When running at scale... we end up hanging/getting blocked by pytorch when downloading the resnet model.
"https://download.pytorch.org/models/resnet34-333f7ec4.pth" to /root/.cache/torch/hub/checkpoints/resnet34-333f7ec4.pth
instead of doing this at runtime, we should do this during docker build so it's not downloaded thousands of time by the batch operators.
When running at scale... we end up hanging/getting blocked by pytorch when downloading the resnet model.
"https://download.pytorch.org/models/resnet34-333f7ec4.pth" to /root/.cache/torch/hub/checkpoints/resnet34-333f7ec4.pthinstead of doing this at runtime, we should do this during docker build so it's not downloaded thousands of time by the batch operators.