Update mlir tests to reflect new namespace change

PiperOrigin-RevId: 638391003

backends/cpu/mlir_tests/mnist/btf_kernels.mlir

@@ -28,15 +28,15 @@ func.func @tensor_io() {
   %one = tfrt.constant.i32 1
   %two = tfrt.constant.i32 2
 
-  %t0 = "btf.read_dense_tensor.i32.2"(%path, %zero) : (!tfrt.string, i32) -> (!t.tensor)
+  %t0 = "btf.read_dense_tensor.i32.2"(%path, %zero) : (!tfrt.string, i32) -> (!tfrt_tensor.tensor)
   // CHECK-NEXT: shape = [2, 2], values = [1, 2, 3, 4]
   %c1 = tfrt_dht.print_tensor %t0, %c0
 
-  %t1 = "btf.read_dense_tensor.i32.1"(%path, %one) : (!tfrt.string, i32) -> (!t.tensor)
+  %t1 = "btf.read_dense_tensor.i32.1"(%path, %one) : (!tfrt.string, i32) -> (!tfrt_tensor.tensor)
   // CHECK-NEXT: shape = [5], values = [0, 1, 2, 3, 4]
   %c2 = tfrt_dht.print_tensor %t1, %c1
 
-  %t2 = "btf.read_dense_tensor.i32.1"(%path, %two) : (!tfrt.string, i32) -> (!t.tensor)
+  %t2 = "btf.read_dense_tensor.i32.1"(%path, %two) : (!tfrt.string, i32) -> (!tfrt_tensor.tensor)
   // CHECK-NEXT: shape = [0], values = []
   %c3 = tfrt_dht.print_tensor %t2, %c2
 
@@ -49,7 +49,7 @@ func.func @tensor_io_invalid_path() {
   %path = "tfrt_test.get_string"() { value = "/tmp/invalid_path" } : () -> !tfrt.string
   %zero = tfrt.constant.i32 0
   // expected-error @+1 {{failed to open file /tmp/invalid_path for reading}}
-  %t0 = "btf.read_dense_tensor.i32.2"(%path, %zero) : (!tfrt.string, i32) -> (!t.tensor)
+  %t0 = "btf.read_dense_tensor.i32.2"(%path, %zero) : (!tfrt.string, i32) -> (!tfrt_tensor.tensor)
 
   tfrt.return
 }

backends/cpu/mlir_tests/resnet/max_pool.mlir

@@ -26,20 +26,20 @@ func.func @test_max_pool_2d_f32_0() {
 
   %input_index = tfrt.constant.i32 0
   %input = "btf.read_dense_tensor.f32.4"(%path, %input_index)
-    : (!tfrt.string, i32) -> (!t.tensor)
+    : (!tfrt.string, i32) -> (!tfrt_tensor.tensor)
 
   %expected_index = tfrt.constant.i32 1
   %expected = "btf.read_dense_tensor.f32.4"(%path, %expected_index)
-    : (!tfrt.string, i32) -> (!t.tensor)
+    : (!tfrt.string, i32) -> (!tfrt_tensor.tensor)
 
   %output = "tfrt_dht.create_uninitialized_tensor.f32.4"() { shape = [2 : i64, 1 : i64, 1 : i64, 6 : i64] }
-    : () -> !t.tensor
+    : () -> !tfrt_tensor.tensor
   %ch1 = "tfrt_test.max_pooling_2d.f32"(%input, %output, %ch0)
     { padding = "valid", pool_size = [3 : i32, 3 : i32], strides = [3 : i32, 3 : i32] }
-    :  (!t.tensor, !t.tensor, !tfrt.chain) -> !tfrt.chain
+    :  (!tfrt_tensor.tensor, !tfrt_tensor.tensor, !tfrt.chain) -> !tfrt.chain
 
   %cmp, %ch2 = "tfrt_dht.tensor_allclose.f32"(%expected, %output, %ch1)
-    : (!t.tensor, !t.tensor, !tfrt.chain) -> (i1, !tfrt.chain)
+    : (!tfrt_tensor.tensor, !tfrt_tensor.tensor, !tfrt.chain) -> (i1, !tfrt.chain)
 
