'मह्लो' बोली

संचालन

mhlo.abs (mhlo::AbsOp)

एब्स ऑपरेशन

वाक्य - विन्यास:

operation ::= `mhlo.abs` $operand attr-dict
              `:` custom<SameOperandsAndResultType>(type($operand), type($result))

operand टेंसर पर तत्व-वार एब्स ऑपरेशन करता है और result टेंसर उत्पन्न करता है।

देखें: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#abs

उदाहरण:

%result = mhlo.abs %operand : tensor<3xi32>

लक्षण: AlwaysSpeculatableImplTrait , Elementwise , SameOperandsAndResultShape

इंटरफेस: ConditionallySpeculatable , InferShapedTypeOpInterface , InferTypeOpInterface , NoMemoryEffect (MemoryEffectOpInterface)

प्रभाव: MemoryEffects::Effect{}

संकार्य:

परिणाम:

mhlo.add (mhlo::AddOp)

ऑपरेशन जोड़ें

वाक्य - विन्यास:

operation ::= `mhlo.add` $lhs `,` $rhs attr-dict
              `:` custom<SameOperandsAndResultType>(type($lhs), type($rhs), type($result))

दो टेंसर lhs और rhs का तत्व-वार जोड़ करता है और एक result टेंसर उत्पन्न करता है।

देखें: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#add

उदाहरण:

%result = mhlo.add %lhs, %rhs : tensor<2x2xi32>

लक्षण: AlwaysSpeculatableImplTrait , Commutative , CompatibleOperandsAndResultType , Elementwise , SameOperandsAndResultShape

इंटरफेस: ConditionallySpeculatable , InferShapedTypeOpInterface , InferTypeOpInterface , NoMemoryEffect (MemoryEffectOpInterface)

प्रभाव: MemoryEffects::Effect{}

संकार्य:

परिणाम:

mhlo.add_dependency (mhlo::AddDependencyOp)

निर्भरता जोड़ें ऑपरेशन

वाक्य - विन्यास:

operation ::= `mhlo.add_dependency` operands attr-dict `:` functional-type(operands, results)

यह ऑपरेशन XLA कंपाइलर के लिए निजी है, इसलिए इसमें अभी तक कोई विनिर्देश नहीं है।

अनौपचारिक रूप से, यह ऑपरेशन दो ऑपरेंड: एक डेटा ऑपरेंड और एक टोकन। ऑपरेशन का आउटपुट डेटा ऑपरेंड है। जब AfterAll के साथ उपयोग किया जाता है तो यह ऑपरेशन गैर-साइड-इफेक्टिंग ऑपरेशंस (वे जो टोकन मान उत्पन्न नहीं करते हैं) को ऑर्डर करने में सक्षम बनाता है।

उदाहरण:

%1 = mhlo.add_dependency %arg0, %0 : (tensor<3x4xf32>, !mhlo.token) -> tensor<3x4xf32>

लक्षण: AlwaysSpeculatableImplTrait

इंटरफेस: ConditionallySpeculatable , InferTypeOpInterface , NoMemoryEffect (MemoryEffectOpInterface)

प्रभाव: MemoryEffects::Effect{}

संकार्य:

परिणाम:

mhlo.after_all (mhlo::AfterAllOp)

आख़िरकार ऑपरेशन

वाक्य - विन्यास:

operation ::= `mhlo.after_all` $inputs attr-dict
              `:` custom<VariadicSameOperandsAndResultType>(ref($inputs), type($inputs), type($result))

यह सुनिश्चित करता है कि inputs उत्पन्न करने वाले ऑपरेशन result पर निर्भर किसी भी ऑपरेशन से पहले निष्पादित किए जाते हैं।

देखें: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#after_all

उदाहरण:

%result = mhlo.after_all %input0, %input1 : !mhlo.token

लक्षण: AlwaysSpeculatableImplTrait

इंटरफेस: ConditionallySpeculatable , InferTypeOpInterface , NoMemoryEffect (MemoryEffectOpInterface)

प्रभाव: MemoryEffects::Effect{}

संकार्य:

परिणाम:

mhlo.all_gather (mhlo::AllGatherOp)

ऑलगैदर ऑपरेशन

प्रक्रिया ग्रिड में प्रत्येक प्रक्रिया समूह के भीतर, प्रत्येक प्रक्रिया से ऑपरेंड टेंसर के मूल्यों को all_gather_dim के साथ जोड़ता है और एक परिणाम टेंसर उत्पन्न करता है। operands में प्रत्येक ऑपरेंड के लिए computation अलग से लागू की जाती है, जिससे प्रति ऑपरेंड एक परिणाम उत्पन्न होता है।

देखें: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#all_gather

उदाहरण:

%result = "mhlo.all_gather"(%operand) {
  all_gather_dim = 1 : i64,
  replica_groups = dense<[[0, 1]]> : tensor<1x2xi64>
  // channel_id = 0
  channel_handle = #mhlo.channel_handle<handle = 0, type = 0>,
  // use_global_device_ids = false
} : (tensor<2x2xf32>) -> tensor<2x4xf32>

लक्षण: SameOperandsAndResultElementType

गुण:

संकार्य:

परिणाम:

mhlo.all_reduce (mhlo::AllReduceOp)

ऑलरिड्यूस ऑपरेशन

प्रक्रिया ग्रिड में प्रत्येक प्रक्रिया समूह के भीतर, प्रत्येक प्रक्रिया से एक ऑपरेंड टेंसर के मानों के लिए एक कमी फ़ंक्शन computation लागू करता है और एक परिणाम टेंसर उत्पन्न करता है। operands में प्रत्येक ऑपरेंड के लिए computation अलग से लागू की जाती है, जिससे प्रति ऑपरेंड एक परिणाम उत्पन्न होता है।

देखें: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#all_reduce

उदाहरण:

%result = "mhlo.all_reduce"(%operand) ({
  ^bb0(%arg0: tensor<f32>, %arg1: tensor<f32>):
    %0 = mhlo.add %arg1, %arg2 : tensor<f32>
    mhlo.return %0 : tensor<f32>
}) {
  replica_groups = dense<[[0, 1]]> : tensor<1x2xi64>
  // channel_id = 0
  channel_handle = #mhlo.channel_handle<handle = 0, type = 0>
  // use_global_device_ids = false
} : (tensor<4xf32>) -> tensor<4xf32>

लक्षण: InferTensorType , SingleBlockImplicitTerminator<ReturnOp> , SingleBlock

इंटरफ़ेस: InferShapedTypeOpInterface , InferTypeOpInterface

गुण:

संकार्य:

परिणाम:

mhlo.all_to_all (mhlo::AllToAllOp)

ऑलटूऑल ऑपरेशन

प्रक्रिया ग्रिड में प्रत्येक प्रक्रिया समूह के भीतर, operand टेंसर के मानों को split_dimension के साथ भागों में विभाजित करता है, प्रक्रियाओं के बीच विभाजित भागों को बिखेरता है, बिखरे हुए हिस्सों को concat_dimension के साथ जोड़ता है और एक result टेंसर उत्पन्न करता है।

देखें: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#all_to_all

उदाहरण:

%result = "mhlo.all_to_all"(%operand) {
  split_dimension = 1 : i64,
  concat_dimension = 0 : i64,
  split_count = 2 : i64,
  replica_groups = dense<[[0, 1]]> : tensor<1x2xi64>
} : (tensor<2x4xf32>) -> tensor<4x2xf32>

लक्षण: AlwaysSpeculatableImplTrait , InferTensorType , SameOperandsElementType , SameOperandsShape , SameVariadicOperandSize

इंटरफेस: ConditionallySpeculatable , InferShapedTypeOpInterface , InferTypeOpInterface , NoMemoryEffect (MemoryEffectOpInterface)

प्रभाव: MemoryEffects::Effect{}

गुण:

संकार्य:

परिणाम:

mhlo.and (mhlo::AndOp)

और ऑपरेशन

वाक्य - विन्यास:

operation ::= `mhlo.and` $lhs `,` $rhs attr-dict
              `:` custom<SameOperandsAndResultType>(type($lhs), type($rhs), type($result))

दो टेंसर lhs और rhs का तत्व-वार AND निष्पादित करता है और एक result टेंसर उत्पन्न करता है

देखें: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#and

उदाहरण:

%result = mhlo.and %lhs, %rhs : tensor<2x2xi32>

लक्षण: AlwaysSpeculatableImplTrait , Commutative , CompatibleOperandsAndResultType , Elementwise , SameOperandsAndResultShape

इंटरफेस: ConditionallySpeculatable , InferShapedTypeOpInterface , InferTypeOpInterface , NoMemoryEffect (MemoryEffectOpInterface)

प्रभाव: MemoryEffects::Effect{}

संकार्य:

परिणाम:

mhlo.async_done (mhlo::AsyncDoneOp)

AsyncDone ऑपरेशन

यह ऑपरेशन XLA कंपाइलर के लिए निजी है, इसलिए इसमें अभी तक कोई विनिर्देश नहीं है।

अनौपचारिक रूप से, यह ऑपरेशन अतुल्यकालिक गणना के अंत तक ब्लॉक हो जाता है। यह अतुल्यकालिक गणना का अंतिम परिणाम लौटाता है।

अधिक जानकारी के लिए AsyncStart के दस्तावेज़ देखें।

इंटरफ़ेस: InferTypeOpInterface

गुण:

संकार्य:

परिणाम:

mhlo.async_start (mhlo::AsyncStartOp)

AsyncStart ऑपरेशन

यह ऑपरेशन XLA कंपाइलर के लिए निजी है, इसलिए इसमें अभी तक कोई विनिर्देश नहीं है।

अनौपचारिक रूप से, यह ऑपरेशन एक अतुल्यकालिक गणना शुरू करता है।

इसका उपयोग तब किया जाता है जब ऐसे फ़ंक्शन होते हैं जिनमें अतुल्यकालिक प्रतीक्षा (जैसे डीएमए) और ऑन-थ्रेड गणना दोनों शामिल होते हैं। उदाहरण के लिए, एक फ़ंक्शन में एक गणना, एक डीएमए, एक अन्य गणना, एक दूसरा डीएमए और एक अंतिम गणना शामिल हो सकती है। इसे async_start के बाद async_update और async_done के रूप में दर्शाया जाएगा। async_start पहली गणना ऑन-थ्रेड करेगा और फिर DMA प्रारंभ करेगा। यदि यह अभी तक पूरा नहीं हुआ है तो async_update डीएमए के पूरा होने की प्रतीक्षा करेगा, फिर फ़ंक्शन में दूसरी गणना निष्पादित करेगा, और दूसरा डीएमए शुरू करेगा। अंत में, async_done इस अंतिम DMA पर प्रतीक्षा करेगा, और फिर अंतिम गणना चलाएगा जिसे थ्रेड पर चलाने की आवश्यकता है और उस अंतिम गणना का परिणाम लौटाएगा।

operands सीधे गणना के लिए पास कर दिया जाता है called_computation वह फ़ंक्शन है जिसे एसिंक्रोनस रूप से चलाया जाएगा execution_thread उस थ्रेड का नाम है जिसमें इसे चलाया जाएगा। मुख्य धागे को "मुख्य" कहा जाता है। सभी धागों के नाम हैं.

यह async ऑप्स के बीच आवश्यक सभी स्थिति लौटाता है। बफ़र असाइनमेंट के बाद, रिटर्न मान इनपुट, परिणाम और एसिंक ऑप द्वारा आवश्यक या संपादित किए गए किसी भी स्क्रैचपैड को रखने के लिए आवश्यक स्थान का प्रतिनिधित्व करता है।

गुण:

संकार्य:

परिणाम:

mhlo.async_update (mhlo::AsyncUpdateOp)

AsyncUpdate ऑपरेशन

यह ऑपरेशन XLA कंपाइलर के लिए निजी है, इसलिए इसमें अभी तक कोई विनिर्देश नहीं है।

अनौपचारिक रूप से, यह ऑपरेशन एक सिंक बाधा तक अतुल्यकालिक गणना पर रोक लगाता है। यह इस पर काम करने के बाद bundle लौटाता है।

अधिक जानकारी के लिए AsyncStart के दस्तावेज़ देखें।

इंटरफ़ेस: InferTypeOpInterface

गुण:

संकार्य:

परिणाम:

mhlo.atan2 (mhlo::Atan2Op)

Atan2 ऑपरेशन

वाक्य - विन्यास:

operation ::= `mhlo.atan2` $lhs `,` $rhs attr-dict
              `:` custom<SameOperandsAndResultType>(type($lhs), type($rhs), type($result))

lhs और rhs टेंसर पर तत्व-वार atan2 ऑपरेशन करता है और result टेंसर उत्पन्न करता है।

देखें: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#atan2

उदाहरण:

%result = mhlo.atan2 %lhs, %rhs : tensor<3xf32>

लक्षण: AlwaysSpeculatableImplTrait , CompatibleOperandsAndResultType , Elementwise , SameOperandsAndResultShape

इंटरफेस: ConditionallySpeculatable , InferShapedTypeOpInterface , InferTypeOpInterface , NoMemoryEffect (MemoryEffectOpInterface)

प्रभाव: MemoryEffects::Effect{}

संकार्य:

परिणाम:

mhlo.batch_norm_grad (mhlo::BatchNormGradOp)

बैचनॉर्मग्रैड ऑपरेशन

grad_output से बैकप्रोपेगेटिंग करके BatchNormTrainingOp के कई इनपुट के ग्रेडिएंट की गणना करता है, और grad_operand , grad_scale और grad_offset टेंसर तैयार करता है।

देखें: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#batch_norm_grad

उदाहरण:

%grad_operand, %grad_scale, %grad_offset =
"mhlo.batch_norm_grad"(%operand, %scale, %mean, %variance, %grad_output) {
  epsilon = 0.0 : f32,
  feature_index = 2 : i64
} : (tensor<2x2x2xf32>, tensor<2xf32>, tensor<2xf32>, tensor<2xf32>,
    tensor<2x2x2xf32>) -> (tensor<2x2x2xf32>, tensor<2xf32>, tensor<2xf32>)

लक्षण: AlwaysSpeculatableImplTrait , InferTensorType

इंटरफेस: ConditionallySpeculatable , InferShapedTypeOpInterface , InferTypeOpInterface , NoMemoryEffect (MemoryEffectOpInterface)

प्रभाव: MemoryEffects::Effect{}

गुण:

संकार्य:

परिणाम:

mhlo.batch_norm_inference (mhlo::BatchNormInferenceOp)

बैचनॉर्मइंफेरेंस ऑपरेशन

feature_index आयाम को छोड़कर सभी आयामों में operand टेंसर को सामान्यीकृत करता है और एक result टेंसर उत्पन्न करता है।

देखें: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#batch_norm_inference

उदाहरण:

%result = "mhlo.batch_norm_inference"(%operand, %scale, %offset, %mean, %variance) {
  epsilon = 0.0 : f32,
  feature_index = 2 : i64
} : (tensor<2x2x2xf32>, tensor<2xf32>, tensor<2xf32>, tensor<2xf32>, tensor<2xf32>) -> tensor<2x2x2xf32>

लक्षण: AlwaysSpeculatableImplTrait , InferTensorType

इंटरफेस: ConditionallySpeculatable , InferShapedTypeOpInterface , InferTypeOpInterface , NoMemoryEffect (MemoryEffectOpInterface)

प्रभाव: MemoryEffects::Effect{}

गुण:

संकार्य:

परिणाम:

mhlo.batch_norm_training (mhlo::BatchNormTrainingOp)

बैचनॉर्मट्रेनिंग ऑपरेशन

बैच और स्थानिक आयामों में माध्य और भिन्नता की गणना करता है और feature_index आयाम में प्रत्येक सुविधा के लिए operand टेंसर को सामान्य करता है और output , batch_mean और batch_var टेंसर उत्पन्न करता है।

देखें: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#batch_norm_training

उदाहरण:

