torchscript_mask_rcnn.cpp: fix compilation error

Summary:
Pull Request resolved: #4240

The API changed on this line to return an rvalue instead of an lvalue (atomic<int>& -> size_t) from D33801501. I've followed that diff to what seems to be the new API.

Reviewed By: ebyrne, eellison, wat3rBro

Differential Revision: D36394835

fbshipit-source-id: 792d80661338d0ad62033c711e75198c226b520c

Author: miguelmartin75

tools/deploy/torchscript_mask_rcnn.cpp

@@ -135,7 +135,8 @@ int main(int argc, const char* argv[]) {
       export_method == "caffe2_tracing" || export_method == "tracing" ||
       export_method == "scripting");
 
-  torch::jit::getBailoutDepth() = 1;
+  torch::jit::FusionStrategy strat = {{torch::jit::FusionBehavior::DYNAMIC, 1}};
+  torch::jit::setFusionStrategy(strat);
   torch::autograd::AutoGradMode guard(false);
   auto module = torch::jit::load(argv[1]);