OSSForks/tensorflow
1
0
Fork 0
mirror of https://github.com/zebrajr/tensorflow.git synced 2025-12-06 00:19:58 +01:00
Code Issues Actions 5 Packages Projects Releases Wiki Activity

Support int64 for mul, this will be needed for i64 tensor_scatter_updates.

Browse Source 
This also enables support for squared_difference for i64

PiperOrigin-RevId: 381160512
Change-Id: I42a4cb34aa510d2d886afd75c5636bd421c3485b
This commit is contained in:
Renjie Liu 2021-06-23 18:53:31 -07:00 committed by TensorFlower Gardener
parent a3b7535ec5
commit 94ee5b9c8f
14 changed files with 40 additions and 15 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

1
RELEASE.md
View File

@ -81,6 +81,7 @@
* `tf.lite`:
* The recommended Android NDK version for building TensorFlow Lite has
been changed from r18b to r19c.
* Supports int64 for mul.
* `tf.saved_model`:
* SavedModels can now save custom gradients. Use the option
`tf.saved_model.SaveOption(experimental_custom_gradients=True)` to

8
tensorflow/compiler/mlir/lite/ir/tfl_ops.cc
View File

@ -248,10 +248,10 @@ bool VerifyMulOpShapeConstraints(MulOp op) {
/*max_bcast_rank=*/4);
}
// Allows I32, QI16 and F32 outputs when the operands have valid shapes, which
// are broadcastable shapes up to four dimension or have same shapes.
if (IsI32Type(element_type) || IsQI16Type(element_type) ||
element_type.isF32()) {
// Allows I32, I64, QI16 and F32 outputs when the operands have valid shapes,
// which are broadcastable shapes up to four dimension or have same shapes.
if (IsI32Type(element_type) || IsI64Type(element_type) ||
IsQI16Type(element_type) || element_type.isF32()) {
return VerifyOperandsHaveSameShapesOrBroadcastableShape(
/*op=*/op.getOperation(), /*indices=*/ArrayRef<unsigned>{0, 1},
/*max_bcast_rank=*/4);

6
tensorflow/compiler/mlir/lite/ir/tfl_ops.td
View File

@ -2338,11 +2338,11 @@ def TFL_MulOp : TFL_Op<"mul", [
}];
let arguments = (
ins TFL_TensorOf<[F32, I32, QI8, QUI8, QI16]>:$lhs,
TFL_TensorOf<[F32, I32, QI8, QUI8, QI16]>:$rhs,
ins TFL_TensorOf<[F32, I32, I64, QI8, QUI8, QI16]>:$lhs,
TFL_TensorOf<[F32, I32, I64, QI8, QUI8, QI16]>:$rhs,
TFL_AFAttr:$fused_activation_function);
let results = (outs TFL_TensorOf<[F32, I32, QI8, QUI8, QI16]>:$output);
let results = (outs TFL_TensorOf<[F32, I32, I64, QI8, QUI8, QI16]>:$output);
let hasFolder = 1;

9
tensorflow/compiler/mlir/lite/tests/legalize-tf.mlir
View File

@ -2149,3 +2149,12 @@ func @conv3d_transpose_unsupported_strides(%arg0: tensor<2x5x6x8x2xf32>, %arg1:
// CHECK-LABEL: conv3d_transpose_unsupported_strides
// CHECK: "tf.Conv3DBackpropInputV2"
}
func @mul_i64(%arg0: tensor<14xi64>, %arg1: tensor<14xi64>) -> tensor<14xi64> {
%0 = "tf.Mul"(%arg0, %arg1) : (tensor<14xi64>, tensor<14xi64>) -> tensor<14xi64>
return %0: tensor<14xi64>
// CHECK-LABEL: mul_i64
// CHECK: tfl.mul %arg0, %arg1 {fused_activation_function = "NONE"} : tensor<14xi64>
// CHECK: return
}

1
tensorflow/lite/experimental/mlir/testing/generate_examples.py
View File

@ -89,7 +89,6 @@ MLIR_CONVERTER_KNOWN_BUGS = {
# int64.
r"div.*dtype=tf\.int64": "119126484",
r"floor_div.*dtype=tf\.int64": "119126484",
r"mul.*dtype=tf\.int64": "119126484",
r"relu.*dtype=tf\.int64": "119126484",
r"squared_difference.*dtype=tf\.int64": "119126484",
# Post-training quantization support missing for below op in mlir.

2
tensorflow/lite/experimental/mlir/testing/generated_examples_zip_test.cc
View File

@ -74,9 +74,7 @@ const std::map<string, string>& GetKnownBrokenTests() {
{R"(^\/where.*1,2,3,1)", "134692786"},
{R"(^\/div.*dtype=tf\.int64)", "119126484"},
{R"(^\/mul.*dtype=tf\.int64)", "119126484"},
{R"(^\/floor_div.*dtype=tf\.int64)", "119126484"},
{R"(^\/squared_difference.*dtype=tf\.int64)", "119126484"},
});
return *kBrokenTests;
}

2
tensorflow/lite/experimental/mlir/testing/op_tests/tensor_scatter_update.py
View File