%output, %batch_mean, %batch_var = "mhlo.batch_norm_training"(%operand, %scale, %offset) {
  epsilon = 0.0 : f32,
  feature_index = 2 : i64
} : (tensor<2x2x2xf32>, tensor<2xf32>, tensor<2xf32>) -> (tensor<2x2x2xf32>, tensor<2xf32>, tensor<2xf32>)

लक्षण: AlwaysSpeculatableImplTrait , InferTensorType

इंटरफेस: ConditionallySpeculatable , InferShapedTypeOpInterface , InferTypeOpInterface , NoMemoryEffect (MemoryEffectOpInterface)

प्रभाव: MemoryEffects::Effect{}

गुण:

संकार्य:

परिणाम:

mhlo.bitcast (mhlo::BitcastOp)

बिटकास्ट ऑपरेशन

वाक्य - विन्यास:

operation ::= `mhlo.bitcast` operands attr-dict `:` functional-type(operands, results)

यह ऑपरेशन XLA कंपाइलर के लिए निजी है, इसलिए इसमें अभी तक कोई विनिर्देश नहीं है।

अनौपचारिक रूप से, यह ऑपरेशन इनपुट के आकार को इस तरह से बदल देता है कि तत्वों की भौतिक व्यवस्था अपरिवर्तित रहती है।

इस ऑपरेशन को "तत्वों की भौतिक व्यवस्था" को समझने के लिए लेआउट जानकारी की आवश्यकता है, और एमएचएलओ में लेआउट समर्थन वर्तमान में प्रगति पर है।

उदाहरण:

%0 = mhlo.bitcast %arg0 : (tensor<3x4xf32>) -> tensor<3x4x1xf32>

लक्षण: AlwaysSpeculatableImplTrait

इंटरफेस: ConditionallySpeculatable , NoMemoryEffect (MemoryEffectOpInterface)

प्रभाव: MemoryEffects::Effect{}

संकार्य:

परिणाम:

mhlo.bitcast_convert (mhlo::BitcastConvertOp)

बिटकास्ट कन्वर्ट ऑपरेशन

वाक्य - विन्यास:

operation ::= `mhlo.bitcast_convert` operands attr-dict `:` functional-type(operands, results)

operand टेंसर पर एक बिटकास्ट ऑपरेशन करता है और एक result टेंसर का उत्पादन करता है जहां पूरे operand टेंसर के बिट्स को result टेंसर के प्रकार का उपयोग करके पुन: व्याख्या किया जाता है।

देखें: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#bitcast_convert

उदाहरण:

%result = mhlo.bitcast_convert %operand : (tensor<2xf32>) -> tensor<2x4xi8>

लक्षण: AlwaysSpeculatableImplTrait

इंटरफेस: ConditionallySpeculatable , InferShapedTypeOpInterface , NoMemoryEffect (MemoryEffectOpInterface)

प्रभाव: MemoryEffects::Effect{}

संकार्य:

परिणाम:

mhlo.broadcast (mhlo::BroadcastOp)

प्रसारण संचालन

यह ऑपरेशन StableHLO से बाहर जा रहा है, इसलिए इसे विनिर्देश में शामिल नहीं किया गया है: https://github.com/openxla/stablehlo/issues/3

अनौपचारिक रूप से, यह ऑपरेशन XLA के प्रसारण के समान ही कार्य करता है: https://www.tensorflow.org/xla/operation_semantics#broadcast

उदाहरण:

%result = mhlo.broadcast %operand, sizes = [1, 2] : (tensor<3xi32>) -> tensor<1x2x3xi32>

लक्षण: AlwaysSpeculatableImplTrait , InferTensorType , SameOperandsAndResultElementType

इंटरफेस: ConditionallySpeculatable , InferShapedTypeOpInterface , InferTypeOpInterface , NoMemoryEffect (MemoryEffectOpInterface)

प्रभाव: MemoryEffects::Effect{}

गुण:

संकार्य:

परिणाम:

mhlo.broadcast_in_dim (mhlo::BroadcastInDimOp)

ब्रॉडकास्टइनडिम ऑपरेशन

operand टेंसर में डेटा को डुप्लिकेट करके इनपुट टेंसर के आयाम और/या रैंक का विस्तार करता है और एक result टेंसर उत्पन्न करता है।

देखें: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#broadcast_in_dim

उदाहरण:

%result = mhlo.broadcast_in_dim %operand, dims = [2, 1] : (tensor<1x3xi32>) -> tensor<2x3x2xi32>

लक्षण: AlwaysSpeculatableImplTrait , HLO_CompatibleOperandsAndResultElementType

इंटरफेस: ConditionallySpeculatable , NoMemoryEffect (MemoryEffectOpInterface)

प्रभाव: MemoryEffects::Effect{}

गुण:

संकार्य:

परिणाम:

mhlo.case (mhlo :: caseop)

केस -ऑपरेशन

index के मूल्य के आधार पर branches से बिल्कुल एक function निष्पादित करने से आउटपुट का उत्पादन करता है।

देखें: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#case

उदाहरण:

%result0, %result1 = "mhlo.case"(%index) ({
  mhlo.return %result_branch0, %result_branch0 : tensor<2xi64>, tensor<2xi64>
}, {
  mhlo.return %result_branch1, %result_branch1 : tensor<2xi64>, tensor<2xi64>
}) : (tensor<i32>) -> (tensor<2xi64>, tensor<2xi64>)

लक्षण: RecursiveMemoryEffects , SingleBlockImplicitTerminator<ReturnOp> , SingleBlock

इंटरफ़ेस: InferTypeOpInterface

संकार्य:

परिणाम:

mhlo.cbrt (mhlo :: cbrtop)

सीबीआरटी ऑपरेशन

वाक्य - विन्यास:

operation ::= `mhlo.cbrt` $operand attr-dict
              `:` custom<SameOperandsAndResultType>(type($operand), type($result))

operand टेंसर पर तत्व-वार क्यूबिक रूट ऑपरेशन करता है और एक result टेंसर का उत्पादन करता है।

देखें: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#cbrt

उदाहरण:

%result = mhlo.cbrt %operand : tensor<4xf32>

लक्षण: AlwaysSpeculatableImplTrait , CompatibleOperandsAndResultType , Elementwise , SameOperandsAndResultShape

इंटरफेस: ConditionallySpeculatable , InferShapedTypeOpInterface , InferTypeOpInterface , NoMemoryEffect (MemoryEffectOpInterface)

प्रभाव: MemoryEffects::Effect{}

संकार्य:

परिणाम:

mhlo.ceil (mhlo :: ceilop)

छत -प्रचालन

वाक्य - विन्यास:

operation ::= `mhlo.ceil` $operand attr-dict
              `:` custom<SameOperandsAndResultType>(type($operand), type($result))

operand टेंसर के तत्व-वार सीआईएल का प्रदर्शन करता है और एक result टेंसर का उत्पादन करता है।

देखें: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#ceil

उदाहरण:

%result = mhlo.ceil %operand : tensor<5xf32>

लक्षण: AlwaysSpeculatableImplTrait , CompatibleOperandsAndResultType , Elementwise , SameOperandsAndResultShape

इंटरफेस: ConditionallySpeculatable , InferShapedTypeOpInterface , InferTypeOpInterface , NoMemoryEffect (MemoryEffectOpInterface)

प्रभाव: MemoryEffects::Effect{}

संकार्य:

परिणाम:

mhlo.cholesky (mhlo :: choleskyop)

चोलीज़ का संचालन

मैट्रिस के एक बैच के चोलस्की अपघटन की गणना करता है।

देखें: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#cholesky

उदाहरण:

%result = mhlo.cholesky %a, lower = true : tensor<3x3xf32>

लक्षण: AlwaysSpeculatableImplTrait , InferTensorType , SameOperandsAndResultElementType

इंटरफेस: ConditionallySpeculatable , InferShapedTypeOpInterface , InferTypeOpInterface , NoMemoryEffect (MemoryEffectOpInterface)

प्रभाव: MemoryEffects::Effect{}

गुण:

संकार्य:

परिणाम:

mhlo.clamp (mhlo :: clampop)

क्लैंप ऑपरेशन

वाक्य - विन्यास:

operation ::= `mhlo.clamp` $min `,` $operand `,` $max attr-dict
              `:` custom<SameOperandsAndResultType>(type($min), type($operand), type($max), type($result))

एक न्यूनतम और अधिकतम मूल्य के बीच operand टेंसर के प्रत्येक तत्व को क्लैंप करता है और एक result टेंसर का उत्पादन करता है।

देखें: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#clamp

उदाहरण:

%result = mhlo.clamp %min, %operand, %max : tensor<3xi32>

लक्षण: AlwaysSpeculatableImplTrait , HLO_BroadcastingElementwise , InferTensorType , SameOperandsAndResultElementType

इंटरफेस: ConditionallySpeculatable , InferShapedTypeOpInterface , InferTypeOpInterface , NoMemoryEffect (MemoryEffectOpInterface)

प्रभाव: MemoryEffects::Effect{}

संकार्य:

परिणाम:

mhlo.collective_broadcast (MHLO :: COLLECTICBROADCASTOP)

सामूहिक -संचालन संचालन

प्रक्रिया ग्रिड में प्रत्येक प्रक्रिया समूह के भीतर, स्रोत प्रक्रिया से operand टेंसर का मूल्य लक्ष्य प्रक्रियाओं तक भेजें और result टेंसर का उत्पादन करें।

देखें: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#collective_broadcast

उदाहरण:

%result = "mhlo.collective_broadcast"(%operand) {
  replica_groups = dense<[[0, 1]]> : tensor<1x2xi64>,
  channel_handle = #mhlo.channel_handle<handle = 0, type = 0>
} : (tensor<1x2xi64>) -> tensor<1x2xi64>

लक्षण: CompatibleOperandsAndResultType

इंटरफेस: InferShapedTypeOpInterface , InferTypeOpInterface

गुण:

संकार्य:

परिणाम:

mhlo.collective_permute (MHLO :: COLLECTICPERMUTEOP)

सामूहिक संचालन

प्रक्रिया ग्रिड में प्रत्येक प्रक्रिया समूह के भीतर, स्रोत प्रक्रिया से operand टेंसर का मूल्य लक्ष्य प्रक्रिया में भेजता है और result टेंसर का उत्पादन करता है।

देखें: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#collective_permute

उदाहरण:

%result = "mhlo.collective_permute"(%operand) {
  source_target_pairs = dense<[[0, 1], [1, 2]]> : tensor<2x2xi64>,
  // channel_id = 0
  channel_handle = #mhlo.channel_handle<handle = 0, type = 0>
} : (tensor<4x2xf32>) -> tensor<4x2xf32>

लक्षण: AlwaysSpeculatableImplTrait , CompatibleOperandsAndResultType

इंटरफेस: ConditionallySpeculatable , InferShapedTypeOpInterface , InferTypeOpInterface , NoMemoryEffect (MemoryEffectOpInterface)

प्रभाव: MemoryEffects::Effect{}

गुण:

संकार्य:

परिणाम:

mhlo.compare (MHLO :: तुलना)

ऑपरेशन की तुलना करें

वाक्य - विन्यास:

operation ::= `mhlo.compare` $comparison_direction `,` $lhs `,` $rhs (`,` $compare_type^)?
              attr-dict `:` functional-type(operands, results)

comparison_direction और compare_type के अनुसार lhs और rhs टेंसर्स की तत्व-वार तुलना करता है, और एक result टेंसर का उत्पादन करता है।

देखें: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#compare

उदाहरण:

%result = mhlo.compare LT, %lhs, %rhs, FLOAT : (tensor<2xf32>, tensor<2xf32>) -> tensor<2xi1>

लक्षण: AlwaysSpeculatableImplTrait , Elementwise , InferTensorType , SameOperandsAndResultShape , SameOperandsElementType

इंटरफेस: ConditionallySpeculatable , InferShapedTypeOpInterface , InferTypeOpInterface , NoMemoryEffect (MemoryEffectOpInterface)

प्रभाव: MemoryEffects::Effect{}

गुण:

संकार्य:

परिणाम:

mhlo.complex (mhlo :: Complexop)

जटिल ऑपरेशन

वाक्य - विन्यास:

operation ::= `mhlo.complex` operands attr-dict
              `:` custom<ComplexOpType>(type($lhs), type($rhs), type($result))

वास्तविक और काल्पनिक मूल्यों, lhs और rhs की एक जोड़ी से एक जटिल मूल्य के लिए तत्व-वार रूपांतरण करता है, और एक result टेंसर का उत्पादन करता है।

देखें: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#complex

उदाहरण:

%result = mhlo.complex %lhs, %rhs : tensor<2xcomplex<f32>>

लक्षण: AlwaysSpeculatableImplTrait , Elementwise , SameOperandsAndResultShape , SameOperandsElementType

इंटरफेस: ConditionallySpeculatable , InferShapedTypeOpInterface , InferTypeOpInterface , NoMemoryEffect (MemoryEffectOpInterface)

प्रभाव: MemoryEffects::Effect{}

संकार्य:

परिणाम:

mhlo.composite (MHLO :: COMMESITEOP)

समग्र संचालन

वाक्य - विन्यास:

operation ::= `mhlo.composite` $name $inputs attr-dict `:` functional-type(operands, results)

अन्य StableHlo संचालन के एक ऑपरेशन (रचित) को एनकैप्सुलेट करता है, inputs और composite_attributes और प्रोड्यूसिंग results लेता है। ओपी के शब्दार्थ को decomposition विशेषता द्वारा लागू किया जाता है। composite ओपी को प्रोग्राम शब्दार्थ को बदलने के बिना इसके अपघटन के साथ प्रतिस्थापित किया जा सकता है। ऐसे मामलों में जहां अपघटन को रोकना एक ही ओपी शब्दार्थ प्रदान नहीं करता है, custom_call उपयोग करना पसंद करते हैं।

version फ़ील्ड ( 0 से डिफॉल्ट्स) का उपयोग एक समग्र शब्दार्थ बदलने पर निरूपित करने के लिए किया जाता है।

देखें: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#composite

उदाहरण:

%results = mhlo.composite "my.op" %arg0, %arg1 {
  decomposition = @my_op,
  composite_attributes = { my_attribute = "my_value" },
  version = 1 : i32
} : (tensor<f32>, tensor<f32>) -> tensor<f32>

इंटरफेस: SymbolUserOpInterface

गुण:

संकार्य:

परिणाम:

mhlo.compute_reshape_shape (mhlo :: computerAshapeshapeop)

कंप्यूटरशैपशेप संचालन

वाक्य - विन्यास:

operation ::= `mhlo.compute_reshape_shape` operands attr-dict `:` functional-type(operands, results)

यह ऑपरेशन प्रगति पर एक काम है, इसलिए यह अभी तक विनिर्देश में शामिल नहीं है: https://github.com/openxla/stablehlo/issues/8

अनौपचारिक रूप से, यह ऑपरेशन डायनेमिक्रशेप के लिए एक आउटपुट_शेप की गणना करता है, जो कि डायनेमिक्स के एक संचालक में तत्वों की संख्या संख्या से num_elements संख्या से है और TF के पुनर्वसन को प्रदान किया गया dynamic_shape आकार: https://www.tensorflow.org/api_docs/python/tf/tf/tf/tf

उदाहरण के लिए, num_elements = 12 और dynamic_shape = [2, -1] के लिए, result [2, 6] होने जा रहा है। यदि ऑपरेंड मान्य नहीं हैं (उदाहरण के लिए यदि आयाम समान रूप से तत्वों की संख्या को विभाजित नहीं करते हैं, या यदि आयामों में कई -1 मान हैं), तो यह अपरिभाषित व्यवहार की ओर जाता है।

उदाहरण:

%result = mhlo.compute_reshape_shape %num_elements, %dynamic_shape
       : (index, tensor<2xi32>) -> tensor<2xi32>

लक्षण: AlwaysSpeculatableImplTrait

इंटरफेस: ConditionallySpeculatable , NoMemoryEffect (MemoryEffectOpInterface)

