r"""ResNetV1d variant described in `Bag of Tricks `_. Compared with default ResNet(ResNetV1b), ResNetV1d replaces the 7x7 conv in the input stem with three 3x3 convs. And in the downsampling block, a 2x2 avg_pool with stride 2 is added before co
source not stored for this graph (policy: none)
no outgoing calls