Add huber loss function & decompose to primitive ops

Signed-off-by: Zahid Wakeel <zahid.wakeel@multicorewareinc.com>

Author: zahidwx

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

@@ -9532,6 +9532,32 @@ def Torch_AtenKlDivOp : Torch_Op<"aten.kl_div", [
   }];
 }
 
+def Torch_AtenHuberLossOp : Torch_Op<"aten.huber_loss", [
+    AllowsTypeRefinement,
+    HasValueSemantics,
+    ReadOnly
+  ]> {
+  let summary = "Generated op for `aten::huber_loss : (Tensor, Tensor, int, float) -> (Tensor)`";
+  let arguments = (ins
+    AnyTorchTensorType:$self,
+    AnyTorchTensorType:$target,
+    Torch_IntType:$reduction,
+    Torch_FloatType:$delta
+  );
+  let results = (outs
+    AnyTorchOptionalTensorType:$result
+  );
+  let hasCustomAssemblyFormat = 1;
+  let extraClassDefinition = [{
+    ParseResult AtenHuberLossOp::parse(OpAsmParser &parser, OperationState &result) {
+      return parseDefaultTorchOp(parser, result, 4, 1);
+    }
+    void AtenHuberLossOp::print(OpAsmPrinter &printer) {
+      printDefaultTorchOp(printer, *this, 4, 1);
+    }
+  }];
+}
+
 def Torch_AtenBincountOp : Torch_Op<"aten.bincount", [
     AllowsTypeRefinement,
     HasValueSemantics,

lib/Dialect/Torch/Transforms/AbstractInterpLibrary.cpp

@@ -10717,6 +10717,31 @@ StringRef mlir::torch::Torch::getAbstractInterpLibrary() {
 "    }\n"
 "    return %2 : !torch.list<int>\n"
 "  }\n"
+"  func.func @\"__torch_mlir_shape_fn.aten.huber_loss\"(%arg0: !torch.list<int>, %arg1: !torch.list<int>, %arg2: !torch.int, %arg3: !torch.float) -> !torch.list<int> {\n"
+"    %none = torch.constant.none\n"
+"    %str = torch.constant.str \"AssertionError: Invalid reduction value.\"\n"
+"    %int0 = torch.constant.int 0\n"
+"    %int1 = torch.constant.int 1\n"
+"    %int2 = torch.constant.int 2\n"
+"    %0 = torch.prim.Uninitialized : !torch.list<int>\n"
+"    %1 = torch.aten.eq.int %arg2, %int0 : !torch.int, !torch.int -> !torch.bool\n"
+"    %2 = torch.prim.If %1 -> (!torch.list<int>) {\n"
+"      %3 = func.call @__torch__.torch.jit._shape_functions.unary(%arg0) : (!torch.list<int>) -> !torch.list<int>\n"
+"      torch.prim.If.yield %3 : !torch.list<int>\n"
+"    } else {\n"
+"      %3 = torch.prim.ListConstruct %int1, %int2 : (!torch.int, !torch.int) -> !torch.list<int>\n"
+"      %4 = torch.aten.__contains__.int_list %3, %arg2 : !torch.list<int>, !torch.int -> !torch.bool\n"
+"      %5 = torch.prim.If %4 -> (!torch.list<int>) {\n"
+"        %6 = torch.prim.ListConstruct  : () -> !torch.list<int>\n"
+"        torch.prim.If.yield %6 : !torch.list<int>\n"
+"      } else {\n"
+"        torch.prim.RaiseException %str, %none : !torch.str, !torch.none\n"
+"        torch.prim.If.yield %0 : !torch.list<int>\n"
+"      }\n"
+"      torch.prim.If.yield %5 : !torch.list<int>\n"
+"    }\n"
+"    return %2 : !torch.list<int>\n"
+"  }\n"
 "  func.func @\"__torch_mlir_shape_fn.aten.nll_loss_forward\"(%arg0: !torch.list<int>, %arg1: !torch.list<int>, %arg2: !torch.optional<list<int>>, %arg3: !torch.int, %arg4: !torch.int) -> !torch.tuple<list<int>, list<int>> {\n"
 "    %0 = call @__torch__.torch.jit._shape_functions.nll_loss_forward(%arg0, %arg1, %arg2, %arg3) : (!torch.list<int>, !torch.list<int>, !torch.optional<list<int>>, !torch.int) -> !torch.tuple<list<int>, list<int>>\n"
 "    return %0 : !torch.tuple<list<int>, list<int>>\n"
@@ -14608,6 +14633,14 @@ StringRef mlir::torch::Torch::getAbstractInterpLibrary() {
 "    %4 = call @__torch__.torch_mlir.jit_ir_importer.build_tools.library_generator.promote_dtypes(%2, %3) : (!torch.list<optional<int>>, !torch.list<int>) -> !torch.int\n"
 "    return %4 : !torch.int\n"
 "  }\n"
+"  func.func @\"__torch_mlir_dtype_fn.aten.huber_loss\"(%arg0: !torch.tuple<int, int>, %arg1: !torch.tuple<int, int>, %arg2: !torch.int, %arg3: !torch.float) -> !torch.int {\n"
+"    %0:2 = torch.prim.TupleUnpack %arg0 : !torch.tuple<int, int> -> !torch.int, !torch.int\n"
+"    %1:2 = torch.prim.TupleUnpack %arg1 : !torch.tuple<int, int> -> !torch.int, !torch.int\n"
+"    %2 = torch.prim.ListConstruct %0#0, %1#0 : (!torch.int, !torch.int) -> !torch.list<optional<int>>\n"
+"    %3 = torch.prim.ListConstruct %0#1, %1#1 : (!torch.int, !torch.int) -> !torch.list<int>\n"
+"    %4 = call @__torch__.torch_mlir.jit_ir_importer.build_tools.library_generator.promote_dtypes(%2, %3) : (!torch.list<optional<int>>, !torch.list<int>) -> !torch.int\n"
+"    return %4 : !torch.int\n"
+"  }\n"
 "  func.func @\"__torch_mlir_dtype_fn.aten.mse_loss\"(%arg0: !torch.tuple<int, int>, %arg1: !torch.tuple<int, int>, %arg2: !torch.int) -> !torch.int {\n"
 "    %none = torch.constant.none\n"
 "    %str = torch.constant.str \"AssertionError: \"\n"