प्रभाव: MemoryEffects::Effect{}

संकार्य:

परिणाम:

mhlo.concatenate (mhlo :: concatenateop)

संप्रदाय संचालन

दिए गए तर्कों के रूप में एक ही क्रम में dimension आयाम के साथ inputs में टेनर्स की एक वैरिडिक संख्या को समेटता है और एक result टेंसर का उत्पादन करता है।

देखें: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#concatenate

उदाहरण:

%result = mhlo.concatenate %input0, %input1, dim = 0 : (tensor<3x2xi64>, tensor<1x2xi64>) -> tensor<4x2xi64>

लक्षण: AlwaysSpeculatableImplTrait , SameOperandsAndResultElementType

इंटरफेस: ConditionallySpeculatable , InferShapedTypeOpInterface , InferTypeOpInterface , NoMemoryEffect (MemoryEffectOpInterface)

प्रभाव: MemoryEffects::Effect{}

गुण:

संकार्य:

परिणाम:

mhlo.constant (mhlo :: constantop)

निरंतर प्रचालन

एक निरंतर value से एक output टेंसर का उत्पादन करता है।

देखें: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#constant

उदाहरण:

%output = mhlo.constant dense<[[0.0, 1.0], [2.0, 3.0]]> : tensor<2x2xf32>

लक्षण: AlwaysSpeculatableImplTrait , ConstantLike

इंटरफेस: ConditionallySpeculatable , InferTypeOpInterface , NoMemoryEffect (MemoryEffectOpInterface)

प्रभाव: MemoryEffects::Effect{}

गुण:

परिणाम:

mhlo.convert (mhlo :: Convertop)

कन्वर्ट ऑपरेशन

वाक्य - विन्यास:

operation ::= `mhlo.convert` $operand attr-dict
              `:` custom<SameOperandsAndResultType>(type($operand), type($result))

एक तत्व-वार रूपांतरण एक तत्व प्रकार से दूसरे के लिए operand टेंसर पर करता है और एक result टेंसर का उत्पादन करता है।

देखें: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#convert

उदाहरण:

%result = mhlo.convert %operand : (tensor<3xi32>) -> tensor<3xcomplex<f32>>

लक्षण: AlwaysSpeculatableImplTrait , Elementwise , SameOperandsAndResultShape

इंटरफेस: ConditionallySpeculatable , InferShapedTypeOpInterface , NoMemoryEffect (MemoryEffectOpInterface)

प्रभाव: MemoryEffects::Effect{}

संकार्य:

परिणाम:

mhlo.convolution (MHLO :: Confolutionop)

कनवल्शन ऑपरेशन

वाक्य - विन्यास:

operation ::= `mhlo.convolution` `(`operands`)`
              `dim_numbers` `=` custom<ConvolutionDimensions>($dimension_numbers) `,`
              `window` `=` `{` custom<WindowAttributes>($window_strides, $padding,
              $lhs_dilation, $rhs_dilation,
              $window_reversal) `}`
              attr-dict `:` functional-type(operands, results)

lhs की खिड़कियों और rhs के स्लाइस के बीच डॉट उत्पादों की गणना करता है और result पैदा करता है।

देखें: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#convolution

उदाहरण:

%result = "mhlo.convolution"(%lhs, %rhs) {
  window_strides = dense<4> : tensor<2xi64>,
  padding = dense<0> : tensor<2x2xi64>,
  lhs_dilation = dense<2> : tensor<2xi64>,
  rhs_dilation = dense<1> : tensor<2xi64>,
  window_reversal = dense<false> : tensor<2xi1>,
  dimension_numbers = #mhlo.conv<[b, 0, 1, f]x[0, 1, i, o]->[b, 0, 1, f]>,
  feature_group_count = 1 : i64,
  batch_group_count = 1 : i64,
  precision_config = [#stablehlo<precision DEFAULT>, #stablehlo<precision DEFAULT>]
} : (tensor<1x4x4x1xi32>, tensor<3x3x1x1xi32>) -> tensor<1x2x2x1xi32>

लक्षण: AlwaysSpeculatableImplTrait

इंटरफेस: ConditionallySpeculatable , NoMemoryEffect (MemoryEffectOpInterface)

प्रभाव: MemoryEffects::Effect{}

गुण:

संकार्य:

परिणाम:

mhlo.copy (mhlo :: copyop)

प्रतिलिपि संचालन

वाक्य - विन्यास:

operation ::= `mhlo.copy` operands attr-dict
              `:` custom<SameOperandsAndResultType>(type($operand), type($result))

यह ऑपरेशन XLA कंपाइलर के लिए निजी है, इसलिए यह अभी तक कोई विनिर्देश नहीं है।

अनौपचारिक रूप से, यह ऑपरेशन operand की एक प्रति है। ऑपरेशन से जुड़े मेटाडेटा के आधार पर, यह नो-ऑप से काफी अलग व्यवहार कर सकता है।

उदाहरण:

%0 = mhlo.copy %arg0 : tensor<f32>

लक्षण: AlwaysSpeculatableImplTrait , CompatibleOperandsAndResultType , Elementwise

इंटरफेस: ConditionallySpeculatable , InferShapedTypeOpInterface , InferTypeOpInterface , NoMemoryEffect (MemoryEffectOpInterface)

प्रभाव: MemoryEffects::Effect{}

गुण:

संकार्य:

परिणाम:

mhlo.cosine (mhlo :: cosineop)

कोसाइन ऑपरेशन

वाक्य - विन्यास:

operation ::= `mhlo.cosine` $operand attr-dict
              `:` custom<SameOperandsAndResultType>(type($operand), type($result))

operand टेंसर पर तत्व-वार कोसाइन ऑपरेशन करता है और एक result टेंसर का उत्पादन करता है।

देखें: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#cosine

उदाहरण:

%result = mhlo.cosine %operand : tensor<2xf32>

लक्षण: AlwaysSpeculatableImplTrait , CompatibleOperandsAndResultType , Elementwise , SameOperandsAndResultShape

इंटरफेस: ConditionallySpeculatable , InferShapedTypeOpInterface , InferTypeOpInterface , NoMemoryEffect (MemoryEffectOpInterface)

प्रभाव: MemoryEffects::Effect{}

संकार्य:

परिणाम:

mhlo.count_leading_zeros (MHLO :: CLZOP)

सीएलजेड प्रचालन

वाक्य - विन्यास:

operation ::= `mhlo.count_leading_zeros` $operand attr-dict
              `:` custom<SameOperandsAndResultType>(type($operand), type($result))

operand टेंसर में अग्रणी शून्य बिट्स की संख्या की तत्व-वार गिनती करता है और एक result टेंसर का उत्पादन करता है।

देखें: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#count_leading_zeros

उदाहरण:

%result = mhlo.count_leading_zeros %operand : tensor<2x2xi8>

लक्षण: AlwaysSpeculatableImplTrait , CompatibleOperandsAndResultType , Elementwise , SameOperandsAndResultShape

इंटरफेस: ConditionallySpeculatable , InferShapedTypeOpInterface , InferTypeOpInterface , NoMemoryEffect (MemoryEffectOpInterface)

प्रभाव: MemoryEffects::Effect{}

संकार्य:

परिणाम:

mhlo.create_token (mhlo :: createTokenop)

CreateToken संचालन

वाक्य - विन्यास:

operation ::= `mhlo.create_token` attr-dict `:` type(results)

यह ऑपरेशन StableHlo से बाहर है, इसलिए यह विनिर्देश में शामिल नहीं है: https://github.com/openxla/stablehlo/issues/3

अनौपचारिक रूप से, यह ऑपरेशन 0 इनपुट के साथ आफ्टरलॉप के रूप में एक ही काम करता है: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#after_all

उदाहरण:

%output = mhlo.create_token : !mhlo.token

लक्षण: AlwaysSpeculatableImplTrait

इंटरफेस: ConditionallySpeculatable , InferTypeOpInterface , NoMemoryEffect (MemoryEffectOpInterface)

प्रभाव: MemoryEffects::Effect{}

परिणाम:

mhlo.cross-replica-sum (mhlo :: crossreplicasumop)

क्रॉस्रीप्लिकसुम संचालन

यह ऑपरेशन StableHlo से बाहर है, इसलिए यह विनिर्देश में शामिल नहीं है: https://github.com/openxla/stablehlo/issues/3

अनौपचारिक रूप से, यह ऑपरेशन channel_id = 0 , use_global_device_ids = false और computation के साथ Allreduceop के समान काम करता है।

उदाहरण:

%result = "mhlo.cross-replica-sum"(%operand) {
  replica_groups = dense<[[0, 1]]> : tensor<1x2xi64>
} : (tensor<4xf32>) -> tensor<4xf32>

लक्षण: AlwaysSpeculatableImplTrait , CompatibleOperandsAndResultType

इंटरफेस: ConditionallySpeculatable , InferShapedTypeOpInterface , InferTypeOpInterface , NoMemoryEffect (MemoryEffectOpInterface)

प्रभाव: MemoryEffects::Effect{}

गुण:

संकार्य:

परिणाम:

mhlo.cstr_reshapable (mhlo::CstrReshapableOp)

CstrReshapable operation

वाक्य - विन्यास:

operation ::= `mhlo.cstr_reshapable` operands attr-dict `:` functional-type(operands, results)

This operation is a work in progress, so it is not yet included in the specification: https://github.com/openxla/stablehlo/issues/8

Informally, this operation creates a witness on the constraint that ComputeReshapeShape would succeed with the provided operands.

उदाहरण:

%result = mhlo.cstr_reshapable %num_elements, %dynamic_shape
       : (index, tensor<3xi32>) -> !shape.witness

लक्षण: AlwaysSpeculatableImplTrait

इंटरफेस: ConditionallySpeculatable , InferTypeOpInterface , NoMemoryEffect (MemoryEffectOpInterface)

प्रभाव: MemoryEffects::Effect{}

संकार्य:

परिणाम:

mhlo.custom_call (mhlo::CustomCallOp)

CustomCall operation

वाक्य - विन्यास:

operation ::= `mhlo.custom_call` custom<CustomCallTarget>($call_target_name) `(` $inputs `)`
              attr-dict `:` functional-type(operands, results)

Encapsulates an implementation-defined operation call_target_name that takes inputs and called_computations and produces results .

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#custom_call

उदाहरण:

%results = "mhlo.custom_call"(%input0) {
  call_target_name = "foo",
  has_side_effect = false,
  backend_config = "bar",
  api_version = 1 : i32,
  called_computations = [@foo]
} : (tensor<f32>) -> tensor<f32>

A custom call invokes code external to XLA. The `inputs` are passed to the
external code, and the external code is expected to produce a result of the
given type. The exact mechanism is backend-specific. For example, in the CPU
backend, a call instruction is emitted which targets a symbol with the name
`call_target_name`.

If XLA runtime is enabled for a backend, then custom calls use the runtime
custom call calling convention to call into the external functions. This
calling convention defines an ABI for encoding arguments, attributes and
results.

Depending on the API version there are two ways to pass extra bits of static
information to the external function:

1. For `API_VERSION_TYPED_FFI` custom calls `backend_config` must be a
   dictionary attribute, that will be encoded according to the custom call
   calling convention and passed to the external function as the attributes
   argument. External code is expected to use declarative bindings (see
   `xla/runtime/custom_call.h`) to decode them at run time. These custom
   calls are only supported if XLA uses XLA runtime.

2. For previous API versions it is the user responsibility to encode extra
   bits of static information as a string `backend_config` attribute, and
   decode it at run time.

Interfaces: MemoryEffectOpInterface

गुण:

संकार्य:

परिणाम:

mhlo.divide (mhlo::DivOp)

Div operation

वाक्य - विन्यास:

operation ::= `mhlo.divide` $lhs `,` $rhs attr-dict
              `:` custom<SameOperandsAndResultType>(type($lhs), type($rhs), type($result))

Performs element-wise division of dividend lhs and divisor rhs tensors and produces a result tensor.

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#divide

उदाहरण:

%result = mhlo.divide %lhs, %rhs : tensor<4xf32>

Traits: AlwaysSpeculatableImplTrait , CompatibleOperandsAndResultType , Elementwise , SameOperandsAndResultShape

Interfaces: ConditionallySpeculatable , InferShapedTypeOpInterface , InferTypeOpInterface , NoMemoryEffect (MemoryEffectOpInterface)

प्रभाव: MemoryEffects::Effect{}

संकार्य:

परिणाम:

mhlo.domain (mhlo::DomainOp)

Domain operation

This operation is private to the XLA compiler, so it is does not yet have a specification.

Informally, these operations are used to group instructions with the same DomainMetadata property. ShardingMetadata is the main use case today to group instructions on the same device. Domain instructions provide two major benefits:

• Prevent unintentionally optimizing instructions across domains.
• Automatically assign the metadata of the instructions created in the domain. Without domain instructions, each HLO optimization pass would have to check and propagate the metadata, which would be easy to miss and also adds complexity to the compiler. Since domain instructions connect two different domains, each domain instruction is associated with two DomainMetadata -- one on the operand side and one on the user side of the domain.

Traits: AlwaysSpeculatableImplTrait , CompatibleOperandsAndResultType

Interfaces: ConditionallySpeculatable , InferShapedTypeOpInterface , InferTypeOpInterface , NoMemoryEffect (MemoryEffectOpInterface)

प्रभाव: MemoryEffects::Effect{}

गुण:

संकार्य:

परिणाम:

mhlo.dot (mhlo::DotOp)

Dot operation

This operation is on its way out of StableHLO, so it is not included in the specification: https://github.com/openxla/stablehlo/issues/3

Informally, this operation does the same thing as XLA's Dot: https://www.tensorflow.org/xla/operation_semantics#dot

उदाहरण:

%0 = mhlo.dot %arg0, %arg1 : (tensor<1x2xi32>, tensor<2x1xi32>) -> tensor<1x1xi32>

लक्षण: AlwaysSpeculatableImplTrait

इंटरफेस: ConditionallySpeculatable , NoMemoryEffect (MemoryEffectOpInterface)

प्रभाव: MemoryEffects::Effect{}

गुण:

संकार्य:

परिणाम:

mhlo.dot_general (mhlo::DotGeneralOp)

DotGeneral operation

Computes dot products between slices of lhs and slices of rhs and produces a result tensor.

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#dot_general

उदाहरण:

%result = "mhlo.dot_general"(%lhs, %rhs) {
  dot_dimension_numbers = #mhlo.dot<
    lhs_batching_dimensions = [0],
    rhs_batching_dimensions = [0],
    lhs_contracting_dimensions = [2],
    rhs_contracting_dimensions = [1]
  >,
  precision_config = [#stablehlo<precision DEFAULT>, #stablehlo<precision DEFAULT>]
} : (tensor<2x2x2xi32>, tensor<2x2x2xi32>) -> tensor<2x2x2xi32>

लक्षण: AlwaysSpeculatableImplTrait

Interfaces: ConditionallySpeculatable , InferShapedTypeOpInterface , NoMemoryEffect (MemoryEffectOpInterface)

प्रभाव: MemoryEffects::Effect{}

गुण:

संकार्य:

परिणाम:

mhlo.dynamic_broadcast_in_dim (mhlo::DynamicBroadcastInDimOp)

DynamicBroadcastInDim operation

This operation is a work in progress, so it is not yet included in the specification: https://github.com/openxla/stablehlo/issues/8

Informally, this operation does the same thing as BroadcastInDimOp except that the result shape is specified dynamically via output_dimensions : https://github.com/openxla/stablehlo/blob/main/docs/spec.md#broadcast_in_dim

It also accepts optional attributes to express static knowledge about the expanding behavior of dimensions. If not specified, all dimensions are assumed to be possibly expanding. The sets of dimensions that are known to be expanding and the set of dimensions that are known to be non-expanding must be disjoint and they must be a subset of the operand's dimensions.

