Unverified50% confidenceFactExact time
文中示例模型总参数约11万，其中Attention模块占1.6万，FFN占3.3万
1
Sources
50%
Confidence
Long-term
Relevance
9/11/2026
First Seen
Sources
Related Entities
Related Claims
Unverified在演示数据集上训练15轮后，文本分类模型准确率达到99%，模型总参数11万（Attention 1.6万，FFN 3.3万，其余为Embedding）75% similarUnverified1000亿参数的模型处理单次推理请求需约80GB显存，百亿参数级别模型仅需约20GB71% similarUnverified模型最终参数量为13,693,824个，落在10-30百万目标区间内69% similarUnverified大模型词汇表通常在32,000到100,000个token之间68% similarUnverified一个15亿参数的模型使用Adam优化器时每个GPU至少需要约24GB的状态信息67% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/904318API
curl https://kongchang.com/api/v1/knowledge/claims/904318MCP
get_claim(id=904318)