c10::optional -> std::optional in a few places (pytorch#144340)

Test Plan: Sandcastle

Pull Request resolved: pytorch#144340
Approved by: https://github.com/malfet

torch/csrc/api/include/torch/nn/parallel/data_parallel.h

@@ -275,7 +275,7 @@ Tensor data_parallel(
     return module->forward(std::move(input)).to(*output_device);
   }
 
-  autograd::Scatter scatter(*devices, /*chunk_sizes=*/nullopt, dim);
+  autograd::Scatter scatter(*devices, /*chunk_sizes=*/std::nullopt, dim);
   auto scattered_inputs = fmap<Tensor>(scatter.apply({std::move(input)}));
   // Input tensor might not be big enough to scale across all available devices
   if (scattered_inputs.size() < devices->size()) {