लक्षण: AlwaysSpeculatableImplTrait

Interfaces: ConditionallySpeculatable , InferShapedTypeOpInterface , NoMemoryEffect (MemoryEffectOpInterface)

प्रभाव: MemoryEffects::Effect{}

गुण:

संकार्य:

परिणाम:

mhlo.dynamic_conv (mhlo::DynamicConvOp)

DynamicConv operation

This operation is a work in progress, so it is not yet included in the specification: https://github.com/openxla/stablehlo/issues/8

Informally, this operation does the same thing as ConvolutionOp except that padding is specified dynamically via d_padding : https://github.com/openxla/stablehlo/blob/main/docs/spec.md#convolution

उदाहरण:

%result = "mhlo.dynamic_conv"(%lhs, %rhs, %d_padding) {
  window_strides = dense<4> : tensor<2xi64>,
  lhs_dilation = dense<2> : tensor<2xi64>,
  rhs_dilation = dense<1> : tensor<2xi64>,
  window_reversal = dense<false> : tensor<2xi1>,
  dimension_numbers = #mhlo.conv<[b, 0, 1, f]x[0, 1, i, o]->[b, 0, 1, f]>,
  feature_group_count = 1 : i64,
  batch_group_count = 1 : i64,
  precision_config = [#stablehlo<precision DEFAULT>, #stablehlo<precision DEFAULT>]
} : (tensor<1x4x4x1xi32>, tensor<3x3x1x1xi32>, tensor<2x2xi64>) -> tensor<1x2x2x1xi32>

लक्षण: AlwaysSpeculatableImplTrait

इंटरफेस: ConditionallySpeculatable , NoMemoryEffect (MemoryEffectOpInterface)

प्रभाव: MemoryEffects::Effect{}

गुण:

संकार्य:

परिणाम:

mhlo.dynamic_gather (mhlo::DynamicGatherOp)

DynamicGather operation

This operation is a work in progress, so it is not yet included in the specification: https://github.com/openxla/stablehlo/issues/8

Informally, this operation does the same thing as GatherOp except that slice_sizes are specified dynamically: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#gather

उदाहरण:

%result = "mhlo.dynamic_gather"(%operand, %start_indices, %slice_sizes) {
  dimension_numbers = #mhlo.gather<
    offset_dims = [2, 3],
    collapsed_slice_dims = [0],
    start_index_map = [0, 2],
    index_vector_dim = 2>,
  indices_are_sorted = false
} : (tensor<3x4x2xi32>, tensor<2x3x2xi64>, tensor<3xi64>) -> tensor<2x3x2x2xi32>

Traits: AlwaysSpeculatableImplTrait , InferTensorType

Interfaces: ConditionallySpeculatable , InferShapedTypeOpInterface , InferTypeOpInterface , NoMemoryEffect (MemoryEffectOpInterface)

प्रभाव: MemoryEffects::Effect{}

गुण:

संकार्य:

परिणाम:

mhlo.dynamic_iota (mhlo::DynamicIotaOp)

DynamicIota operation

This operation is a work in progress, so it is not yet included in the specification: https://github.com/openxla/stablehlo/issues/8

Informally, this operation does the same thing as IotaOp except that the result shape is specified dynamically via output_shape : https://github.com/openxla/stablehlo/blob/main/docs/spec.md#iota

उदाहरण:

%0 = mhlo.dynamic_iota %arg0, dim = 0 : (tensor<1xindex>) -> tensor<4xi32>

लक्षण: AlwaysSpeculatableImplTrait

Interfaces: ConditionallySpeculatable , InferShapedTypeOpInterface , NoMemoryEffect (MemoryEffectOpInterface)

प्रभाव: MemoryEffects::Effect{}

गुण:

संकार्य:

परिणाम:

mhlo.dynamic_pad (mhlo::DynamicPadOp)

DynamicPad operation

वाक्य - विन्यास:

operation ::= `mhlo.dynamic_pad` operands attr-dict `:` functional-type(operands, results)

Dynamically Pads the operand , with amount of padding added at low-end/high-end/interior is passed through input tensors.

लक्षण: AlwaysSpeculatableImplTrait

Interfaces: ConditionallySpeculatable , InferShapedTypeOpInterface , NoMemoryEffect (MemoryEffectOpInterface)

प्रभाव: MemoryEffects::Effect{}

संकार्य:

परिणाम:

mhlo.dynamic_reshape (mhlo::DynamicReshapeOp)

DynamicReshape operation

वाक्य - विन्यास:

operation ::= `mhlo.dynamic_reshape` operands attr-dict `:` functional-type(operands, results)

This operation is a work in progress, so it is not yet included in the specification: https://github.com/openxla/stablehlo/issues/8

Informally, this operation does the same thing as ReshapeOp except that the result shape is specified dynamically via output_shape : https://github.com/openxla/stablehlo/blob/main/docs/spec.md#reshape

उदाहरण:

%0 = mhlo.dynamic_reshape %arg0, %shape : (tensor<?xf32>, tensor<2xindex>) -> tensor<?x?xf32>

लक्षण: AlwaysSpeculatableImplTrait

Interfaces: ConditionallySpeculatable , InferShapedTypeOpInterface , NoMemoryEffect (MemoryEffectOpInterface)

प्रभाव: MemoryEffects::Effect{}

संकार्य:

परिणाम:

mhlo.dynamic_slice (mhlo::DynamicSliceOp)

DynamicSlice operation

Extracts a slice from the operand using dynamically-computed starting indices and produces a result tensor.

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#dynamic_slice

उदाहरण:

%result = mhlo.dynamic_slice %operand, %start_indices0, %start_indices1, sizes = [2, 2]
  : (tensor<4x4xi32>, tensor<i64>, tensor<i64>) -> tensor<2x2xi32>

Traits: AlwaysSpeculatableImplTrait , InferTensorType

Interfaces: ConditionallySpeculatable , InferShapedTypeOpInterface , InferTypeOpInterface , NoMemoryEffect (MemoryEffectOpInterface)

प्रभाव: MemoryEffects::Effect{}

गुण:

संकार्य:

परिणाम:

mhlo.dynamic_update_slice (mhlo::DynamicUpdateSliceOp)

DynamicUpdateSlice operation

वाक्य - विन्यास:

operation ::= `mhlo.dynamic_update_slice` operands attr-dict `:` functional-type(operands, results)

Produces a result tensor which is equal to the operand tensor except that the slice starting at start_indices is updated with the values in update .

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#dynamic_update_slice

उदाहरण:

%result = mhlo.dynamic_update_slice %operand, %update, %start_indices0, %start_indices1
  : (tensor<4x4xi32>, tensor<2x2xi32>, tensor<i64>, tensor<i64>) -> tensor<4x4xi32>

Traits: AlwaysSpeculatableImplTrait , InferTensorType

Interfaces: ConditionallySpeculatable , InferShapedTypeOpInterface , InferTypeOpInterface , NoMemoryEffect (MemoryEffectOpInterface)

प्रभाव: MemoryEffects::Effect{}

संकार्य:

परिणाम:

mhlo.einsum (mhlo::EinsumOp)

Einsum operation

This operation is on its way out of StableHLO, so it is not included in the specification: https://github.com/openxla/stablehlo/issues/3

Informally, this operation does the same thing as TF's einsum: https://www.tensorflow.org/api_docs/python/tf/einsum

उदाहरण:

%result = "mhlo.einsum"(%lhs, %rhs) {
  einsum_config = "ab,bc->ac"
} : (tensor<4x16xf32>, tensor<16x4xf32>) -> tensor<4x4xf32>

लक्षण: AlwaysSpeculatableImplTrait

इंटरफेस: ConditionallySpeculatable , NoMemoryEffect (MemoryEffectOpInterface)

प्रभाव: MemoryEffects::Effect{}

गुण:

संकार्य:

परिणाम:

mhlo.erf (mhlo::ErfOp)

Erf operation

वाक्य - विन्यास:

operation ::= `mhlo.erf` $operand attr-dict
              `:` custom<SameOperandsAndResultType>(type($operand), type($result))

Performs element-wise erf operation on operand tensor and produces a result tensor.

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#erf

उदाहरण:

%result = mhlo.erf %operand : tensor<2x2xf32>

Traits: AlwaysSpeculatableImplTrait , CompatibleOperandsAndResultType , Elementwise , SameOperandsAndResultShape

Interfaces: ConditionallySpeculatable , InferShapedTypeOpInterface , InferTypeOpInterface , NoMemoryEffect (MemoryEffectOpInterface)

प्रभाव: MemoryEffects::Effect{}

संकार्य:

परिणाम:

mhlo.exponential (mhlo::ExpOp)

Exp operation

वाक्य - विन्यास:

operation ::= `mhlo.exponential` $operand attr-dict
              `:` custom<SameOperandsAndResultType>(type($operand), type($result))

Performs element-wise exponential operation on operand tensor and produces a result tensor.

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#exponential

उदाहरण:

%result = mhlo.exponential %operand : tensor<2x2xf64>

Traits: AlwaysSpeculatableImplTrait , CompatibleOperandsAndResultType , Elementwise , SameOperandsAndResultShape

Interfaces: ConditionallySpeculatable , InferShapedTypeOpInterface , InferTypeOpInterface , NoMemoryEffect (MemoryEffectOpInterface)

प्रभाव: MemoryEffects::Effect{}

संकार्य:

परिणाम:

mhlo.exponential_minus_one (mhlo::Expm1Op)

Expm1 operation

वाक्य - विन्यास:

operation ::= `mhlo.exponential_minus_one` $operand attr-dict
              `:` custom<SameOperandsAndResultType>(type($operand), type($result))

Performs element-wise exponential minus one operation on operand tensor and produces a result tensor.

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#exponential_minus_one

उदाहरण:

%result = mhlo.exponential_minus_one %operand : tensor<2xf32>

Traits: AlwaysSpeculatableImplTrait , CompatibleOperandsAndResultType , Elementwise , SameOperandsAndResultShape

Interfaces: ConditionallySpeculatable , InferShapedTypeOpInterface , InferTypeOpInterface , NoMemoryEffect (MemoryEffectOpInterface)

प्रभाव: MemoryEffects::Effect{}

संकार्य:

परिणाम:

mhlo.fft (mhlo::FftOp)

Fft operation

Performs the forward and inverse Fourier transforms for real and complex inputs/outputs.

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#fft

उदाहरण:

%result = mhlo.fft %operand, type = FFT, length = [4] : (tensor<4xcomplex<f32>>) -> tensor<4xcomplex<f32>>

Traits: AlwaysSpeculatableImplTrait , InferTensorType

Interfaces: ConditionallySpeculatable , InferShapedTypeOpInterface , InferTypeOpInterface , NoMemoryEffect (MemoryEffectOpInterface)

प्रभाव: MemoryEffects::Effect{}

गुण:

संकार्य:

परिणाम:

mhlo.floor (mhlo::FloorOp)

Floor operation

वाक्य - विन्यास:

operation ::= `mhlo.floor` $operand attr-dict
              `:` custom<SameOperandsAndResultType>(type($operand), type($result))

Performs element-wise floor of operand tensor and produces a result tensor.

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#floor

उदाहरण:

%result = mhlo.floor %operand : tensor<2xf32>

Traits: AlwaysSpeculatableImplTrait , CompatibleOperandsAndResultType , Elementwise , SameOperandsAndResultShape

Interfaces: ConditionallySpeculatable , InferShapedTypeOpInterface , InferTypeOpInterface , NoMemoryEffect (MemoryEffectOpInterface)

प्रभाव: MemoryEffects::Effect{}

संकार्य:

परिणाम:

mhlo.fusion (mhlo::FusionOp)

Fusion operation

This operation is private to the XLA compiler, so it is does not yet have a specification.

Informally, this operation consists of a group of basic ops (represented as a region attached to it). It serves as a hint to the backend that it is beneficial to emit the contained ops into a single loop nest or kernel.

गुण:

संकार्य:

परिणाम:

mhlo.gather (mhlo::GatherOp)

Gather operation

Gathers slices from operand tensor from offsets specified in start_indices and produces a result tensor.

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#gather

उदाहरण:

%result = "mhlo.gather"(%operand, %start_indices) {
  dimension_numbers = #mhlo.gather<
    offset_dims = [2, 3],
    collapsed_slice_dims = [0],
    start_index_map = [0, 2],
    index_vector_dim = 2>,
  slice_sizes = dense<[0, 2, 2]> : tensor<3xi64>,
  indices_are_sorted = false
} : (tensor<3x4x2xi32>, tensor<2x3x2xi64>) -> tensor<2x3x2x2xi32>

Traits: AlwaysSpeculatableImplTrait , InferTensorType

Interfaces: ConditionallySpeculatable , InferShapedTypeOpInterface , InferTypeOpInterface , NoMemoryEffect (MemoryEffectOpInterface)

प्रभाव: MemoryEffects::Effect{}

गुण:

संकार्य:

परिणाम:

mhlo.get_dimension_size (mhlo::GetDimensionSizeOp)

GetDimensionSize operation

Produces the size of the given dimension of the operand .

See https://github.com/openxla/stablehlo/blob/main/docs/spec.md#get_dimension_size

उदाहरण:

%result = mhlo.get_dimension_size %operand, dim = 1 : (tensor<2x3xf32>) -> tensor<i32>

Traits: AlwaysSpeculatableImplTrait , InferTensorType

Interfaces: ConditionallySpeculatable , InferShapedTypeOpInterface , InferTypeOpInterface , NoMemoryEffect (MemoryEffectOpInterface)

प्रभाव: MemoryEffects::Effect{}

गुण:

संकार्य:

परिणाम:

mhlo.get_tuple_element (mhlo::GetTupleElementOp)

GetTupleElement operation

वाक्य - विन्यास:

operation ::= `mhlo.get_tuple_element` $operand `[` $index `]` attr-dict `:` functional-type(operands, results)

Extracts element at index position of the operand tuple and produces a result .

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#get_tuple_element

उदाहरण:

%result = mhlo.get_tuple_element %operand[0] : (tuple<tensor<2xf32>, tuple<tensor<i32>>>) -> tensor<2xf32>

लक्षण: AlwaysSpeculatableImplTrait

इंटरफेस: ConditionallySpeculatable , InferTypeOpInterface , NoMemoryEffect (MemoryEffectOpInterface)

प्रभाव: MemoryEffects::Effect{}

गुण:

संकार्य:

परिणाम:

mhlo.if (mhlo::IfOp)

यदि ऑपरेशन

Produces the output from executing exactly one branch from true_branch or false_branch depending on the value of pred .

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#if

Example: %result = "mhlo.if"(%pred) ({ "mhlo.return"(%result_true_branch) : (tensor ) -> () }, { "mhlo.return"(%result_false_branch) : (tensor ) -> () }) : (tensor ) -> tensor

Traits: RecursiveMemoryEffects , SingleBlockImplicitTerminator<ReturnOp> , SingleBlock

इंटरफ़ेस: InferTypeOpInterface

संकार्य:

परिणाम:

mhlo.imag (mhlo::ImagOp)

Imag operation

वाक्य - विन्यास:

operation ::= `mhlo.imag` $operand attr-dict
              `:` custom<SameOperandsAndResultType>(type($operand), type($result))

Extracts the imaginary part, element-wise, from the operand and produces a result tensor.

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#imag

उदाहरण:

%result = mhlo.imag %operand : (tensor<2xcomplex<f32>>) -> tensor<2xf32>

Traits: AlwaysSpeculatableImplTrait , Elementwise , SameOperandsAndResultShape

Interfaces: ConditionallySpeculatable , InferShapedTypeOpInterface , InferTypeOpInterface , NoMemoryEffect (MemoryEffectOpInterface)

प्रभाव: MemoryEffects::Effect{}

संकार्य:

परिणाम:

mhlo.infeed (mhlo::InfeedOp)

Infeed operation

Reads data from the infeed and produces results .

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#infeed

उदाहरण:

%results:2 = "mhlo.infeed"(%token) {
  infeed_config = ""
} : (!mhlo.token) -> (tensor<3x3x3xi32>, !mhlo.token)

गुण:

संकार्य:

परिणाम:

mhlo.iota (mhlo::IotaOp)

Iota operation

Fills an output tensor with values in increasing order starting from zero along the iota_dimension dimension.