   // CHECK: int1 = 1
   tfrt.print.i1 %cmp, %ch2
@@ -57,20 +57,20 @@ func.func @test_max_pool_2d_f32_1() {
 
   %input_index = tfrt.constant.i32 2
   %input = "btf.read_dense_tensor.f32.4"(%path, %input_index)
-    : (!tfrt.string, i32) -> (!t.tensor)
+    : (!tfrt.string, i32) -> (!tfrt_tensor.tensor)
 
   %expected_index = tfrt.constant.i32 3
   %expected = "btf.read_dense_tensor.f32.4"(%path, %expected_index)
-    : (!tfrt.string, i32) -> (!t.tensor)
+    : (!tfrt.string, i32) -> (!tfrt_tensor.tensor)
 
   %output = "tfrt_dht.create_uninitialized_tensor.f32.4"() { shape = [2 : i64, 3 : i64, 3 : i64, 6 : i64] }
-    : () -> !t.tensor
+    : () -> !tfrt_tensor.tensor
   %ch1 = "tfrt_test.max_pooling_2d.f32"(%input, %output, %ch0)
     { padding = "same", pool_size = [3 : i32, 3 : i32], strides = [2 : i32, 2 : i32] }
-    :  (!t.tensor, !t.tensor, !tfrt.chain) -> !tfrt.chain
+    :  (!tfrt_tensor.tensor, !tfrt_tensor.tensor, !tfrt.chain) -> !tfrt.chain
 
   %cmp, %ch2 = "tfrt_dht.tensor_allclose.f32"(%expected, %output, %ch1)
-    : (!t.tensor, !t.tensor, !tfrt.chain) -> (i1, !tfrt.chain)
+    : (!tfrt_tensor.tensor, !tfrt_tensor.tensor, !tfrt.chain) -> (i1, !tfrt.chain)
 
   // CHECK: int1 = 1
   tfrt.print.i1 %cmp, %ch2

backends/cpu/mlir_tests/resnet/resnet_tensor_kernels.mlir

@@ -20,16 +20,16 @@ func.func @test_flatten_f32() {
 
   %t1 = "tfrt_dht.create_uninitialized_tensor.f32.4"()
     { shape = [1 : i64, 2 : i64, 2 : i64, 1 : i64] }
-    : () -> !t.tensor
+    : () -> !tfrt_tensor.tensor
   %ch1 = "tfrt_dht.set_tensor_with_constant_values.f32"(%t1, %ch0)
     { values = [1.0 : f32, 2.0 : f32, 3.0 : f32, 4.0 : f32] }
-    : (!t.tensor, !tfrt.chain) -> !tfrt.chain
+    : (!tfrt_tensor.tensor, !tfrt.chain) -> !tfrt.chain
 
   %t2 = "tfrt_dht.create_uninitialized_tensor.f32.2"()
     { shape = [1 : i64, 4 : i64] }
-    : () -> !t.tensor
+    : () -> !tfrt_tensor.tensor
   %ch2 = "tfrt_test.flatten.f32"(%t1, %t2, %ch1)
-    : (!t.tensor, !t.tensor, !tfrt.chain) -> !tfrt.chain
+    : (!tfrt_tensor.tensor, !tfrt_tensor.tensor, !tfrt.chain) -> !tfrt.chain
 
   // CHECK: shape = [1, 4], values = [1.000000e+00, 2.000000e+00, 3.000000e+00, 4.000000e+00]
   tfrt_dht.print_tensor %t2, %ch2
@@ -43,15 +43,15 @@ func.func @test_max_pool_2d_f32_padding_error() {
 
   %input = "tfrt_dht.create_uninitialized_tensor.f32.4"()
     { shape = [2 : i64, 3 : i64, 3 : i64, 6 : i64] }
-    : () -> !t.tensor
+    : () -> !tfrt_tensor.tensor
   %output = "tfrt_dht.create_uninitialized_tensor.f32.4"()
     { shape = [2 : i64, 1 : i64, 1 : i64, 6 : i64] }
-    : () -> !t.tensor
+    : () -> !tfrt_tensor.tensor
 
