Add aten.std.correction op and its decomposition (#1731)

e2e_testing/xfail_sets.py

@@ -628,7 +628,6 @@ LTC_XFAIL_SET = {
     "AdaptiveAvgPool2dNonUnitOutputSizeStaticModule_basic",
     "AddIntModule_basic",
     "BernoulliFloatModule_basic",
-    "BernoulliModule_basic",
     "BernoulliTensorModule_basic",
     "BincountMinlengthModule_basic",
     "BincountModule_basic",
@@ -639,7 +638,6 @@ LTC_XFAIL_SET = {
     "BoolIntTrueModule_basic",
     "CeilFloatModule_basic",
     "DivFloatModule_basic",
-    "DropoutTrainModule_basic",
     "ElementwiseAtenFloorDivideBroadcastModule_basic",
     "ElementwiseAtenFloorDivideModule_basic",
     "EqIntModule_basic",
@@ -712,13 +710,6 @@ LTC_XFAIL_SET = {
     "SliceOutOfUpperBoundIndexModule_basic",
     "SliceStartEqEndModule_basic",
     "SqrtIntModule_basic",
-    "StdBiasedModule_basic",
-    "StdDimBiasedModule_basic",
-    "StdDimKeepDimFalseModule_basic",
-    "StdDimKeepDimTrueModule_basic",
-    "StdDimEmptyDimModule_basic",
-    "StdDimNoneDimModule_basic",
-    "StdUnbiasedModule_basic",
     "SubFloatModule_basic",
     "SubIntModule_basic",
     "TensorsConcatNegativeDimModule_basic",

include/torch-mlir/Dialect/Torch/IR/GeneratedTorchOps.td

@@ -4905,6 +4905,32 @@ def Torch_AtenStdDimOp : Torch_Op<"aten.std.dim", [
   }];
 }
 
+def Torch_AtenStdCorrectionOp : Torch_Op<"aten.std.correction", [
+    AllowsTypeRefinement,
+    HasValueSemantics,
+    ReadOnly
+  ]> {
+  let summary = "Generated op for `aten::std.correction : (Tensor, int[]?, int?, bool) -> (Tensor)`";
+  let arguments = (ins
+    AnyTorchTensorType:$self,
+    AnyTorchOptionalListOfTorchIntType:$dim,
+    AnyTorchOptionalIntType:$correction,
+    Torch_BoolType:$keepdim
+  );
+  let results = (outs
+    AnyTorchTensorType:$result
+  );
+  let hasCustomAssemblyFormat = 1;
+  let extraClassDefinition = [{
+    ParseResult AtenStdCorrectionOp::parse(OpAsmParser &parser, OperationState &result) {
+      return parseDefaultTorchOp(parser, result, 4, 1);
+    }
+    void AtenStdCorrectionOp::print(OpAsmPrinter &printer) {
+      printDefaultTorchOp(printer, *this, 4, 1);
+    }
+  }];
+}
+
 def Torch_AtenVarOp : Torch_Op<"aten.var", [
     AllowsTypeRefinement,
     HasValueSemantics,

lib/Dialect/Torch/Transforms/AbstractInterpLibrary.cpp

@@ -5839,6 +5839,12 @@ StringRef mlir::torch::Torch::getAbstractInterpLibrary() {
 "    %1 = call @__torch__.torch.jit._shape_functions.sum_mean_dim(%arg0, %arg1, %arg3, %0) : (!torch.list<int>, !torch.optional<list<int>>, !torch.bool, !torch.any) -> !torch.list<int>\n"
 "    return %1 : !torch.list<int>\n"
 "  }\n"
+"  func.func @\"__torch_mlir_shape_fn.aten.std.correction\"(%arg0: !torch.list<int>, %arg1: !torch.optional<list<int>>, %arg2: !torch.optional<int>, %arg3: !torch.bool) -> !torch.list<int> {\n"
+"    %none = torch.constant.none\n"
+"    %0 = torch.derefine %none : !torch.none to !torch.any\n"
+"    %1 = call @__torch__.torch.jit._shape_functions.sum_mean_dim(%arg0, %arg1, %arg3, %0) : (!torch.list<int>, !torch.optional<list<int>>, !torch.bool, !torch.any) -> !torch.list<int>\n"
+"    return %1 : !torch.list<int>\n"
+"  }\n"
 "  func.func @\"__torch_mlir_shape_fn.aten.argmax\"(%arg0: !torch.list<int>, %arg1: !torch.optional<int>, %arg2: !torch.bool) -> !torch.list<int> {\n"
 "    %none = torch.constant.none\n"
 "    %0 = torch.aten.__is__ %arg1, %none : !torch.optional<int>, !torch.none -> !torch.bool\n"