lib/Dialect/Torch/Transforms/DecomposeComplexOps.cpp

@@ -10707,6 +10707,91 @@ class DecomposeAtenKlDivOp : public OpRewritePattern<AtenKlDivOp> {
 };
 } // namespace
 
+namespace {
+class DecomposeAtenHuberLossOp : public OpRewritePattern<AtenHuberLossOp> {
+  using OpRewritePattern::OpRewritePattern;
+  LogicalResult matchAndRewrite(AtenHuberLossOp op,
+                                PatternRewriter &rewriter) const override {
+    Location loc = op.getLoc();
+    Value self = op.getSelf();
+    Value target = op.getTarget();
+    Value reductionValue = op.getReduction();
+    Value deltaValue = op.getDelta();
+
+    auto selfTy = cast<ValueTensorType>(self.getType());
+    auto targetTy = cast<ValueTensorType>(target.getType());
+    auto outTy = cast<ValueTensorType>(op.getType());
+    if (!selfTy.hasSizes() || !targetTy.hasSizes() || !outTy.hasSizes()) {
+      return rewriter.notifyMatchFailure(
+          op, "require self, target and output having sizes!");
+    }
+    if (!selfTy.hasDtype() || !targetTy.hasDtype() || !outTy.hasDtype()) {
+      return rewriter.notifyMatchFailure(
+          op, "require self, target and output having dtype!");
+    }
+
+    // Extract delta float value from delta argument
+    // double delta;
+    // if (!matchPattern(deltaValue, m_TorchConstantFloat(&delta))) {
+    //   return rewriter.notifyMatchFailure(op,
+    //                                      "delta should be a constant
+    //                                      float!");
+    // }
+
+    // Squared term: 0.5 * (input - target)^2
+    Value constOne =
+        rewriter.create<ConstantIntOp>(loc, rewriter.getI64IntegerAttr(1));
+    Value constHalf =
+        rewriter.create<ConstantFloatOp>(loc, rewriter.getF64FloatAttr(0.5));
+    Value inputMinusTarget =
+        rewriter.create<AtenSubTensorOp>(loc, selfTy, self, target, constOne);
+    Value squaredValue =
+        rewriter.create<AtenSquareOp>(loc, selfTy, inputMinusTarget);
+    Value squaredTerm =
+        rewriter.create<AtenMulScalarOp>(loc, selfTy, squaredValue, constHalf);
+
+    // Delta scaled term: delta * (|input - target| - 0.5 * delta)
+    Value absDiffValue =
+        rewriter.create<AtenAbsOp>(loc, selfTy, inputMinusTarget);
+    Value halfOfDelta = rewriter.create<AtenMulOp>(
+        loc, rewriter.getType<Torch::FloatType>(), constHalf, deltaValue);
+    Value absDiffMinusDeltaHalf = rewriter.create<AtenSubScalarOp>(
+        loc, selfTy, absDiffValue, halfOfDelta, constOne);
+    Value deltaScaledTerm = rewriter.create<AtenMulScalarOp>(
+        loc, selfTy, absDiffMinusDeltaHalf, deltaValue);
+
+    // Loss calculation based on the condition: |input - target| < delta
+    ValueTensorType boolTy = ValueTensorType::get(
+        op.getContext(), selfTy.getSizes(), rewriter.getI1Type());
+    Value cmpValue =
+        rewriter.create<AtenLeScalarOp>(loc, boolTy, absDiffValue, deltaValue);
+    Value lossPointwise = rewriter.create<AtenWhereSelfOp>(
+        loc, selfTy, cmpValue, squaredTerm, deltaScaledTerm);
+
+    // Extract reduction int value from reduction argument
+    int64_t reduction;
+    if (!matchPattern(reductionValue, m_TorchConstantInt(&reduction))) {
+      return rewriter.notifyMatchFailure(op,
+                                         "reduction should be a constant int!");
+    }
+    Value loss;
+    Value none = rewriter.create<ConstantNoneOp>(loc);
+    // reduction: mean
+    if (reduction == 1) {
+      loss = rewriter.create<AtenMeanOp>(loc, outTy, lossPointwise, none);
+    } else if (reduction == 2) {
+      // reduction: sum
+      loss = rewriter.create<AtenSumOp>(loc, outTy, lossPointwise, none);
+    } else {
+      // reduction: none
+      loss = lossPointwise;
+    }
+    rewriter.replaceOp(op, loss);
+    return success();
+  }
+};
+} // namespace
+
 namespace {
 class DecomposeAtenBinaryCrossEntropyWithLogitsOp
     : public OpRewritePattern<AtenBinaryCrossEntropyWithLogitsOp> {
@@ -12625,6 +12710,7 @@ class DecomposeComplexOpsPass
     addPatternIfTargetOpIsIllegal<DecomposeAtenBinaryCrossEntropyWithLogitsOp>(
         patterns);
     addPatternIfTargetOpIsIllegal<DecomposeAtenKlDivOp>(patterns);
+    addPatternIfTargetOpIsIllegal<DecomposeAtenHuberLossOp>(patterns);
     addPatternIfTargetOpIsIllegal<DecomposeAtenVarMeanDimOp>(patterns);
     addPatternIfTargetOpIsIllegal<DecomposeAtenTopkOp>(patterns);
     addPatternIfTargetOpIsIllegal<DecomposeAtenArgsortOp>(patterns);