@ -29,7 +29,7 @@ def make_tensor_scatter_update_tests(options):
"""Make a set of tests to do tensor_scatter_update."""
test_parameters = [{
"input_dtype": [tf.float32, tf.int32],
"input_dtype": [tf.float32, tf.int32, tf.int64],
"input_shape": [[14], [2, 4, 7]],
"updates_count": [1, 3, 5],
}]

9
tensorflow/lite/kernels/mul.cc
View File

@ -157,6 +157,12 @@ void EvalMul(TfLiteContext* context, TfLiteNode* node, TfLiteMulParams* params,
TF_LITE_MUL(optimized_ops, Mul, float);
}
}
} else if (output->type == kTfLiteInt64) {
if (need_broadcast) {
TF_LITE_MUL(reference_ops, BroadcastMul4DSlow, int64_t);
} else {
TF_LITE_MUL(reference_ops, Mul, int64_t);
}
}
#undef TF_LITE_MUL
}
@ -275,7 +281,8 @@ TfLiteStatus Eval(TfLiteContext* context, TfLiteNode* node) {
TF_LITE_ENSURE_OK(context,
GetOutputSafe(context, node, kOutputTensor, &output));
if (output->type == kTfLiteFloat32 || output->type == kTfLiteInt32) {
if (output->type == kTfLiteFloat32 || output->type == kTfLiteInt32 ||
output->type == kTfLiteInt64) {
EvalMul<kernel_type>(context, node, params, data, input1, input2, output);
} else if (output->type == kTfLiteUInt8 || output->type == kTfLiteInt8 ||
output->type == kTfLiteInt16) {

2
tensorflow/lite/kernels/register.cc
View File

@ -100,7 +100,7 @@ BuiltinOpResolver::BuiltinOpResolver() {
/* min_version = */ 1,
/* max_version = */ 3);
AddBuiltin(BuiltinOperator_MUL, Register_MUL(), /* min_version = */ 1,
/* max_version = */ 4);
/* max_version = */ 5);
AddBuiltin(BuiltinOperator_L2_NORMALIZATION, Register_L2_NORMALIZATION(),
/* min_version = */ 1,
/* max_version = */ 2);

2
tensorflow/lite/kernels/register_ref.cc
View File

@ -265,7 +265,7 @@ BuiltinRefOpResolver::BuiltinRefOpResolver() {
/* min_version = */ 1,
/* max_version = */ 3);
AddBuiltin(BuiltinOperator_MUL, Register_MUL_REF(), /* min_version = */ 1,
/* max_version = */ 4);
/* max_version = */ 5);
AddBuiltin(BuiltinOperator_L2_NORMALIZATION, Register_L2NORM_REF(),
/* min_version = */ 1,
/* max_version = */ 2);

1
tensorflow/lite/testing/generated_examples_zip_test.cc
View File

@ -101,7 +101,6 @@ const std::map<string, string>& GetKnownBrokenTests() {
{R"(^\/floor_mod.*activation=True.*dtype=tf\.int64)", "112968789"},
{R"(^\/div.*dtype=tf\.int64)", "119126484"},
{R"(^\/mul.*dtype=tf\.int64)", "119126484"},
{R"(^\/floor_div.*dtype=tf\.int64)", "119126484"},
{R"(^\/squared_difference.*dtype=tf\.int64)", "119126484"},
});

4
tensorflow/lite/tools/versioning/op_version.cc
View File

@ -228,6 +228,10 @@ int GetBuiltinOperatorVersion(const OpSignature& op_sig) {
}
case BuiltinOperator_MUL:
// Version 5 supports int64 inputs
if (op_sig.inputs.at(0).type == kTfLiteInt64) {
return 5;
}
// Version 4 supports int16 inputs
if (op_sig.inputs.at(0).type == kTfLiteInt16) {
return 4;

7
tensorflow/lite/tools/versioning/op_version_test.cc
View File

@ -432,6 +432,13 @@ TEST(OpVersionTest, VersioningSubTest) {
SimpleVersioningTest(BuiltinOperator_SUB);
}
TEST(OpVersionTest, VersioningMUL5Test) {
OpSignature fake_op_sig;
fake_op_sig.op = BuiltinOperator_MUL;
fake_op_sig.inputs = CreateOpSignatureTensorSpecs(kTfLiteInt64);
EXPECT_EQ(GetBuiltinOperatorVersion(fake_op_sig), 5);
}
TEST(OpVersionTest, VersioningSub4Test) {
OpSignature fake_op_sig = {
.op = BuiltinOperator_SUB,

1
tensorflow/lite/tools/versioning/runtime_version.cc
View File

@ -146,6 +146,7 @@ std::string FindMinimumRuntimeVersionForOp(tflite::BuiltinOperator op_code,
{{BuiltinOperator_MUL, 2}, "1.14.0"},
{{BuiltinOperator_MUL, 3}, "1.15.0"},
{{BuiltinOperator_MUL, 4}, "2.3.0"},
{{BuiltinOperator_MUL, 5}, "2.6.0"},
{{BuiltinOperator_NON_MAX_SUPPRESSION_V4, 1}, "2.1.0"},
{{BuiltinOperator_NON_MAX_SUPPRESSION_V5, 1}, "2.1.0"},
{{BuiltinOperator_PAD, 1}, "1.5.0"},