   // expected-error @+1 {{padding 'invalid' is not recognized}}
   "tfrt_test.max_pooling_2d.f32"(%input, %output, %ch0)
     { padding = "invalid", pool_size = [3 : i32, 3 : i32], strides = [3 : i32, 3 : i32] }
-    :  (!t.tensor, !t.tensor, !tfrt.chain) -> !tfrt.chain
+    :  (!tfrt_tensor.tensor, !tfrt_tensor.tensor, !tfrt.chain) -> !tfrt.chain
 
   tfrt.return
 }
@@ -62,15 +62,15 @@ func.func @test_max_pool_2d_f32_shape_error() {
 
   %input = "tfrt_dht.create_uninitialized_tensor.f32.4"()
     { shape = [2 : i64, 3 : i64, 3 : i64, 6 : i64] }
-    : () -> !t.tensor
+    : () -> !tfrt_tensor.tensor
   %output = "tfrt_dht.create_uninitialized_tensor.f32.4"()
     { shape = [2 : i64, 2 : i64, 1 : i64, 6 : i64] }
-    : () -> !t.tensor
+    : () -> !tfrt_tensor.tensor
 
   // expected-error @+1 {{output shape [2, 2, 1, 6] does not match the expected output shape [2, 1, 1, 6]}}
   "tfrt_test.max_pooling_2d.f32"(%input, %output, %ch0)
     { padding = "valid", pool_size = [3 : i32, 3 : i32], strides = [3 : i32, 3 : i32] }
-    :  (!t.tensor, !t.tensor, !tfrt.chain) -> !tfrt.chain
+    :  (!tfrt_tensor.tensor, !tfrt_tensor.tensor, !tfrt.chain) -> !tfrt.chain
 
   tfrt.return
 }

include/tfrt/tensor/opdefs/tensor.h

@@ -28,10 +28,9 @@ using namespace mlir;
 namespace tfrt {
 namespace tfrt_tensor {
 
-// TODO (b/341154040): Pass in "tfrt_tensor" instead of "t".
 class TensorDialect : public Dialect {
  public:
-  static StringRef getDialectNamespace() { return "t"; }
+  static StringRef getDialectNamespace() { return "tfrt_tensor"; }
   explicit TensorDialect(MLIRContext *context);
 
   Type parseType(DialectAsmParser &parser) const override;
@@ -43,7 +42,7 @@ class TensorDialect : public Dialect {
 class TensorType : public Type::TypeBase<TensorType, Type, TypeStorage> {
  public:
   using Base::Base;
-  static constexpr StringLiteral name = "tfrt.t.tensor";
+  static constexpr StringLiteral name = "tfrt.tfrt_tensor.tensor";
 };
 
 }  // namespace tfrt_tensor

lib/tensor/opdefs/tensor.cc

@@ -32,10 +32,8 @@ namespace tfrt_tensor {
 // TensorShape Dialect
 //===----------------------------------------------------------------------===//
 
-// TODO (b/341154040): Pass in "tfrt_tensor" into the Dialect constructor
-// instead of "t".
 TensorDialect::TensorDialect(MLIRContext *context)
-    : Dialect(/*name=*/"t", context, TypeID::get<TensorDialect>()) {
+    : Dialect(/*name=*/"tfrt_tensor", context, TypeID::get<TensorDialect>()) {
   allowUnknownTypes();
   addTypes<TensorType>();
   addOperations<

mlir_tests/tensor/string_host_tensor.mlir

@@ -19,7 +19,7 @@ func.func @basic() {
   %c0 = tfrt.new.chain
 
   %a = "tfrt_sht.create_tensor"()
-    {shape = [2], values = ["string", "tensor"]} : () -> !t.tensor
+    {shape = [2], values = ["string", "tensor"]} : () -> !tfrt_tensor.tensor
 
   // CHECK: shape = [2], values = ["string", "tensor"]
   %c1 = tfrt_dht.print_tensor %a, %c0