lib/Dialect/Torch/Transforms/DecomposeComplexOps.cpp

@@ -1710,6 +1710,32 @@ public:
 };
 } // namespace
 
+// Decompose aten.std.correction to sqrt(var.correction(x))
+namespace {
+class DecomposeAtenStdCorrectionOp
+    : public OpRewritePattern<AtenStdCorrectionOp> {
+public:
+  using OpRewritePattern::OpRewritePattern;
+  LogicalResult matchAndRewrite(AtenStdCorrectionOp op,
+                                PatternRewriter &rewriter) const override {
+    Value self = op.getSelf();
+    BaseTensorType inputTensorType = self.getType().cast<BaseTensorType>();
+    if (!inputTensorType.hasDtype() ||
+        !inputTensorType.getDtype().isa<mlir::FloatType>()) {
+      return rewriter.notifyMatchFailure(
+          op,
+          "aten.std.correction expects input tensor of floating-point type");
+    }
+
+    Value varCorrection = rewriter.create<AtenVarCorrectionOp>(
+        op->getLoc(), op.getType(), self, op.getDim(), op.getCorrection(),
+        op.getKeepdim());
+    rewriter.replaceOpWithNewOp<AtenSqrtOp>(op, op.getType(), varCorrection);
+    return success();
+  }
+};
+} // namespace
+
 // Hardsigmoid(x) = max(0, min(1, (x+3)/6))
 namespace {
 class DecomposeAtenHardsigmoidOp : public OpRewritePattern<AtenHardsigmoidOp> {
@@ -3511,6 +3537,7 @@ public:
     addPatternIfTargetOpIsIllegal<DecomposeAtenAmaxOp>(patterns);
     addPatternIfTargetOpIsIllegal<DecomposeAtenVarCorrectionOp>(patterns);
     addPatternIfTargetOpIsIllegal<DecomposeAtenStdDimOp>(patterns);
+    addPatternIfTargetOpIsIllegal<DecomposeAtenStdCorrectionOp>(patterns);
     addPatternIfTargetOpIsIllegal<DecomposeAtenNarrowOp>(patterns);
     addPatternIfTargetOpIsIllegal<DecomposeAten_EmbeddingBagOp>(patterns);
     addPatternIfTargetOpIsIllegal<DecomposeAtenLiftFreshCopyOp>(patterns);

lib/Dialect/Torch/Transforms/LowerToBackendContract.cpp

@@ -429,6 +429,7 @@ static void markDecomposedOpsAsIllegal(MLIRContext *context,
   target.addIllegalOp<AtenAmaxOp>();
   target.addIllegalOp<AtenVarCorrectionOp>();
   target.addIllegalOp<AtenStdDimOp>();
+  target.addIllegalOp<AtenStdCorrectionOp>();
   target.addIllegalOp<AtenNarrowOp>();
   target.addIllegalOp<Aten_EmbeddingBagOp>();
   target.addIllegalOp<AtenLiftFreshCopyOp>();

lib/Dialect/Torch/Transforms/RefineTypes.cpp

@@ -976,8 +976,8 @@ void TypeAnalysis::visitOperation(Operation *op,
     Type dtype = operands[0]->getValue().dtype;
     visitReductionAlongAllDimsOp(op, dtype, operands);
     return;
-  } else if (isa<AtenStdOp, AtenStdDimOp, AtenVarOp, AtenVarDimOp,
+  } else if (isa<AtenStdOp, AtenStdDimOp, AtenStdCorrectionOp, AtenVarOp,
-                 AtenVarCorrectionOp>(op)) {
+                 AtenVarDimOp, AtenVarCorrectionOp>(op)) {
     auto input = operands[0]->getValue();
     visitReductionAlongAllDimsOp(op, input.dtype, operands);
     return;

python/torch_mlir/dialects/torch/importer/jit_ir/build_tools/abstract_interp_lib_gen.py

@@ -320,6 +320,9 @@ def aten〇std〡shape(self: List[int], unbiased: bool = True) -> List[int]:
 def aten〇std〇dim〡shape(self: List[int], dim: Optional[List[int]], unbiased: bool = True, keepdim: bool = False) -> List[int]:
     return upstream_shape_functions.sum_mean_dim(self, dim, keepdim, None)
 
+def aten〇std〇correction〡shape(self: List[int], dim: Optional[List[int]] = None, correction: Optional[int] = None, keepdim: bool = False) -> List[int]:
+    return upstream_shape_functions.sum_mean_dim(self, dim, keepdim, None)
+
 def _reduce_along_dim(self: List[int], dim: int, keepdim: bool):
     dim = upstream_shape_functions.maybe_wrap_dim(dim, len(self))
     out: List[int] = []

python/torch_mlir/dialects/torch/importer/jit_ir/build_tools/torch_ods_gen.py

@@ -405,6 +405,7 @@ def emit_ops(emitter_td: TextEmitter, registry: Registry):
     emit("aten::mean : (Tensor, int?) -> (Tensor)")
     emit("aten::std : (Tensor, bool) -> (Tensor)")
     emit("aten::std.dim : (Tensor, int[]?, bool, bool) -> (Tensor)")
+    emit("aten::std.correction : (Tensor, int[]?, int?, bool) -> (Tensor)")
     emit("aten::var : (Tensor, bool) -> (Tensor)")
     emit("aten::var.dim : (Tensor, int[]?, bool, bool) -> (Tensor)")
     emit("aten::var.correction : (Tensor, int[]?, int?, bool) -> (Tensor)")

python/torch_mlir_e2e_test/test_suite/stats.py

@@ -405,6 +405,163 @@ def StdDimNoneDimModule_basic(module, tu: TestUtils):
 # ==============================================================================
 