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#iota

उदाहरण:

%output = mhlo.iota dim = 0 : tensor<4x5xi32>

लक्षण: AlwaysSpeculatableImplTrait

इंटरफेस: ConditionallySpeculatable , NoMemoryEffect (MemoryEffectOpInterface)

प्रभाव: MemoryEffects::Effect{}

गुण:

परिणाम:

mhlo.is_finite (mhlo::IsFiniteOp)

IsFinite operation

वाक्य - विन्यास:

operation ::= `mhlo.is_finite` $x attr-dict `:` functional-type(operands, results)

Performs element-wise check whether the value in x is finite (ie is neither +Inf, -Inf, nor NaN) and produces a y tensor.

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#is_finite

उदाहरण:

%y = mhlo.is_finite %x : (tensor<7xf32>) -> tensor<7xi1>

Traits: AlwaysSpeculatableImplTrait , Elementwise , SameOperandsAndResultShape

Interfaces: ConditionallySpeculatable , InferShapedTypeOpInterface , InferTypeOpInterface , NoMemoryEffect (MemoryEffectOpInterface)

प्रभाव: MemoryEffects::Effect{}

संकार्य:

परिणाम:

mhlo.log (mhlo::LogOp)

Log operation

वाक्य - विन्यास:

operation ::= `mhlo.log` $operand attr-dict
              `:` custom<SameOperandsAndResultType>(type($operand), type($result))

Performs element-wise logarithm operation on operand tensor and produces a result tensor.

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#log

उदाहरण:

%result = mhlo.log %operand : tensor<2x2xf64>

Traits: AlwaysSpeculatableImplTrait , CompatibleOperandsAndResultType , Elementwise , SameOperandsAndResultShape

Interfaces: ConditionallySpeculatable , InferShapedTypeOpInterface , InferTypeOpInterface , NoMemoryEffect (MemoryEffectOpInterface)

प्रभाव: MemoryEffects::Effect{}

संकार्य:

परिणाम:

mhlo.log_plus_one (mhlo::Log1pOp)

Log1p operation

वाक्य - विन्यास:

operation ::= `mhlo.log_plus_one` $operand attr-dict
              `:` custom<SameOperandsAndResultType>(type($operand), type($result))

Performs element-wise logarithm plus one operation on operand tensor and produces a result tensor.

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#log_plus_one

उदाहरण:

%result = mhlo.log_plus_one %operand : tensor<6xf32>

Traits: AlwaysSpeculatableImplTrait , CompatibleOperandsAndResultType , Elementwise , SameOperandsAndResultShape

Interfaces: ConditionallySpeculatable , InferShapedTypeOpInterface , InferTypeOpInterface , NoMemoryEffect (MemoryEffectOpInterface)

प्रभाव: MemoryEffects::Effect{}

संकार्य:

परिणाम:

mhlo.logistic (mhlo::LogisticOp)

Logistic operation

वाक्य - विन्यास:

operation ::= `mhlo.logistic` $operand attr-dict
              `:` custom<SameOperandsAndResultType>(type($operand), type($result))

Performs element-wise logistic operation on operand tensor and produces a result tensor.

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#logistic

उदाहरण:

%result = mhlo.logistic %operand : tensor<2x2xf32>

Traits: AlwaysSpeculatableImplTrait , CompatibleOperandsAndResultType , Elementwise , SameOperandsAndResultShape

Interfaces: ConditionallySpeculatable , InferShapedTypeOpInterface , InferTypeOpInterface , NoMemoryEffect (MemoryEffectOpInterface)

प्रभाव: MemoryEffects::Effect{}

संकार्य:

परिणाम:

mhlo.map (mhlo::MapOp)

Map operation

Applies a map function computation to inputs along the dimensions and produces a result tensor.

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#map

उदाहरण:

%result = "mhlo.map"(%input0, %input1) ({
  ^bb0(%arg0: tensor<i32>, %arg1: tensor<i32>):
    %0 = mhlo.multiply %arg0, %arg1 : tensor<i32>
    mhlo.return %0 : tensor<i32>
}) {
  dimensions = dense<[0, 1]> : tensor<2xi64>
} : (tensor<2x2xi32>, tensor<2x2xi32>) -> tensor<2x2xi32>

Traits: InferTensorType , RecursiveMemoryEffects , SameOperandsAndResultShape , SingleBlockImplicitTerminator<ReturnOp> , SingleBlock

Interfaces: InferShapedTypeOpInterface , InferTypeOpInterface

गुण:

संकार्य:

परिणाम:

mhlo.maximum (mhlo::MaxOp)

मैक्स ऑपरेशन

वाक्य - विन्यास:

operation ::= `mhlo.maximum` $lhs `,` $rhs attr-dict
              `:` custom<SameOperandsAndResultType>(type($lhs), type($rhs), type($result))

Performs element-wise max operation on tensors lhs and rhs and produces a result tensor.

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#maximum

उदाहरण:

%result = mhlo.maximum %lhs, %rhs : tensor<4xf32>

Traits: AlwaysSpeculatableImplTrait , Commutative , CompatibleOperandsAndResultType , Elementwise , SameOperandsAndResultShape

Interfaces: ConditionallySpeculatable , InferShapedTypeOpInterface , InferTypeOpInterface , NoMemoryEffect (MemoryEffectOpInterface)

प्रभाव: MemoryEffects::Effect{}

संकार्य:

परिणाम:

mhlo.minimum (mhlo::MinOp)

न्यूनतम ऑपरेशन

वाक्य - विन्यास:

operation ::= `mhlo.minimum` $lhs `,` $rhs attr-dict
              `:` custom<SameOperandsAndResultType>(type($lhs), type($rhs), type($result))

Performs element-wise min operation on tensors lhs and rhs and produces a result tensor.

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#minimum

उदाहरण:

%result = mhlo.minimum %lhs, %rhs : tensor<4xf32>

Traits: AlwaysSpeculatableImplTrait , Commutative , CompatibleOperandsAndResultType , Elementwise , SameOperandsAndResultShape

Interfaces: ConditionallySpeculatable , InferShapedTypeOpInterface , InferTypeOpInterface , NoMemoryEffect (MemoryEffectOpInterface)

प्रभाव: MemoryEffects::Effect{}

संकार्य:

परिणाम:

mhlo.multiply (mhlo::MulOp)

Mul operation

वाक्य - विन्यास:

operation ::= `mhlo.multiply` $lhs `,` $rhs attr-dict
              `:` custom<SameOperandsAndResultType>(type($lhs), type($rhs), type($result))

Performs element-wise product of two tensors lhs and rhs and produces a result tensor.

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#multiply

उदाहरण:

%result = mhlo.multiply %lhs, %rhs : tensor<2xi32>

Traits: AlwaysSpeculatableImplTrait , Commutative , CompatibleOperandsAndResultType , Elementwise , SameOperandsAndResultShape

Interfaces: ConditionallySpeculatable , InferShapedTypeOpInterface , InferTypeOpInterface , NoMemoryEffect (MemoryEffectOpInterface)

प्रभाव: MemoryEffects::Effect{}

संकार्य:

परिणाम:

mhlo.negate (mhlo::NegOp)

Neg operation

वाक्य - विन्यास:

operation ::= `mhlo.negate` $operand attr-dict
              `:` custom<SameOperandsAndResultType>(type($operand), type($result))

Performs element-wise negation of operand tensor and produces a result tensor.

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#negate

उदाहरण:

%result = mhlo.negate %operand : tensor<2x3xi32>

Traits: AlwaysSpeculatableImplTrait , CompatibleOperandsAndResultType , Elementwise , SameOperandsAndResultShape

Interfaces: ConditionallySpeculatable , InferShapedTypeOpInterface , InferTypeOpInterface , NoMemoryEffect (MemoryEffectOpInterface)

प्रभाव: MemoryEffects::Effect{}

संकार्य:

परिणाम:

mhlo.not (mhlo::NotOp)

Not operation

वाक्य - विन्यास:

operation ::= `mhlo.not` $operand attr-dict
              `:` custom<SameOperandsAndResultType>(type($operand), type($result))

Performs element-wise NOT of tensor operand of type integer and produces a result tensor.

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#not

उदाहरण:

%result = mhlo.not %operand : tensor<5x3x1xi1>

Traits: AlwaysSpeculatableImplTrait , CompatibleOperandsAndResultType , Elementwise , SameOperandsAndResultShape

Interfaces: ConditionallySpeculatable , InferShapedTypeOpInterface , InferTypeOpInterface , NoMemoryEffect (MemoryEffectOpInterface)

प्रभाव: MemoryEffects::Effect{}

संकार्य:

परिणाम:

mhlo.optimization_barrier (mhlo::OptimizationBarrierOp)

OptimizationBarrier operation

वाक्य - विन्यास:

operation ::= `mhlo.optimization_barrier` attr-dict ($operand^ `:` custom<PairwiseOpType>(type($operand), type($result))):(`(` `)`)?

Ensures that the operations that produce the operand are executed before any operations that depend on the result and prevents compiler transformations from moving operations across the barrier. Other than that, the operation is an identity, ie result = operand .

See https://github.com/openxla/stablehlo/blob/main/docs/spec.md#optimization_barrier

उदाहरण:

%result0, %result1 = mhlo.optimization_barrier %operand0, %operand1 : tensor<f32>, tensor<f32>

Traits: AlwaysSpeculatableImplTrait , HLO_PairwiseSameOperandAndResultType

इंटरफेस: ConditionallySpeculatable , InferTypeOpInterface , NoMemoryEffect (MemoryEffectOpInterface)

प्रभाव: MemoryEffects::Effect{}

संकार्य:

परिणाम:

mhlo.or (mhlo::OrOp)

Or operation

वाक्य - विन्यास:

operation ::= `mhlo.or` $lhs `,` $rhs attr-dict
              `:` custom<SameOperandsAndResultType>(type($lhs), type($rhs), type($result))

Performs element-wise OR of two tensors lhs and rhs and produces a result tensor.

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#or

उदाहरण:

%result = mhlo.or %lhs, %rhs : tensor<2xi1>

Traits: AlwaysSpeculatableImplTrait , Commutative , CompatibleOperandsAndResultType , Elementwise , SameOperandsAndResultShape

Interfaces: ConditionallySpeculatable , InferShapedTypeOpInterface , InferTypeOpInterface , NoMemoryEffect (MemoryEffectOpInterface)

प्रभाव: MemoryEffects::Effect{}

संकार्य:

परिणाम:

mhlo.outfeed (mhlo::OutfeedOp)

Outfeed operation

Writes inputs to the outfeed and produces a result token.

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#outfeed

उदाहरण:

%result = "mhlo.outfeed"(%input0, %token) {
  outfeed_config = ""
} : (tensor<3x3x3xi32>, !mhlo.token) -> !mhlo.token

इंटरफ़ेस: InferTypeOpInterface

गुण:

संकार्य:

परिणाम:

mhlo.pad (mhlo::PadOp)

Pad operation

Expands operand by padding around the tensor as well as between the elements of the tensor with the given padding_value .

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#pad

उदाहरण:

%0 = mhlo.pad %arg0, %arg1, low = [0, 1], high = [2, 1], interior = [1, 2]
  : (tensor<2x3xi32>, tensor<i32>) -> tensor<5x9xi32>

Traits: AlwaysSpeculatableImplTrait , SameOperandsAndResultElementType

Interfaces: ConditionallySpeculatable , InferShapedTypeOpInterface , InferTypeOpInterface , NoMemoryEffect (MemoryEffectOpInterface)

प्रभाव: MemoryEffects::Effect{}

गुण:

संकार्य:

परिणाम:

mhlo.partition_id (mhlo::PartitionIdOp)

PartitionId operation

वाक्य - विन्यास:

operation ::= `mhlo.partition_id` attr-dict `:` type(results)

Produces partition_id of the current process.

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#partition_id

उदाहरण:

%result = mhlo.partition_id : tensor<ui32>

इंटरफ़ेस: InferTypeOpInterface

परिणाम:

mhlo.popcnt (mhlo::PopulationCountOp)

PopulationCount operation

वाक्य - विन्यास:

operation ::= `mhlo.popcnt` $operand attr-dict
              `:` custom<SameOperandsAndResultType>(type($operand), type($result))

Performs element-wise count of the number of bits set in the operand tensor and produces a result tensor.

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#popcnt

उदाहरण:

%result = mhlo.popcnt %operand : tensor<4xi8>

Traits: AlwaysSpeculatableImplTrait , CompatibleOperandsAndResultType , Elementwise , SameOperandsAndResultShape

Interfaces: ConditionallySpeculatable , InferShapedTypeOpInterface , InferTypeOpInterface , NoMemoryEffect (MemoryEffectOpInterface)

प्रभाव: MemoryEffects::Effect{}

संकार्य:

परिणाम:

mhlo.power (mhlo::PowOp)

Pow operation

वाक्य - विन्यास:

operation ::= `mhlo.power` $lhs `,` $rhs attr-dict
              `:` custom<SameOperandsAndResultType>(type($lhs), type($rhs), type($result))

Performs element-wise exponentiation of lhs tensor by rhs tensor and produces a result tensor.

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#power

उदाहरण:

%result = mhlo.power %lhs, %rhs : tensor<6xf32>

Traits: AlwaysSpeculatableImplTrait , CompatibleOperandsAndResultType , Elementwise , SameOperandsAndResultShape

Interfaces: ConditionallySpeculatable , InferShapedTypeOpInterface , InferTypeOpInterface , NoMemoryEffect (MemoryEffectOpInterface)

प्रभाव: MemoryEffects::Effect{}

संकार्य:

परिणाम:

mhlo.real (mhlo::RealOp)

Real operation

वाक्य - विन्यास:

operation ::= `mhlo.real` $operand attr-dict
              `:` custom<SameOperandsAndResultType>(type($operand), type($result))

Extracts the real part, element-wise, from the operand and produces a result tensor.

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#real

उदाहरण:

%result = mhlo.real %operand : (tensor<2xcomplex<f32>>) -> tensor<2xf32>

Traits: AlwaysSpeculatableImplTrait , Elementwise , SameOperandsAndResultShape

Interfaces: ConditionallySpeculatable , InferShapedTypeOpInterface , InferTypeOpInterface , NoMemoryEffect (MemoryEffectOpInterface)

प्रभाव: MemoryEffects::Effect{}

संकार्य:

परिणाम:

mhlo.real_dynamic_slice (mhlo::RealDynamicSliceOp)

RealDynamicSlice operation

वाक्य - विन्यास:

operation ::= `mhlo.real_dynamic_slice` operands attr-dict `:` functional-type(operands, results)

This operation is a work in progress, so it is not yet included in the specification: https://github.com/openxla/stablehlo/issues/8

Informally, this operation does the same thing as SliceOp except that start_indices , limit_indices and strides are specified dynamically: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#slice

उदाहरण:

%result = mhlo.real_dynamic_slice %operand,
            %start_indices, %limit_indices, %strides
       : (tensor<256x?xf32>, tensor<2xindex>, tensor<2xindex>, tensor<2xindex>) -> tensor<256x?xf32>

लक्षण: AlwaysSpeculatableImplTrait

Interfaces: ConditionallySpeculatable , InferShapedTypeOpInterface , NoMemoryEffect (MemoryEffectOpInterface)

प्रभाव: MemoryEffects::Effect{}

संकार्य:

परिणाम:

mhlo.recv (mhlo::RecvOp)

Recv operation

Receives data from a channel with channel_id and produces results .

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#recv

उदाहरण:

%results:2 = "mhlo.recv"(%token) {
  // channel_id = 5 : i64,
  // channel_type = #stablehlo<channel_type HOST_TO_DEVICE>,
  channel_handle = #mhlo.channel_handle<handle = 5, type = 3>,
  is_host_transfer = true
} : (!mhlo.token) -> (tensor<3x4xi32>, !mhlo.token)

गुण:

संकार्य:

परिणाम:

mhlo.reduce (mhlo::ReduceOp)

Reduce operation

Applies a reduction function body to inputs and init_values along the dimensions and produces a result tensor.

