 2020, combining the global self-attention mechanism of Transformers with the local feature extraction capability of convolutional neural networks.74% similarVerified思维链（Chain-of-Thought）提示技术由Google Research于2022年提出72% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/585372API
curl https://kongchang.com/api/v1/knowledge/claims/585372MCP
get_claim(id=585372)