lib/Dialect/Torch/Transforms/LowerToBackendContract.cpp

@@ -588,6 +588,7 @@ static void markDecomposedOpsAsIllegal(MLIRContext *context,
   target.addIllegalOp<AtenLogaddexpOp>();
   target.addIllegalOp<AtenLogaddexp2Op>();
   target.addIllegalOp<AtenKlDivOp>();
+  target.addIllegalOp<AtenHuberLossOp>();
 
   for (auto &opName : backendLegalOpsSet) {
     target.addLegalOp(

projects/pt1/python/torch_mlir/jit_ir_importer/build_tools/abstract_interp_lib_gen.py

@@ -2182,6 +2182,14 @@ def aten〇kl_div〡shape(self: List[int], target: List[int], reduction: int = 1
     else:
         assert False, "Invalid reduction value."
 
+def aten〇huber_loss〡shape(self: List[int], target: List[int], reduction: int = 1, delta: float = 1.) -> List[int]:
+    if reduction == 0:
+        return upstream_shape_functions.unary(self)
+    elif reduction in [1, 2]:
+        return []
+    else:
+        assert False, "Invalid reduction value."
+
 @check_shape_function([
     Invocation(TensorOfShape(2, 3), LongTensorOfShape(2), None, 1, -100), # Basic case.
     Invocation(TensorOfShape(3), LongTensorOfShape(), None, 1, -100), # No batch dim.
@@ -4568,6 +4576,14 @@ def aten〇kl_div〡dtype(self_rank_dtype: Tuple[int, int], target_rank_dtype: T
     promoted_dtype = promote_dtypes(ranks, dtypes)
     return promoted_dtype
 
