3 ms·
FWIW, I believe that the current state of the art for batch-size 1, fp32 inference for ResNet-50 on Intel CPUs is AWS's work in https://arxiv.org/abs/1809.02697
by ajtulloch 8y ago
FWIW, I believe that the current state of the art for batch-size 1, fp32 inference for ResNet-50 on Intel CPUs is AWS's work in https://arxiv.org/abs/1809.02697 https://arxiv.org/abs/1809.02697. After the low-hanging fruit outside of model execution are picked, this kind of work is probably quite relevant.
- human_afterall 8y agoHey! Author here, thanks for linking the paper. The article was from an infrastructure perspective, but we're definitely diving deeper into graph execution optimizations after this:)
- rahulun 8y agoAre there any particular optimizations you are looking into ?