Update BERT model optimization python script (#2521)

Add support of GPT2 model optimization:
* Match subgraph of Gelu Approximation (using Tanh).
* Fuse LayerNormalization if SkipLayerNormalization is not ready.
* Output model even if embedding layer is not fused.
* Improve Reshape Fusion to improve coverage.
* Refine constant input checking, and output fused op counter.

Update script according to latest op improvements:
* Fusion of Add Bias and Gelu.
* Fuse SkipLayerNormalization and Add Bias.

Other:
* Add ReduceSum for mask as intermediate step.
* Refactor verbose setting.

onnxruntime/python/tools/bert/README.md

@@ -52,4 +52,7 @@ See below for description of all the options:
 - **input_int32**: (*optional*)
     Exported model ususally uses int64 tensor as input. If this flag is specified, int32 tensors will be used as input, and it could avoid un-necessary Cast nodes and get better performance.
 - **float16**: (*optional*)
-    By default, model uses float32 in computation. If this flag is specified, half-precision float will be used. This option is recommended for NVidia GPU with Tensor Core like V100 and T4. For older GPUs, float32 is likely faster.
+    By default, model uses float32 in computation. If this flag is specified, half-precision float will be used. This option is recommended for NVidia GPU with Tensor Core like V100 and T4. For older GPUs, float32 is likely faster.
+- **verbose**: (*optional*)
+    Print verbose information when this flag is specified.
+