+def aten〇huber_loss〡dtype(self_rank_dtype: Tuple[int, int], target_rank_dtype: Tuple[int, int], reduction: int = 1, delta: float = 1.) -> int:
+    self_rank, self_dtype = self_rank_dtype
+    target_rank, target_dtype = target_rank_dtype
+    ranks: List[Optional[int]] = [self_rank, target_rank]
+    dtypes = [self_dtype, target_dtype]
+    promoted_dtype = promote_dtypes(ranks, dtypes)
+    return promoted_dtype
+
 @check_dtype_function(_check_two_tensor_op(
     output_error_types={torch.bool, torch.int8, torch.uint8, torch.int16, torch.int32, torch.int64}))
 def aten〇mse_loss〡dtype(self_rank_dtype: Tuple[int, int], target_rank_dtype: Tuple[int, int], reduction: int = 1) -> int:

projects/pt1/python/torch_mlir/jit_ir_importer/build_tools/torch_ods_gen.py

@@ -765,6 +765,7 @@ def emit_with_mutating_variants(key, **kwargs):
         "aten::poisson_nll_loss : (Tensor, Tensor, bool, bool, float, int) -> (Tensor)"
     )
     emit("aten::kl_div : (Tensor, Tensor, int, bool) -> (Tensor)")
+    emit("aten::huber_loss : (Tensor, Tensor, int, float) -> (Tensor)")
     emit("aten::bincount : (Tensor, Tensor?, int) -> (Tensor)")
     emit("aten::linalg_vector_norm : (Tensor, Scalar, int[]?, bool, int?) -> (Tensor)")
     emit("aten::linalg_norm : (Tensor, Scalar?, int[]?, bool, int?) -> (Tensor)")

projects/pt1/python/torch_mlir_e2e_test/test_suite/__init__.py

@@ -63,3 +63,4 @@ def register_all_tests():
     from . import meshgrid
     from . import timeout
     from . import kl_div_loss
+    from . import huber_loss

projects/pt1/python/torch_mlir_e2e_test/test_suite/huber_loss.py

@@ -0,0 +1,108 @@
+# Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+# See https://llvm.org/LICENSE.txt for license information.
+# SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+# Also available under a BSD-style license. See LICENSE.
+
+import torch
+
+from torch_mlir_e2e_test.framework import TestUtils
+from torch_mlir_e2e_test.registry import register_test_case
+from torch_mlir_e2e_test.annotations import annotate_args, export
+
+# ==============================================================================
+
+
+class HuberLossModule_default(torch.nn.Module):
+    def __init__(self):
+        super().__init__()
+
+    @export
+    @annotate_args(
+        [
+            None,
+            ([-1, -1, -1], torch.float32, True),
+            ([-1, -1, -1], torch.float32, True),
+        ]
+    )
+    def forward(self, x, y):
+        return torch.ops.aten.huber_loss(x, y)
+
+
+@register_test_case(module_factory=lambda: HuberLossModule_default())
+def HuberLossModule_default_basic(module, tu: TestUtils):
+    module.forward(tu.rand(3, 5, 2), tu.rand(3, 5, 2))
+
+
+# ==============================================================================
+
+
+class HuberLossModule_reduction_is_none(torch.nn.Module):
+    def __init__(self):
+        super().__init__()
+
+    @export
+    @annotate_args(
+        [
+            None,
+            ([-1, -1, -1], torch.float32, True),
+            ([-1, -1, -1], torch.float32, True),
+        ]
+    )
+    def forward(self, x, y):
+        return torch.ops.aten.huber_loss(x, y, delta=2.3, reduction=0)
+
+
+@register_test_case(module_factory=lambda: HuberLossModule_reduction_is_none())
+def HuberLossModule_reduction_is_none_basic(module, tu: TestUtils):
+    module.forward(tu.rand(3, 5, 2), tu.rand(3, 5, 2))
+
+
+# ==============================================================================
+
+
+class HuberLossModule_mean_reduction(torch.nn.Module):
+    def __init__(self):
+        super().__init__()
+
+    @export
+    @annotate_args(
+        [
+            None,
+            ([-1, -1, -1], torch.float32, True),
+            ([-1, -1, -1], torch.float32, True),
+        ]
+    )
+    def forward(self, x, y):
+        return torch.ops.aten.huber_loss(x, y, reduction=1)
+
+
+@register_test_case(module_factory=lambda: HuberLossModule_mean_reduction())
+def HuberLossModule_mean_reduction_basic(module, tu: TestUtils):
+    module.forward(tu.rand(3, 5, 2), tu.rand(3, 5, 2))
+
+
+# ==============================================================================
+
+
+class HuberLossModule_sum_reduction(torch.nn.Module):
+    def __init__(self):
+        super().__init__()
+
+    @export
+    @annotate_args(
+        [
+            None,
+            ([-1, -1, -1], torch.float32, True),
+            ([-1, -1, -1], torch.float32, True),
+        ]
+    )
+    def forward(self, x, y):
+        return torch.ops.aten.huber_loss(x, y, reduction=2)
+
+
+@register_test_case(module_factory=lambda: HuberLossModule_sum_reduction())
+def HuberLossModule_sum_reduction_basic(module, tu: TestUtils):
+    module.forward(tu.rand(3, 5, 2), tu.rand(3, 5, 2))
+
+
+# ==============================================================================