Unverified50% confidenceFactExact time
Gshard（2020）首次实现了万亿参数MoE模型的分布式训练，Switch Transformer（2022）将每个token的路由选择收敛为单一专家
1
Sources
50%
Confidence
Long-term
Relevance
7/20/2026
First Seen
Sources
Related Claims
UnverifiedSwitch Transformer将专家数量扩展至2048个，2022年谷歌GLaM模型进一步验证了MoE在多任务泛化上的优势77% similarUnverifiedSwitch Transformer, proposed in 2021, simplified the number of experts routed to per token to 1.70% similarVerifiedGoogle's Switch Transformer, released in 2021, applied MoE at scale to Transformer language models, expanding model capacity to the trillion-parameter level.70% similarVerifiedGoogle proposed the Transformer architecture in 2017, BERT emerged in 2018, and the GPT series of models appeared after 2020, pushing NLP into the era of large models with pre-training plus fine-tuning.68% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/567255API
curl https://kongchang.com/api/v1/knowledge/claims/567255MCP
get_claim(id=567255)