update time stamp for function signature: due to the while loop, the …

…lstm function signature will have like tensor<1x?x10xf32>, but if we know the input, we can derive the timestamp.

PiperOrigin-RevId: 299236850
Change-Id: I4bf8d93fd724919a6829e72cae19af551bfeb873

tensorflow/compiler/mlir/lite/utils/lstm_utils.cc

@@ -588,16 +588,6 @@ LogicalResult CreateEqualSizeSplitVOp(Value input, int axis, int splits,
   return success();
 }
 
-void UpdateFuncSignature(int batch, int time, int output,
-                         mlir::FuncOp* func_op) {
-  SmallVector<int64_t, 4> output_shape{batch, time, output};
-  auto input_types = func_op->getType().getInputs();
-  auto element_type = input_types[0].cast<RankedTensorType>().getElementType();
-  auto output_type = mlir::RankedTensorType::get(output_shape, element_type);
-  func_op->setType(
-      mlir::FunctionType::get(input_types, output_type, func_op->getContext()));
-}
-
 // TODO(b/147436982): Consider refactor this to be more general.
 LogicalResult ConvertKerasLSTMLayer(mlir::FuncOp func_op, OpBuilder* builder) {
   // For argument order, please check out standard_lstm under
@@ -729,19 +719,29 @@ LogicalResult ConvertKerasLSTMLayer(mlir::FuncOp func_op, OpBuilder* builder) {
   }
 
   SmallVector<Value, 5> outputs;
+  SmallVector<Type, 5> output_types;
 
+  // Due to the existence of the while loop, the timestamp may be unknown
+  // for the signature, for us, since we know the inputs, we can infer the time
+  // steps.
   for (int i = 0; i < 5; ++i) {
     if (i == 1) {
       // only this one is the real output.
       outputs.push_back(final_output);
+      output_types.push_back(final_output.getType());
     } else {
       auto result_type =
           func_op.getCallableResults()[i].dyn_cast<RankedTensorType>();
       outputs.push_back(CreatTfF32ConstOp(builder, result_type.getShape(), 0.0f,
                                           func_op.getLoc()));
+      output_types.push_back(result_type);
     }
   }
 
+  // Update function signatures.
+  func_op.setType(mlir::FunctionType::get(func_op.getType().getInputs(),
+                                          output_types, func_op.getContext()));
+
   builder->create<mlir::ReturnOp>(func_op.getLoc(), outputs);
   return success();
 }