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#reduce

उदाहरण:

%result = "mhlo.reduce"(%input, %init_value) ({
  ^bb0(%arg0: tensor<i32>, %arg1: tensor<i32>):
    %0 = "mhlo.add"(%arg0, %arg1) : (tensor<i32>, tensor<i32>) -> tensor<i32>
    "mhlo.return"(%0) : (tensor<i32>) -> ()
}) {
  dimensions = dense<1> : tensor<1xi64>
} : (tensor<1x6xi32>, tensor<i32>) -> tensor<1xi32>

Traits: InferTensorType , RecursiveMemoryEffects , SameVariadicOperandSize , SingleBlockImplicitTerminator<ReturnOp> , SingleBlock

Interfaces: InferShapedTypeOpInterface , InferTypeOpInterface

गुण:

संकार्य:

परिणाम:

mhlo.reduce_precision (mhlo::ReducePrecisionOp)

ReducePrecision operation

वाक्य - विन्यास:

operation ::= `mhlo.reduce_precision` $operand `,` `format` `=` custom<ExponentMantissa>($exponent_bits, $mantissa_bits)
              attr-dict `:` custom<SameOperandsAndResultType>(type($operand), type($output))

Performs element-wise conversion of operand to another floating-point type that uses exponent_bits and mantissa_bits and back to the original floating-point type and produces an output tensor.

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#reduce_precision

उदाहरण:

%output = mhlo.reduce_precision %operand, format = e5m2 : tensor<6xf32>

Traits: AlwaysSpeculatableImplTrait , CompatibleOperandsAndResultType , Elementwise

Interfaces: ConditionallySpeculatable , InferShapedTypeOpInterface , InferTypeOpInterface , NoMemoryEffect (MemoryEffectOpInterface)

प्रभाव: MemoryEffects::Effect{}

गुण:

संकार्य:

परिणाम:

mhlo.reduce_scatter (mhlo::ReduceScatterOp)

रिड्यूसस्कैटर ऑपरेशन

Within each process group in the process grid, performs reduction, using computations , over the values of the operand tensor from each process, splits the reduction result along scatter_dimension into parts, and scatters the split parts between the processes to produce the result .

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#reduce_scatter

उदाहरण:

%result = "mhlo.reduce_scatter"(%operand) ({
  ^bb0(%arg0: tensor<f32>, %arg1: tensor<f32>):
  %0 = mhlo.add %arg0, %arg1 : tensor<f32>
  mhlo.return %0 : tensor<f32>
}) {
  scatter_dimension = 1 : i64,
  replica_groups = dense<[[0, 1]]> : tensor<1x2xi64>,
  // channel_id = 0
  channel_handle = #mhlo.channel_handle<handle = 0, type = 0>
  // use_global_device_ids = false
} : (tensor<2x4xf32>) -> tensor<2x2xf32>

गुण:

संकार्य:

परिणाम:

mhlo.reduce_window (mhlo::ReduceWindowOp)

ReduceWindow operation

Applies a reduction function body to windows of inputs and init_values and produces results .

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#reduce_window

उदाहरण:

%result = "mhlo.reduce_window"(%input, %init_value) ({
  ^bb0(%arg0: tensor<i32>, %arg1: tensor<i32>):
    %0 = mhlo.add %arg0, %arg1 : tensor<i32>
    mhlo.return %0 : tensor<i32>
}) {
  window_dimensions = dense<[2, 1]> : tensor<2xi64>,
  window_strides = dense<[4, 1]> : tensor<2xi64>,
  base_dilations = dense<[2, 1]> : tensor<2xi64>,
  window_dilations = dense<[3, 1]> : tensor<2xi64>,
  padding = dense<[[2, 1], [0, 0]]> : tensor<2x2xi64>
} : (tensor<3x2xi32>, tensor<i32>) -> tensor<2x2xi32>

Traits: InferTensorType , RecursiveMemoryEffects , SameVariadicOperandSize , SingleBlockImplicitTerminator<ReturnOp> , SingleBlock

Interfaces: InferShapedTypeOpInterface , InferTypeOpInterface

गुण:

संकार्य:

परिणाम:

mhlo.remainder (mhlo::RemOp)

Rem operation

वाक्य - विन्यास:

operation ::= `mhlo.remainder` $lhs `,` $rhs attr-dict
              `:` custom<SameOperandsAndResultType>(type($lhs), type($rhs), type($result))

Performs element-wise remainder of dividend lhs and divisor rhs tensors and produces a result tensor.

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#remainder

उदाहरण:

%result = mhlo.remainder %lhs, %rhs : tensor<4xi64>

Traits: AlwaysSpeculatableImplTrait , CompatibleOperandsAndResultType , Elementwise , SameOperandsAndResultShape

Interfaces: ConditionallySpeculatable , InferShapedTypeOpInterface , InferTypeOpInterface , NoMemoryEffect (MemoryEffectOpInterface)

प्रभाव: MemoryEffects::Effect{}

संकार्य:

परिणाम:

mhlo.replica_id (mhlo::ReplicaIdOp)

ReplicaId operation

वाक्य - विन्यास:

operation ::= `mhlo.replica_id` attr-dict `:` type(results)

Produces replica_id of the current process.

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#replica_id

उदाहरण:

%result = mhlo.replica_id : tensor<ui32>

लक्षण: AlwaysSpeculatableImplTrait

इंटरफेस: ConditionallySpeculatable , InferTypeOpInterface , NoMemoryEffect (MemoryEffectOpInterface)

प्रभाव: MemoryEffects::Effect{}

परिणाम:

mhlo.reshape (mhlo::ReshapeOp)

Reshape operation

वाक्य - विन्यास:

operation ::= `mhlo.reshape` operands attr-dict `:` functional-type(operands, results)

Performs reshape of operand tensor to a result tensor.

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#reshape

उदाहरण:

%result = mhlo.reshape %operand : (tensor<2xf32>) -> tensor<1x2xf32>

Traits: AlwaysSpeculatableImplTrait , HLO_CompatibleOperandsAndResultElementType

इंटरफेस: ConditionallySpeculatable , NoMemoryEffect (MemoryEffectOpInterface)

प्रभाव: MemoryEffects::Effect{}

संकार्य:

परिणाम:

mhlo.return (mhlo::ReturnOp)

_This operation is a work in progress, so it is not yet included in the specification: https://github.com/openxla/stablehlo/issues/425

Informally, this operation serves as a terminator for regions defined by
the StableHLO ops. Non-StableHLO ops, e.g. `func.func`, have their own
terminators, e.g. `func.return`.

Example:

    ```mlir
    %result = "mhlo.reduce"(%input, %init_value) ({
      ^bb0(%arg0: tensor<i32>, %arg1: tensor<i32>):
        %0 = "mhlo.add"(%arg0, %arg1) : (tensor<i32>, tensor<i32>) -> tensor<i32>
        "mhlo.return"(%0) : (tensor<i32>) -> ()
    }) {
      dimensions = dense<1> : tensor<1xi64>
    } : (tensor<1x6xi32>, tensor<i32>) -> tensor<1xi32>
    ```_


Syntax:

```

operation ::= mhlo.return $results attr-dict ( : type($results)^)?

Traits: `AlwaysSpeculatableImplTrait`, `Terminator`

Interfaces: `ConditionallySpeculatable`, `NoMemoryEffect (MemoryEffectOpInterface)`

Effects: `MemoryEffects::Effect{}`

#### Operands:

| Operand | Description |
| :-----: | ----------- |
| `results` | variadic of ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values or token or nested tuple with any combination of ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer values or ranked tensor of 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values or token values


### `mhlo.reverse` (mhlo::ReverseOp)

_Reverse operation_

Reverses the order of elements in the `operand` along the specified
`dimensions` and produces a `result` tensor.

See:
<a href="https://github.com/openxla/stablehlo/blob/main/docs/spec.md#reverse">https://github.com/openxla/stablehlo/blob/main/docs/spec.md#reverse</a>

Example:
```mlir
%result = mhlo.reverse %operand, dims = [1] : tensor<3x2xi32>

Traits: AlwaysSpeculatableImplTrait , CompatibleOperandsAndResultType

Interfaces: ConditionallySpeculatable , InferShapedTypeOpInterface , InferTypeOpInterface , NoMemoryEffect (MemoryEffectOpInterface)

प्रभाव: MemoryEffects::Effect{}

गुण:

संकार्य:

परिणाम:

mhlo.rng (mhlo::RngOp)

Rng operation

Generates random numbers using the rng_distribution algorithm and produces a result tensor of a given shape shape .

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#rng

उदाहरण:

%result = mhlo.rng %a, %b, %shape, distribution = NORMAL : (tensor<i32>, tensor<i32>, tensor<2xi64>) -> tensor<3x3xi32>

Traits: InferTensorType

Interfaces: InferShapedTypeOpInterface , InferTypeOpInterface

गुण:

संकार्य:

परिणाम:

mhlo.rng_bit_generator (mhlo::RngBitGeneratorOp)

RngBitGenerator operation

Returns an output filled with uniform random data and an updated output state output_state given an initial state initial_state using the pseudorandom number generator algorithm rng_algorithm .

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#rng_bit_generator

उदाहरण:

%output_state, %output = mhlo.rng_bit_generator %initial_state, algorithm = THREE_FRY : (tensor<2xui64>) -> (tensor<2xui64>, tensor<2x2xui64>)

लक्षण: AlwaysSpeculatableImplTrait

इंटरफेस: ConditionallySpeculatable , NoMemoryEffect (MemoryEffectOpInterface)

प्रभाव: MemoryEffects::Effect{}

गुण:

संकार्य:

परिणाम:

mhlo.round_nearest_afz (mhlo::RoundOp)

Round operation

वाक्य - विन्यास:

operation ::= `mhlo.round_nearest_afz` $operand attr-dict
              `:` custom<SameOperandsAndResultType>(type($operand), type($result))

Performs element-wise rounding towards the nearest integer, breaking ties away from zero, on the operand tensor and produces a result tensor.

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#round_nearest_afz

उदाहरण:

%result = mhlo.round_nearest_afz %operand : tensor<5xf32>

Traits: AlwaysSpeculatableImplTrait , CompatibleOperandsAndResultType , Elementwise , SameOperandsAndResultShape

Interfaces: ConditionallySpeculatable , InferShapedTypeOpInterface , InferTypeOpInterface , NoMemoryEffect (MemoryEffectOpInterface)

प्रभाव: MemoryEffects::Effect{}

संकार्य:

परिणाम:

mhlo.round_nearest_even (mhlo::RoundNearestEvenOp)

RoundNearestEven operation

वाक्य - विन्यास:

operation ::= `mhlo.round_nearest_even` $operand attr-dict
              `:` custom<SameOperandsAndResultType>(type($operand), type($result))

Performs element-wise rounding towards the nearest integer, breaking ties towards the even integer, on the operand tensor and produces a result tensor.

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#round_nearest_even

उदाहरण:

%result = mhlo.round_nearest_even %operand : tensor<5xf32>

Traits: AlwaysSpeculatableImplTrait , CompatibleOperandsAndResultType , Elementwise , SameOperandsAndResultShape

Interfaces: ConditionallySpeculatable , InferShapedTypeOpInterface , InferTypeOpInterface , NoMemoryEffect (MemoryEffectOpInterface)

प्रभाव: MemoryEffects::Effect{}

संकार्य:

परिणाम:

mhlo.rsqrt (mhlo::RsqrtOp)

Rsqrt operation

वाक्य - विन्यास:

operation ::= `mhlo.rsqrt` $operand attr-dict
              `:` custom<SameOperandsAndResultType>(type($operand), type($result))

Performs element-wise reciprocal square root operation on operand tensor and produces a result tensor, implementing the rSqrt operation from the IEEE-754 specification.

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#rsqrt

उदाहरण:

%result = mhlo.rsqrt %operand : tensor<2x2xf32>

Traits: AlwaysSpeculatableImplTrait , CompatibleOperandsAndResultType , Elementwise , SameOperandsAndResultShape

Interfaces: ConditionallySpeculatable , InferShapedTypeOpInterface , InferTypeOpInterface , NoMemoryEffect (MemoryEffectOpInterface)

प्रभाव: MemoryEffects::Effect{}

संकार्य:

परिणाम:

mhlo.scatter (mhlo::ScatterOp)

Scatter operation

Produces results tensors which are equal to inputs tensors except that several slices specified by scatter_indices are updated with the values updates using update_computation .

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#scatter

उदाहरण:

%result = "mhlo.scatter"(%input, %scatter_indices, %update) ({
  ^bb0(%arg0: tensor<i32>, %arg1: tensor<i32>):
    %0 = mhlo.add %arg0, %arg1 : tensor<i32>
    mhlo.return %0 : tensor<i32>
}) {
  scatter_dimension_numbers = #mhlo.scatter<
    update_window_dims = [2,3],
    inserted_window_dims = [0],
    scatter_dims_to_operand_dims = [1, 0],
    index_vector_dim = 2>,
  indices_are_sorted = false,
  unique_indices = false
} : (tensor<3x4x2xi32>, tensor<2x3x2xi64>, tensor<2x3x2x2xi32>) -> tensor<3x4x2xi32>

Traits: RecursiveMemoryEffects , SameVariadicOperandSize

इंटरफ़ेस: InferTypeOpInterface

गुण:

संकार्य:

परिणाम:

mhlo.select (mhlo::SelectOp)

ऑपरेशन का चयन करें

वाक्य - विन्यास:

operation ::= `mhlo.select` operands attr-dict `:`
              custom<SelectOpType>(type($pred), type($on_true), type($on_false), type($result))

Produces a result tensor where each element is selected from on_true or on_false tensor based on the value of the corresponding element of pred .

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#select

उदाहरण:

%result = mhlo.select %pred, %on_true, %on_false : tensor<2x2xi1>, tensor<2x2xi32>

Traits: AlwaysSpeculatableImplTrait , HLO_BroadcastingElementwise , InferTensorType

Interfaces: ConditionallySpeculatable , InferShapedTypeOpInterface , InferTypeOpInterface , NoMemoryEffect (MemoryEffectOpInterface)

प्रभाव: MemoryEffects::Effect{}

संकार्य:

परिणाम:

mhlo.select_and_scatter (mhlo::SelectAndScatterOp)

SelectAndScatter operation

Scatters the values from the source tensor using scatter based on the outcome of reduce_window of the input tensor using select and produces a result tensor.

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#select_and_scatter

उदाहरण:

%result = "mhlo.select_and_scatter"(%operand, %source, %init_value) ({
  ^bb0(%arg0: tensor<i32>, %arg1: tensor<i32>):
    %0 = "mhlo.compare"(%arg0, %arg1) {
      comparison_direction = #stablehlo<comparison_direction GE>
    } : (tensor<i32>, tensor<i32>) -> tensor<i1>
    "mhlo.return"(%0) : (tensor<i1>) -> ()
}, {
  ^bb0(%arg0: tensor<i32>, %arg1: tensor<i32>):
    %0 = "mhlo.add"(%arg0, %arg1) : (tensor<i32>, tensor<i32>) -> tensor<i32>
    "mhlo.return"(%0) : (tensor<i32>) -> ()
}) {
  window_dimensions = dense<[3, 1]> : tensor<2xi64>,
  window_strides = dense<[2, 1]> : tensor<2xi64>,
  padding = dense<[[0, 1], [0, 0]]> : tensor<2x2xi64>
} : (tensor<4x2xi32>, tensor<2x2xi32>, tensor<i32>) -> tensor<4x2xi32>

Traits: RecursiveMemoryEffects

इंटरफ़ेस: InferTypeOpInterface

गुण:

संकार्य:

परिणाम:

mhlo.send (mhlo::SendOp)

Send operation

Sends inputs to a channel channel_id and produces a result token.

