Fix -Wmissing-braces warnings in f8f8bf16_tensorwise.cu

Reviewed By: wenxin0319

Differential Revision: D72594079

fbshipit-source-id: 78eba167cd68055f4b3b2a20c16114ae5bbf4cd3

Author: r-barnes

fbgemm_gpu/experimental/gen_ai/src/quantize/cutlass_extensions/f8f8bf16_tensorwise.cu

@@ -197,7 +197,7 @@ at::Tensor f8f8bf16_tensorwise_impl(
        stride_output}};
 
   arguments.epilogue.thread = {
-      {float(scale)}, // scale
+      {{float(scale)}}, // scale
       {}, // Accumulator
       {}, // Multiplies
   };