removing nvtx range used for debugging (#485)

apex/parallel/optimized_sync_batchnorm_kernel.py

@@ -8,7 +8,6 @@ class SyncBatchnormFunction(Function):
 
     @staticmethod
     def forward(ctx, input, z, weight, bias, running_mean, running_variance, eps, track_running_stats = True, momentum = 1.0, process_group = None, channel_last = False, fuse_relu = False):
-        torch.cuda.nvtx.range_push("sync_BN_fw")
         input = input.contiguous()
         world_size = 0
 
@@ -64,13 +63,11 @@ class SyncBatchnormFunction(Function):
         else:
             out = syncbn.batchnorm_forward(input, mean, inv_std, weight, bias)
 
-        torch.cuda.nvtx.range_pop()
         return out
 
     @staticmethod
     def backward(ctx, grad_output):
         grad_output = grad_output.contiguous()
-        torch.cuda.nvtx.range_push("sync_BN_bw")
         # mini batch mean & var are calculated by forward path.
         # mu = 1./N*np.sum(h, axis = 0)
         # var = 1./N*np.sum((h-mu)**2, axis = 0)
@@ -113,5 +110,4 @@ class SyncBatchnormFunction(Function):
         if weight is None or not ctx.needs_input_grad[3]:
             grad_bias = None
 
-        torch.cuda.nvtx.range_pop()
         return grad_input, grad_z, grad_weight, grad_bias, None, None, None, None, None, None, None, None