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#send

उदाहरण:

%result = "mhlo.send"(%operand, %token) {
  // channel_id = 5 : i64,
  // channel_type = #stablehlo<channel_type DEVICE_TO_HOST>,
  channel_handle = #mhlo.channel_handle<handle = 5, type = 2>,
  is_host_transfer = true
} : (tensor<3x4xi32>, !mhlo.token) -> !mhlo.token

इंटरफ़ेस: InferTypeOpInterface

गुण:

संकार्य:

परिणाम:

mhlo.set_dimension_size (mhlo::SetDimensionSizeOp)

SetDimensionSize operation

This operation is a work in progress, so it is not yet included in the specification: https://github.com/openxla/stablehlo/issues/8

Informally, this operation does the same thing as XLA's SetDimensionSize: https://www.tensorflow.org/xla/operation_semantics#setdimensionsize

उदाहरण:

%0 = mhlo.set_dimension_size %arg0, %arg1, dim = 1 : (tensor<4x2xf32>, tensor<i32>) -> tensor<4x2xf32>

Traits: AlwaysSpeculatableImplTrait , InferTensorType

Interfaces: ConditionallySpeculatable , InferShapedTypeOpInterface , InferTypeOpInterface , NoMemoryEffect (MemoryEffectOpInterface)

प्रभाव: MemoryEffects::Effect{}

गुण:

संकार्य:

परिणाम:

mhlo.shift_left (mhlo::ShiftLeftOp)

ShiftLeft operation

वाक्य - विन्यास:

operation ::= `mhlo.shift_left` $lhs `,` $rhs attr-dict
              `:` custom<SameOperandsAndResultType>(type($lhs), type($rhs), type($result))

Performs element-wise left-shift operation on the lhs tensor by rhs number of bits and produces a result tensor.

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#shift_left

उदाहरण:

%result = mhlo.shift_left %lhs, %rhs : tensor<6xi8>

Traits: AlwaysSpeculatableImplTrait , CompatibleOperandsAndResultType , Elementwise , SameOperandsAndResultShape

Interfaces: ConditionallySpeculatable , InferShapedTypeOpInterface , InferTypeOpInterface , NoMemoryEffect (MemoryEffectOpInterface)

प्रभाव: MemoryEffects::Effect{}

संकार्य:

परिणाम:

mhlo.shift_right_arithmetic (mhlo::ShiftRightArithmeticOp)

ShiftRightArithmetic operation

वाक्य - विन्यास:

operation ::= `mhlo.shift_right_arithmetic` $lhs `,` $rhs attr-dict
              `:` custom<SameOperandsAndResultType>(type($lhs), type($rhs), type($result))

Performs element-wise arithmetic right-shift operation on the lhs tensor by rhs number of bits and produces a result tensor.

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#shift_right_arithmetic

उदाहरण:

%result = mhlo.shift_right_arithmetic %lhs, %rhs : tensor<6xi8>

Traits: AlwaysSpeculatableImplTrait , CompatibleOperandsAndResultType , Elementwise , SameOperandsAndResultShape

Interfaces: ConditionallySpeculatable , InferShapedTypeOpInterface , InferTypeOpInterface , NoMemoryEffect (MemoryEffectOpInterface)

प्रभाव: MemoryEffects::Effect{}

संकार्य:

परिणाम:

mhlo.shift_right_logical (mhlo::ShiftRightLogicalOp)

ShiftRightLogical operation

वाक्य - विन्यास:

operation ::= `mhlo.shift_right_logical` $lhs `,` $rhs attr-dict
              `:` custom<SameOperandsAndResultType>(type($lhs), type($rhs), type($result))

Performs element-wise logical right-shift operation on the lhs tensor by rhs number of bits and produces a result tensor.

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#shift_right_logical

उदाहरण:

%result = mhlo.shift_right_logical %lhs, %rhs : tensor<6xi8>

Traits: AlwaysSpeculatableImplTrait , CompatibleOperandsAndResultType , Elementwise , SameOperandsAndResultShape

Interfaces: ConditionallySpeculatable , InferShapedTypeOpInterface , InferTypeOpInterface , NoMemoryEffect (MemoryEffectOpInterface)

प्रभाव: MemoryEffects::Effect{}

संकार्य:

परिणाम:

mhlo.sign (mhlo::SignOp)

Sign operation

वाक्य - विन्यास:

operation ::= `mhlo.sign` $operand attr-dict
              `:` custom<SameOperandsAndResultType>(type($operand), type($result))

Returns the sign of the operand element-wise and produces a result tensor.

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#sign

उदाहरण:

%result = mhlo.sign %operand : tensor<7xf32>

Traits: AlwaysSpeculatableImplTrait , CompatibleOperandsAndResultType , Elementwise , SameOperandsAndResultShape

Interfaces: ConditionallySpeculatable , InferShapedTypeOpInterface , InferTypeOpInterface , NoMemoryEffect (MemoryEffectOpInterface)

प्रभाव: MemoryEffects::Effect{}

संकार्य:

परिणाम:

mhlo.sine (mhlo::SineOp)

Sine operation

वाक्य - विन्यास:

operation ::= `mhlo.sine` $operand attr-dict
              `:` custom<SameOperandsAndResultType>(type($operand), type($result))

Performs element-wise sine operation on operand tensor and produces a result tensor.

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#sine

उदाहरण:

%result = mhlo.sine %operand : tensor<2xf32>

Traits: AlwaysSpeculatableImplTrait , CompatibleOperandsAndResultType , Elementwise , SameOperandsAndResultShape

Interfaces: ConditionallySpeculatable , InferShapedTypeOpInterface , InferTypeOpInterface , NoMemoryEffect (MemoryEffectOpInterface)

प्रभाव: MemoryEffects::Effect{}

संकार्य:

परिणाम:

mhlo.slice (mhlo::SliceOp)

स्लाइस ऑपरेशन

Extracts a slice from the operand using statically-computed starting indices and produces a result tensor.

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#slice

उदाहरण:

%result = "mhlo.slice" (%operand) {
  start_indices = dense<[1, 2]> : tensor<2xi64>,
  limit_indices = dense<[3, 4]> : tensor<2xi64>,
  strides = dense<1> : tensor<2xi64>
} : (tensor<3x4xi64>) -> tensor<2x2xi64>

Traits: AlwaysSpeculatableImplTrait , SameOperandsAndResultElementType

इंटरफेस: ConditionallySpeculatable , InferTypeOpInterface , NoMemoryEffect (MemoryEffectOpInterface)

प्रभाव: MemoryEffects::Effect{}

गुण:

संकार्य:

परिणाम:

mhlo.sort (mhlo::SortOp)

Sort operation

Sorts a variadic number of tensors in inputs together, according to a custom comparator , along the given dimension and produces a variadic number of tensors as results .

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#sort

उदाहरण:

%result0, %result1 = "mhlo.sort"(%input0, %input1) ({
  ^bb0(%arg0: tensor<i32>, %arg1: tensor<i32>, %arg2: tensor<i32>, %arg3: tensor<i32>):
    %predicate = "mhlo.compare"(%arg0, %arg1) {
      comparison_direction = #stablehlo<comparison_direction GT>
      } : (tensor<i32>, tensor<i32>) -> tensor<i1>
    "mhlo.return"(%predicate) : (tensor<i1>) -> ()
}) {
  dimension = 0 : i64,
  is_stable = true
} : (tensor<2x3xi32>, tensor<2x3xi32>) -> (tensor<2x3xi32>, tensor<2x3xi32>)

Traits: InferTensorType , RecursiveMemoryEffects , SameOperandsAndResultShape

Interfaces: InferShapedTypeOpInterface , InferTypeOpInterface

गुण:

संकार्य:

परिणाम:

mhlo.sparse_dot (mhlo::SparseDotOp)

Sparse dot operation

Similar to dot_general operation, with one or both of the operands being sparse. An additional argument provides sparsity meta information. Disclaimer: this op is experimental / a work in progress.

लक्षण: AlwaysSpeculatableImplTrait

इंटरफेस: ConditionallySpeculatable , NoMemoryEffect (MemoryEffectOpInterface)

प्रभाव: MemoryEffects::Effect{}

गुण:

संकार्य:

परिणाम:

mhlo.sqrt (mhlo::SqrtOp)

Sqrt operation

वाक्य - विन्यास:

operation ::= `mhlo.sqrt` $operand attr-dict
              `:` custom<SameOperandsAndResultType>(type($operand), type($result))

Performs element-wise square root operation on operand tensor and produces a result tensor.

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#sqrt

उदाहरण:

%result = mhlo.sqrt %operand : tensor<2x2xf32>

Traits: AlwaysSpeculatableImplTrait , CompatibleOperandsAndResultType , Elementwise , SameOperandsAndResultShape

Interfaces: ConditionallySpeculatable , InferShapedTypeOpInterface , InferTypeOpInterface , NoMemoryEffect (MemoryEffectOpInterface)

प्रभाव: MemoryEffects::Effect{}

संकार्य:

परिणाम:

mhlo.stochastic_convert (mhlo::StochasticConvertOp)

StochasticConvert operation

This operation is a work in progress, so it is not yet included in the specification: https://github.com/openxla/stablehlo/issues/295

Informally, this operation performs element-wise conversion of values from a bigger type to a smaller one with stochastic rounding using the random number passed in.

Traits: AlwaysSpeculatableImplTrait , Elementwise

इंटरफेस: ConditionallySpeculatable , NoMemoryEffect (MemoryEffectOpInterface)

प्रभाव: MemoryEffects::Effect{}

संकार्य:

परिणाम:

mhlo.subtract (mhlo::SubtractOp)

घटाव संक्रिया

वाक्य - विन्यास:

operation ::= `mhlo.subtract` $lhs `,` $rhs attr-dict
              `:` custom<SameOperandsAndResultType>(type($lhs), type($rhs), type($result))

Performs element-wise subtraction of two tensors lhs and rhs and produces a result tensor.

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#subtract

उदाहरण:

%result = mhlo.subtract %lhs, %rhs : tensor<2xi32>

Traits: AlwaysSpeculatableImplTrait , CompatibleOperandsAndResultType , Elementwise , SameOperandsAndResultShape

Interfaces: ConditionallySpeculatable , InferShapedTypeOpInterface , InferTypeOpInterface , NoMemoryEffect (MemoryEffectOpInterface)

प्रभाव: MemoryEffects::Effect{}

संकार्य:

परिणाम:

mhlo.tan (mhlo::TanOp)

Tan operation

वाक्य - विन्यास:

operation ::= `mhlo.tan` $operand attr-dict
              `:` custom<SameOperandsAndResultType>(type($operand), type($result))

This operation is a work in progress, so it is not yet included in the specification: https://github.com/openxla/stablehlo/issues/954

Informally, this operation returns Tan(operand) element-wise.

उदाहरण:

%0 = mhlo.tan %arg0 : tensor<2xf32>

Traits: AlwaysSpeculatableImplTrait , CompatibleOperandsAndResultType , Elementwise , SameOperandsAndResultShape

Interfaces: ConditionallySpeculatable , InferShapedTypeOpInterface , InferTypeOpInterface , NoMemoryEffect (MemoryEffectOpInterface)

प्रभाव: MemoryEffects::Effect{}

संकार्य:

परिणाम:

mhlo.tanh (mhlo::TanhOp)

Tanh operation

वाक्य - विन्यास:

operation ::= `mhlo.tanh` $operand attr-dict
              `:` custom<SameOperandsAndResultType>(type($operand), type($result))

Performs element-wise hyperbolic tangent operation on operand tensor and produces a result tensor.

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#tanh

उदाहरण:

%result = mhlo.tanh %operand : tensor<2xf32>

Traits: AlwaysSpeculatableImplTrait , CompatibleOperandsAndResultType , Elementwise , SameOperandsAndResultShape

Interfaces: ConditionallySpeculatable , InferShapedTypeOpInterface , InferTypeOpInterface , NoMemoryEffect (MemoryEffectOpInterface)

प्रभाव: MemoryEffects::Effect{}

संकार्य:

परिणाम:

mhlo.topk (mhlo::TopKOp)

TopK operation

वाक्य - विन्यास:

operation ::= `mhlo.topk` `(`$operand `,` `k` `=` $k (`,` `largest` `=` $largest^)? `)` attr-dict `:`
              type($operand) `->` `(`type($values)`,` type($indices)`)`

Returns top k values and their indices, along the last dimension of the operand if largest=true or the bottom k values if largest=false .

See: https://www.tensorflow.org/xla/operation_semantics#top-k

उदाहरण:

%values, %indices = mhlo.topk(%operand, k=5, largest=true)
  : tensor<100xf32> -> (tensor<5xf32>, tensor<5xi32>)

Traits: InferTensorType , RecursiveMemoryEffects

Interfaces: InferShapedTypeOpInterface , InferTypeOpInterface

गुण:

संकार्य:

परिणाम:

mhlo.torch_index_select (mhlo::TorchIndexSelectOp)

TorchIndexSelect operation

This operation is on its way out of StableHLO, so it is not included in the specification: https://github.com/openxla/stablehlo/issues/3

Informally, this operation does the same thing as PyTorch's index_select, augmented with support for batch dimensions: https://pytorch.org/docs/stable/generated/torch.index_select.html

The batch_dims attribute specifies the number of major batch dimensions (0 or more) that act like a multidimensional loop over both the operand and the index.

उदाहरण:

%result = "mhlo.torch_index_select"(%operand, %index) {
  dim = 2 : i64,
  batch_dims = 1 : i64
} : (tensor<8x128x3072x64xf32>, tensor<8x16x1024xi32>) -> tensor<8x128x16x1024x64xf32>

लक्षण: AlwaysSpeculatableImplTrait

इंटरफेस: ConditionallySpeculatable , NoMemoryEffect (MemoryEffectOpInterface)

प्रभाव: MemoryEffects::Effect{}

गुण:

संकार्य:

परिणाम:

mhlo.trace (mhlo::TraceOp)

Trace operation

वाक्य - विन्यास:

operation ::= `mhlo.trace` $operand `,` $tag attr-dict `:` type($operand)

This operation is on its way out of StableHLO, so it is not included in the specification: https://github.com/openxla/stablehlo/issues/604

It is not used by JAX, PyTorch or TensorFlow, so it looks like we should've classified it as "Private to XLA" and not included it in StableHLO in the first place. With that in mind, its semantics will not be documented here.

उदाहरण:

mhlo.trace %arg0, "In test code." : tensor<5x1x5xi32>

गुण:

संकार्य:

mhlo.transpose (mhlo::TransposeOp)

ट्रांसपोज़ ऑपरेशन

Permutes the dimensions of operand tensor using permutation and produces a result tensor.

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#transpose

उदाहरण:

%0 = mhlo.transpose %arg0, dims = [2, 1, 0] : (tensor<1x2x3xi32>) -> tensor<3x2x1xi32>

Traits: AlwaysSpeculatableImplTrait , HLO_CompatibleOperandsAndResultElementType

Interfaces: ConditionallySpeculatable , InferShapedTypeOpInterface , InferTypeOpInterface , NoMemoryEffect (MemoryEffectOpInterface)

प्रभाव: MemoryEffects::Effect{}

गुण:

संकार्य:

परिणाम:

mhlo.triangular_solve (mhlo::TriangularSolveOp)

TriangularSolve operation

Solves batches of systems of linear equations with lower or upper triangular coefficient matrices.

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#triangular_solve

उदाहरण:

%result = "mhlo.triangular_solve"(%a, %b) {
  left_side = true,
  lower = true,
  unit_diagonal = false,
  transpose_a = #stablehlo<transpose NO_TRANSPOSE>
} : (tensor<3x3xf32>, tensor<3x3xf32>) -> tensor<3x3xf32>

Traits: AlwaysSpeculatableImplTrait , InferTensorType , SameOperandsAndResultElementType

Interfaces: ConditionallySpeculatable , InferShapedTypeOpInterface , InferTypeOpInterface , NoMemoryEffect (MemoryEffectOpInterface)

प्रभाव: MemoryEffects::Effect{}

गुण:

संकार्य:

परिणाम:

mhlo.tuple (mhlo::TupleOp)

टुपल ऑपरेशन

वाक्य - विन्यास:

operation ::= `mhlo.tuple` $val attr-dict `:` custom<TupleOpType>(type($val), type($result))

Produces a result tuple from values val .