 
+class StdCorrectionModule(torch.nn.Module):
+
+    def __init__(self):
+        super().__init__()
+
+    @export
+    @annotate_args([
+        None,
+        ([-1, -1, -1], torch.float32, True),
+    ])
+    def forward(self, x):
+        return torch.ops.aten.std(x, dim=None, correction=2)
+
+
+@register_test_case(module_factory=lambda: StdCorrectionModule())
+def StdCorrectionModule_basic(module, tu: TestUtils):
+    module.forward(tu.rand(3, 4, 7))
+
+
+# ==============================================================================
+
+
+class StdCorrectionSingleDimReduceModule(torch.nn.Module):
+
+    def __init__(self):
+        super().__init__()
+
+    @export
+    @annotate_args([
+        None,
+        ([-1, -1, -1], torch.float32, True),
+    ])
+    def forward(self, x):
+        return torch.ops.aten.std(x, dim=[1], correction=1)
+
+
+@register_test_case(module_factory=lambda: StdCorrectionSingleDimReduceModule())
+def StdCorrectionSingleDimReduceModule_basic(module, tu: TestUtils):
+    module.forward(tu.rand(3, 4, 7))
+
+
+# ==============================================================================
+
+
+class StdCorrectionAllDimReduceModule(torch.nn.Module):
+
+    def __init__(self):
+        super().__init__()
+
+    @export
+    @annotate_args([
+        None,
+        ([-1, -1, -1], torch.float32, True),
+    ])
+    def forward(self, x):
+        return torch.ops.aten.std(x,
+                                  dim=[0, 1, 2],
+                                  correction=10,
+                                  keepdim=False)
+
+
+@register_test_case(module_factory=lambda: StdCorrectionAllDimReduceModule())
+def StdCorrectionAllDimReduceModule_basic(module, tu: TestUtils):
+    module.forward(tu.rand(3, 4, 7))
+
+
+# ==============================================================================
+
+
+class StdCorrectionKeepDimModule(torch.nn.Module):
+
+    def __init__(self):
+        super().__init__()
+
+    @export
+    @annotate_args([
+        None,
+        ([-1, -1, -1], torch.float32, True),
+    ])
+    def forward(self, x):
+        return torch.ops.aten.std(x, dim=[0, 1], correction=None, keepdim=True)
+
+
+@register_test_case(module_factory=lambda: StdCorrectionKeepDimModule())
+def StdCorrectionKeepDimModule_basic(module, tu: TestUtils):
+    module.forward(tu.rand(3, 4, 7))
+
+
+# ==============================================================================
+
+
+class StdCorrectionNoneModule(torch.nn.Module):
+
+    def __init__(self):
+        super().__init__()
+
+    @export
+    @annotate_args([
+        None,
+        ([-1, -1, -1], torch.float32, True),
+    ])
+    def forward(self, x):
+        return torch.ops.aten.std(x, dim=None, correction=None)
+
+
+@register_test_case(module_factory=lambda: StdCorrectionNoneModule())
+def StdCorrectionNoneModule_basic(module, tu: TestUtils):
+    module.forward(tu.rand(3, 4, 7))
+
+
+# ==============================================================================
+
+
+class StdCorrectionEmptyDimModule(torch.nn.Module):
+
+    def __init__(self):
+        super().__init__()
+
+    @export
+    @annotate_args([
+        None,
+        ([-1, -1, -1], torch.float32, True),
+    ])
+    def forward(self, x):
+        return torch.ops.aten.std(x, dim=[], correction=2)
+
+
+@register_test_case(module_factory=lambda: StdCorrectionEmptyDimModule())
+def StdCorrectionEmptyDimModule_basic(module, tu: TestUtils):
+    module.forward(tu.rand(3, 4, 7))
+
+
+# ==============================================================================
+
+
+class StdCorrectionLargeInputModule(torch.nn.Module):
+
+    def __init__(self):
+        super().__init__()
+
+    @export
+    @annotate_args([
+        None,
+        ([-1, -1, -1, -1], torch.float32, True),
+    ])
+    def forward(self, x):
+        return torch.ops.aten.std(x, dim=[2, 3], correction=2)
+
+
+@register_test_case(module_factory=lambda: StdCorrectionLargeInputModule())
+def StdCorrectionLargeInputModule_basic(module, tu: TestUtils):
+    module.forward(tu.rand(3, 4, 1024, 8192, low=100.0, high=101.0))
+
+
+# ==============================================================================
+
+
 class VarDimModule(torch.nn.Module):
 
     def __init__(self):
@@ -754,7 +911,7 @@ class VarCorrectionLargeInputModule(torch.nn.Module):
 
 @register_test_case(module_factory=lambda: VarCorrectionLargeInputModule())
 def VarCorrectionLargeInputModule_basic(module, tu: TestUtils):
-    module.forward(100 + tu.rand(3, 4, 1024, 8192))
+    module.forward(tu.rand(3, 4, 1024, 8192, low=100.0, high=101.0))
 
 
 # ==============================================================================