The ImageNet architectures scale from 18 to 152 weighted layers while retaining lower reported forward complexity than VGG-16/19 at the deepest setting; the 34-layer baseline is reported as 18% of VGG-19's complexity.

The ImageNet architectures scale from 18 to 152 weighted layers while retaining lower reported forward complexity than VGG-16/19 at the deepest setting; the 34-layer baseline is reported as 18% of VGG-19's complexity. · CiteArk

测量结果

已评估 0/8 项

获得支持 0有争议或冲突 0尚未评估 8

保留同一结论下的完整结果组。已有评估排在前面；单项判断不替代整组结论。展开行查看条件、来源与历次证据。

每行是一项测量，分别显示论文报告、最近可用实测和累计证据状态。
论文报告	最近可用实测	证据判断
1,800,000,000 FLOPs (multiply-adds)	—	尚未评估
3,600,000,000 FLOPs (multiply-adds)	—	尚未评估
3,800,000,000 FLOPs (multiply-adds)	—	尚未评估
7,600,000,000 FLOPs (multiply-adds)	—	尚未评估
11,300,000,000 FLOPs (multiply-adds)	—	尚未评估
15,300,000,000 FLOPs (multiply-adds)	—	尚未评估
19,600,000,000 FLOPs (multiply-adds)	—	尚未评估
18 percent of VGG-19 FLOPs	—	尚未评估

The ImageNet architectures scale from 18 to 152 weighted layers while retaining lower reported forward complexity than VGG-16/19 at the deepest setting; the 34-layer baseline is reported as 18% of VGG-19's complexity.

报告指标与观测值

Assessment（0）

关联运行（0）

测量结果