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#tuple

उदाहरण:

%result = mhlo.tuple %val0, %val1 : tuple<tensor<2xf32>, tuple<tensor<i32>>>

लक्षण: AlwaysSpeculatableImplTrait

इंटरफेस: ConditionallySpeculatable , InferTypeOpInterface , NoMemoryEffect (MemoryEffectOpInterface)

प्रभाव: MemoryEffects::Effect{}

संकार्य:

परिणाम:

mhlo.unary_einsum (mhlo::UnaryEinsumOp)

UnaryEinsum operation

This operation is on its way out of StableHLO, so it is not included in the specification: https://github.com/openxla/stablehlo/issues/3

Informally, this operation does the same thing as TF's einsum: https://www.tensorflow.org/api_docs/python/tf/einsum

उदाहरण:

%result = "mhlo.unary_einsum"(%operand) {
  einsum_config = "ab->a"
} : (tensor<4x16xf32>) -> tensor<4xf32>

लक्षण: AlwaysSpeculatableImplTrait

इंटरफेस: ConditionallySpeculatable , NoMemoryEffect (MemoryEffectOpInterface)

प्रभाव: MemoryEffects::Effect{}

गुण:

संकार्य:

परिणाम:

mhlo.uniform_dequantize (mhlo::UniformDequantizeOp)

UniformDequantize operation

वाक्य - विन्यास:

operation ::= `mhlo.uniform_dequantize` $operand attr-dict
              `:` custom<SameOperandsAndResultType>(type($operand), type($result))

Performs element-wise conversion of quantized tensor operand to a floating-point tensor result according to the quantization parameters defined by the operand type.

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#uniform_dequantize

उदाहरण:

%result = mhlo.uniform_dequantize %operand : (tensor<16x16x!quant.uniform<i8:f32, 34.0:16>>) -> tensor<16x16xf32>

Traits: AlwaysSpeculatableImplTrait , Elementwise , InferTensorType , SameOperandsAndResultShape

Interfaces: ConditionallySpeculatable , InferShapedTypeOpInterface , InferTypeOpInterface , NoMemoryEffect (MemoryEffectOpInterface)

प्रभाव: MemoryEffects::Effect{}

संकार्य:

परिणाम:

mhlo.uniform_quantize (mhlo::UniformQuantizeOp)

UniformQuantize operation

वाक्य - विन्यास:

operation ::= `mhlo.uniform_quantize` $operand attr-dict
              `:` custom<SameOperandsAndResultType>(type($operand), type($result))

Performs element-wise conversion of floating-point tensor or quantized tensor operand to a quantized tensor result according to the quantization parameters defined by the result type.

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#uniform_quantize

उदाहरण:

%result = mhlo.uniform_quantize %operand : (tensor<16x16xf32>) -> tensor<16x16x!quant.uniform<ui8:f32, 34.0:16>>

Traits: AlwaysSpeculatableImplTrait , Elementwise , SameOperandsAndResultShape

Interfaces: ConditionallySpeculatable , InferShapedTypeOpInterface , NoMemoryEffect (MemoryEffectOpInterface)

प्रभाव: MemoryEffects::Effect{}

संकार्य:

परिणाम:

mhlo.while (mhlo::WhileOp)

While operation

Produces the output from executing body function 0 or more times while the cond function outputs true .

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#while

उदाहरण:

%results0, %results1 = "mhlo.while"(%operand0, %operand1) ({
  ^bb0(%arg0: tensor<i32>, %arg1: tensor<i32>):
    %0 = "mhlo.compare"(%arg0, %arg1) {
      comparison_direction = #stablehlo<comparison_direction LT>
    } : (tensor<i32>, tensor<i32>) -> tensor<i1>
    "mhlo.return"(%0) : (tensor<i1>) -> ()
}, {
  ^bb0(%arg0: tensor<i32>, %arg1: tensor<i32>):
    %0 = "mhlo.add"(%arg0, %constant0) : (tensor<i32>, tensor<i32>) -> tensor<i32>
    "mhlo.return"(%0, %arg1) : (tensor<i32>, tensor<i32>) -> ()
}) : (tensor<i32>, tensor<i32>) -> (tensor<i32>, tensor<i32>)

Traits: RecursiveMemoryEffects , SingleBlockImplicitTerminator<ReturnOp> , SingleBlock

Interfaces: InferTypeOpInterface , OpAsmOpInterface

संकार्य:

परिणाम:

mhlo.xla.rng_get_and_update_state (mhlo::XlaRngGetAndUpdateStateOp)

XlaRngGetAndUpdateState operation

वाक्य - विन्यास:

operation ::= `mhlo.xla.rng_get_and_update_state` attr-dict

This operation is private to the XLA compiler, so it is does not yet have a specification.

Informally, this operation represents the change of the global random number generator state for rng instructions. The global state is incremented by delta and the old state is returned.

The output is currently defined for a single output type. If this changes in the future to support multiple types, lowering to use of a global memref must ensure that a single memref is still used and updated appropriately.

इंटरफ़ेस: InferTypeOpInterface

गुण:

परिणाम:

mhlo.xor (mhlo::XorOp)

Xor operation

वाक्य - विन्यास:

operation ::= `mhlo.xor` $lhs `,` $rhs attr-dict
              `:` custom<SameOperandsAndResultType>(type($lhs), type($rhs), type($result))

Performs element-wise XOR of two tensors lhs and rhs and produces a result tensor.

See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#xor

उदाहरण:

%result = mhlo.xor %lhs, %rhs : tensor<2xi32>

Traits: AlwaysSpeculatableImplTrait , Commutative , CompatibleOperandsAndResultType , Elementwise , SameOperandsAndResultShape

Interfaces: ConditionallySpeculatable , InferShapedTypeOpInterface , InferTypeOpInterface , NoMemoryEffect (MemoryEffectOpInterface)

प्रभाव: MemoryEffects::Effect{}

संकार्य:

परिणाम:

गुण

ArgResultAliasAttr

Attribute that models the alias relationship of entry function argument

This attribute captures the alias relationship of an MHLO main function argument to one of the results, denoted by resultIndex . The argTupleIndices and resultTupleIndices are used to index into nested tuples in operand and result respectively. If isMustAlias is true then the operand-result pair must alias.

This is meant to be used as an attribute on a function argument in MHLO. For example, in the following code it expresses that %arg1 may alias 0-th result.

func @main(%arg0: tensor<2xf32>, %arg1: tensor<3xf32> {mhlo.result_alias =
    mhlo.result_alias<result_index = [2], ...>}
  ) -> tensor<2xf32>, tensor<3xf32> {
  // function body ...
}

पैरामीटर:

ChannelHandleAttr

two 64-bit integers 'handle' and 'type'

वाक्य - विन्यास:

#mhlo.channel_handle<
  int64_t,   # handle
  int64_t   # type
>

पैरामीटर:

ComparisonDirectionAttr

Which comparison operation to perform.

वाक्य - विन्यास:

#mhlo.comparison_direction<
  ::mlir::mhlo::ComparisonDirection   # value
>

एनम मामले:

• EQ ( EQ )
• NE ( NE )
• GE ( GE )
• GT ( GT )
• LE ( LE )
• LT ( LT ) #### Parameters:

ComparisonTypeAttr

Which comparison type to use.

वाक्य - विन्यास:

#mhlo.comparison_type<
  ::mlir::mhlo::ComparisonType   # value
>

एनम मामले:

• NOTYPE ( NOTYPE )
• FLOAT ( FLOAT )
• TOTALORDER ( TOTALORDER )
• SIGNED ( SIGNED )
• UNSIGNED ( UNSIGNED ) #### Parameters:

ConvDimensionNumbersAttr

Structure of dimension information for conv op

पैरामीटर:

CrossProgramPrefetchAttr

Argument that is prefetched from another program

वाक्य - विन्यास:

#mhlo.cross_program_prefetch<
  int64_t,   # parameter
  ::llvm::ArrayRef<int64_t>,   # indices
  std::optional<int64_t>   # offset
>

This attribute captures an argument that is prefetched from another program. For a given CrossProgramPrefetchAttr , parameter tells us which argument of the main function of the module is prefetched, and indices is a shape index telling us what subshape of that argument is prefetched.

A shape has a subshape iff it is a tuple. In that case, the subshape of the tuple by indices is the shape achieved after indexing by each element of indices in turn. For example, the [1,0] subshape of tuple<tuple<token, token>, tuple<tensor<i32>, token>> is tensor<i32> .

An empty value for indices means the whole shape is prefetched.

उदाहरण के लिए,

module attributes { mhlo.cross_program_prefetch = [ #mhlo.cross_program_prefetch< parameter = 0, indices = [0]> ]} {
  func.func @copy(%arg0 : tuple<tensor<2x3xi32>, tensor<i32>>) -> tuple<tensor<2x3xi32>, tensor<i32>> {
    %0 = "mhlo.copy"(%arg0) {is_cross_program_prefetch}
    return %0 : tuple<tensor<2x3xi32>, tensor<i32>>
  }
  func.func @main(%arg0 : tuple<tensor<2x3xi32>, tensor<i32>>) -> tuple<tensor<2x3xi32>, tensor<i32>> {
    %1 = "mhlo.async_start"(%arg0) {called_computation=@copy}
    %2 = "mhlo.async_done"(%1) {called_computation=@copy}
    return %2 : tuple<tensor<2x3xi32>, tensor<i32>>
  }
}

The parameter = 0 tells us that the async copy of the 0 th parameter is a cross_program_prefetch , while the index of [0] tells us that the 0 th element of the tuple is prefetched while the other element of the tuple is not.

पैरामीटर:

CustomCallScheduleAttr

Specifies the desired schedule for the custom-call.

वाक्य - विन्यास:

#mhlo.custom_call_schedule<
  ::mlir::mhlo::CustomCallSchedule   # value
>

एनम मामले:

• कोई नहीं ( NONE )
• LATEST ( LATEST )
• EARLIEST ( EARLIEST ) #### Parameters:

DequantizeModeAttr

Dequantization mode. Only MIN_COMBINED is supported.

वाक्य - विन्यास:

#mhlo.dequantize_mode<
  ::mlir::mhlo::DequantizeMode   # value
>

एनम मामले:

• MIN_COMBINED ( MIN_COMBINED ) #### Parameters:

DomainKindAttr

Kind of domain metatdata attached to an HLO domain.

वाक्य - विन्यास:

#mhlo.kind<
  ::mlir::mhlo::DomainKind   # value
>

एनम मामले:

• sharding ( sharding ) #### Parameters:

DotDimensionNumbersAttr

Attribute that models the dimension information for dot.

पैरामीटर:

FftTypeAttr

XLA fast fourier transform type.

वाक्य - विन्यास:

#mhlo.fft_type<
  ::mlir::mhlo::FftType   # value
>

एनम मामले:

• FFT ( FFT )
• IFFT ( IFFT )
• RFFT ( RFFT )
• IRFFT ( IRFFT ) #### Parameters:

FusionKindAttr

fusion kind

वाक्य - विन्यास:

#mhlo.fusion_kind<
  ::mlir::mhlo::FusionKind   # value
>

एनम मामले:

• kLoop ( kLoop )
• kInput ( kInput )
• kOutput ( kOutput )
• kCustom ( kCustom ) #### Parameters:

GatherDimensionNumbersAttr

Attribute that models the dimension information for gather

पैरामीटर:

OutputOperandAliasAttr

Attribute that models the alias relationship of output and operand of a CustomCall op

वाक्य - विन्यास:

#mhlo.output_operand_alias<
  ::llvm::ArrayRef<int64_t>,   # outputTupleIndices
  int64_t,   # operandIndex
  ::llvm::ArrayRef<int64_t>   # operandTupleIndices
>

This attribute captures the alias relationship of the output to one of the operands for a CustomCall op, denoted by operand_index . The output_tuple_indices and operand_tuple_indices are used to index into output and operand types. These indices lists are empty if the corresponding types are not tuple types, and can be arbitrarily long in case of arbitrarily nested tuple types.

See https://www.tensorflow.org/xla/aliasing

Example when used as array with in mhlo.custom-call:

%0 = "mhlo.custom_call"(%arg0, %arg1) {
  // other attributes
  output_operand_alias = [
    #mhlo.output_operand_alias<output_tuple_indices = [0],
                               operand_index = 0,
                               operand_tuple_indices = [1]>
  ]
} : (tuple<tensor<1x1xf32>, tensor<2x3xf32>>, tensor<5x5xf32>) -> tuple<tensor<2x3xf32>>

The output and the 0th operand are both tuples. The aliasing shows the
relationship between the 0th element in output tuple with the 1st element in
the 0th operand. And both of them are of the same type: tensor<2x3xf32>.

पैरामीटर:

PrecisionAttr

XLA precision for an operand. Has backend specific meaning.

वाक्य - विन्यास:

#mhlo.precision<
  ::mlir::mhlo::Precision   # value
>

एनम मामले:

• DEFAULT ( DEFAULT )
• HIGH )
• HIGHEST ( HIGHEST )
• PACKED_NIBBLE ( PACKED_NIBBLE ) #### Parameters:

RngAlgorithmAttr

XLA PRNG algorithm to be used.

वाक्य - विन्यास:

#mhlo.rng_algorithm<
  ::mlir::mhlo::RngAlgorithm   # value
>

एनम मामले:

• DEFAULT ( DEFAULT )
• THREE_FRY ( THREE_FRY )
• PHILOX ( PHILOX ) #### Parameters:

RngDistributionAttr

XLA PRNG distribution to be used.

वाक्य - विन्यास:

#mhlo.rng_distribution<
  ::mlir::mhlo::RngDistribution   # value
>

एनम मामले:

• UNIFORM ( UNIFORM )
• NORMAL ( NORMAL ) #### Parameters:

ScatterDimensionNumbersAttr

Attribute that models the dimension information for scatter

पैरामीटर:

SparsityDescriptorAttr

Describes structured (N:M) sparsity configuration

वाक्य - विन्यास:

#mhlo.sparsity<
  int64_t,   # dimension
  int64_t,   # n
  int64_t   # m
>

This attribute is defined for a sparse dot operation with a structured sparse input tensor. With (N=2,M=4), every 4 consecutive logical elements have exactly 2 non-zero physical elements in the input tensor.

$dimension defines the index of the contracting dimension that is sparse (it has to be the most minor dimension). The additional metadata operand in the sparse dot operation defines which logical elements are zeroed out.

पैरामीटर:

TransposeAttr

Transpose options

वाक्य - विन्यास:

#mhlo.transpose<
  ::mlir::mhlo::Transpose   # value
>

एनम मामले:

• TRANSPOSE_INVALID ( TRANSPOSE_INVALID )
• NO_TRANSPOSE ( NO_TRANSPOSE )
• ट्रांसपोज़ ( TRANSPOSE )
• ADJOINT ( ADJOINT ) #### Parameters:

TypeExtensionsAttr

Attribute that extends tensor type with MHLO type properties.

वाक्य - विन्यास:

#mhlo.type_extensions<
  ::llvm::ArrayRef<int64_t>   # bounds
>

This attribute is used to extend MLIR tensor type with MHLO tensor specific properties. These properties aren't modeled in the MLIR type. This attribute is set in the encoding field of the tensor type.

See HLO_BoundedAttrInterface for documentation for bounds .

पैरामीटर:

प्रकार

AsyncBundleType

Opaque collection of other types

वाक्य - विन्यास:

!mhlo.async_bundle<
  ::llvm::ArrayRef<Type>   # types
>

पैरामीटर: