[torch] Add support for torch.split_with_sizes via decompose (llvm#…

…2979)

Convert to individiual slices and tuple together as a list.

---------

Co-authored-by: Scott Todd <scott.todd0@gmail.com>

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

@@ -12910,6 +12910,30 @@ def Torch_AtenSplitWithSizesOp : Torch_Op<"aten.split_with_sizes", [
   }];
 }
 
+def Torch_AtenSplitSizesOp : Torch_Op<"aten.split.sizes", [
+    AllowsTypeRefinement,
+    ReadOnly
+  ]> {
+  let summary = "Generated op for `aten::split.sizes : (Tensor, int[], int) -> (Tensor[])`";
+  let arguments = (ins
+    AnyTorchTensorType:$self,
+    AnyTorchListOfTorchIntType:$split_size,
+    Torch_IntType:$dim
+  );
+  let results = (outs
+    AnyTorchListOfTensorType:$result
+  );
+  let hasCustomAssemblyFormat = 1;
+  let extraClassDefinition = [{
+    ParseResult AtenSplitSizesOp::parse(OpAsmParser &parser, OperationState &result) {
+      return parseDefaultTorchOp(parser, result, 3, 1);
+    }
+    void AtenSplitSizesOp::print(OpAsmPrinter &printer) {
+      printDefaultTorchOp(printer, *this, 3, 1);
+    }
+  }];
+}
+
 def Torch_AtenUnbindIntOp : Torch_Op<"aten.unbind.int", [
     AllowsTypeRefinement,
     ReadOnly

lib/Dialect/Torch/Transforms/DecomposeComplexOps.cpp

@@ -693,6 +693,131 @@ class DecomposeAtenSelectIntOp : public OpRewritePattern<AtenSelectIntOp> {
 };
 } // namespace
 
+namespace {
+class DecomposePrimTolistOp : public OpRewritePattern<PrimTolistOp> {
+  using OpRewritePattern::OpRewritePattern;
+  LogicalResult matchAndRewrite(PrimTolistOp op,
+                                PatternRewriter &rewriter) const override {
+    auto loc = op.getLoc();
+    auto self = op.getOperands()[0];
+    auto selfTy = dyn_cast<Torch::BaseTensorType>(self.getType());
+    if (!selfTy || !selfTy.hasSizes())
+      return rewriter.notifyMatchFailure(op, "Unknown self shape");
+
+    int64_t rank = selfTy.getSizes().size();
+    if (rank != 1)
+      return rewriter.notifyMatchFailure(op, "Expected rank-1");
+
+    int64_t length = selfTy.getSizes().back();
+    if (length == Torch::kUnknownSize)
+      return rewriter.notifyMatchFailure(op, "Tolist length is unknown");
+
+    auto resultTy = dyn_cast<Torch::ListType>(op.getType(0));
+    if (!resultTy)
+      return rewriter.notifyMatchFailure(op, "Result type is not list");
+
+    auto scalarTy = resultTy.getContainedType();
+    Value zero =
+        rewriter.create<ConstantIntOp>(loc, rewriter.getI64IntegerAttr(0));
+    auto extractTy = rewriter.getType<ValueTensorType>(
+        llvm::SmallVector<int64_t>{1}, selfTy.getOptionalDtype());
+    llvm::SmallVector<Value> results;
+    llvm::SmallVector<int64_t> sizes(selfTy.getSizes());
+    for (int64_t i = 0; i < length; ++i) {
+      Value iv =
+          rewriter.create<ConstantIntOp>(loc, rewriter.getI64IntegerAttr(i));
+      Value extract = rewriter.create<AtenSelectIntOp>(
+          loc, extractTy, self, /*dim=*/zero, /*index=*/iv);
+      Value scalar = rewriter.create<AtenItemOp>(loc, scalarTy, extract);
+      results.push_back(scalar);
+    }
+
+    rewriter.replaceOpWithNewOp<PrimListConstructOp>(op, resultTy, results);
+    return failure();
+  }
+};
+} // namespace
+
+namespace {
+class DecomposeAtenSplitSizesOp : public OpRewritePattern<AtenSplitSizesOp> {
+  using OpRewritePattern::OpRewritePattern;
+  LogicalResult matchAndRewrite(AtenSplitSizesOp op,
+                                PatternRewriter &rewriter) const override {
+    rewriter.replaceOpWithNewOp<AtenSplitWithSizesOp>(
+        op, op->getResultTypes(), op.getSelf(), op.getSplitSize(), op.getDim());
+    return success();
+  }
+};
+} // namespace
+
+namespace {
+class DecomposeAtenSplitWithSizesOp
+    : public OpRewritePattern<AtenSplitWithSizesOp> {
+  using OpRewritePattern::OpRewritePattern;
+  LogicalResult matchAndRewrite(AtenSplitWithSizesOp op,
+                                PatternRewriter &rewriter) const override {
+    auto loc = op.getLoc();
+    Value self = op.getSelf();
+    SmallVector<Value> splitSizes;
+    if (!getListConstructElements(op.getSplitSizes(), splitSizes))
+      return rewriter.notifyMatchFailure(op, "Unable to get sizes");
+
+    if (splitSizes.empty())
+      return rewriter.notifyMatchFailure(op, "No split sizes");
+
+    auto selfTy = dyn_cast<BaseTensorType>(self.getType());
+    if (!selfTy || !selfTy.hasSizes())
+      return rewriter.notifyMatchFailure(op, "Self shape unknown");
+
+    int64_t rank = selfTy.getSizes().size();
+    auto resultTy = dyn_cast<Torch::ListType>(op.getResult().getType());
+    if (!resultTy)
+      return rewriter.notifyMatchFailure(op, "Result type not a list");
+
+    auto sliceTy =
+        dyn_cast_or_null<Torch::BaseTensorType>(resultTy.getContainedType());
+    if (!isa<Torch::BaseTensorType>(sliceTy))
+      return rewriter.notifyMatchFailure(op, "Slice type is unknown");
+
+    int64_t dimInt = 0;
+    bool hasDim = matchPattern(op.getDim(), m_TorchConstantInt(&dimInt));
+    if (dimInt < 0)
+      dimInt += rank;
+
+    auto intTy = rewriter.getType<Torch::IntType>();
+    Value one =
+        rewriter.create<ConstantIntOp>(loc, rewriter.getI64IntegerAttr(1));
+    Value begin =
+        rewriter.create<ConstantIntOp>(loc, rewriter.getI64IntegerAttr(0));
+
+    llvm::SmallVector<Value> slices;
+    llvm::SmallVector<int64_t> sliceSizes(sliceTy.getSizes());
+    int64_t defaultLength = !hasDim ? Torch::kUnknownSize : sliceSizes[dimInt];
+    for (auto size : splitSizes) {
+      Value end = rewriter.create<AtenAddIntOp>(loc, intTy, begin, size);
+
+      int64_t sizeInt;
+      if (hasDim && matchPattern(size, m_TorchConstantInt(&sizeInt))) {
+        sliceSizes[dimInt] = sizeInt;
+      } else if (hasDim) {
+        sliceSizes[dimInt] = defaultLength;
+      }
+
+      sliceTy = rewriter.getType<ValueTensorType>(sliceSizes,
+                                                  sliceTy.getOptionalDtype());
+      Value slice = rewriter.create<AtenSliceTensorOp>(
+          loc, sliceTy, op.getSelf(),
+          /*dim=*/op.getDim(), /*start=*/begin, /*end=*/end, /*step=*/one);
+      slices.push_back(slice);
+      begin = end;
+    }
+
+    rewriter.replaceOpWithNewOp<PrimListConstructOp>(op, resultTy, slices);
+    return success();
+  }
+};
+} // namespace
+
 namespace {
 class DecomposeAtenNarrowOp : public OpRewritePattern<AtenNarrowOp> {
 public:
@@ -7008,6 +7133,8 @@ class DecomposeComplexOpsPass
     addPatternIfTargetOpIsIllegal<DecomposeAtenVarCorrectionOp>(patterns);
     addPatternIfTargetOpIsIllegal<DecomposeAtenStdDimOp>(patterns);
     addPatternIfTargetOpIsIllegal<DecomposeAtenStdCorrectionOp>(patterns);
+    addPatternIfTargetOpIsIllegal<DecomposeAtenSplitSizesOp>(patterns);
+    addPatternIfTargetOpIsIllegal<DecomposeAtenSplitWithSizesOp>(patterns);
     addPatternIfTargetOpIsIllegal<DecomposeAtenNarrowOp>(patterns);
     addPatternIfTargetOpIsIllegal<DecomposeAtenNarrowTensorOp>(patterns);
     addPatternIfTargetOpIsIllegal<DecomposeAtenGluOp>(patterns);
@@ -7035,6 +7162,7 @@ class DecomposeComplexOpsPass
     addPatternIfTargetOpIsIllegal<DecomposeAtenNewEmptyStridedOp>(patterns);
     addPatternIfTargetOpIsIllegal<DecomposeAtenEmptyStridedOp>(patterns);
     addPatternIfTargetOpIsIllegal<DecomposeAtenBucketizeTensorOp>(patterns);
+    addPatternIfTargetOpIsIllegal<DecomposePrimTolistOp>(patterns);
     addPatternIfTargetOpIsIllegal<DecomposePrimsSqueezeOp>(patterns);
     addPatternIfTargetOpIsIllegal<DecomposeAtenMovedimIntOp>(patterns);
     addPatternIfTargetOpIsIllegal<DecomposeAtenOneHotOp>(patterns);

projects/pt1/e2e_testing/xfail_sets.py

@@ -20,7 +20,8 @@
     # 'linalg.depthwise_conv_2d_nchw_chw' op inferred input/output operand #1 has shape's dimension #0 to be 4, but found 8
     "Conv2dWithPaddingDilationStrideStaticModule_depthwise_multiplier",
     "IscloseStaticModule_basic",
-    "IscloseStaticModuleTrue_basic"
+    "IscloseStaticModuleTrue_basic",
+    "SplitWithSizes_Module_basic",
 }
 
 TORCHDYNAMO_XFAIL_SET = {
@@ -1478,15 +1479,6 @@
     "VarBiasedModule_basic",
     "VarMeanBiasedModule_basic",
 
-    # Failure - constant int lowering
-    "SplitTensorGetItem_Module_basic",
-    "SplitTensorLastSmallerModule_basic",
-    "SplitTensorListUnpackModule_basic",
-    "SplitTensorNegativeDimModule_basic",
-    "SplitWithSizesListUnpackModule_basic",
-    "UnbindIntGetItem_Module_basic",
-    "UnbindIntListUnpack_Module_basic",
-
     # Failure - incorrect numerics
     "AdaptiveAvgPool1dUnitOutputSizeDynamicModule_basic",
     "AdaptiveAvgPool2dUnitOutputSizeDynamicModule_basic",

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

@@ -741,6 +741,7 @@ def emit_with_mutating_variants(key, **kwargs):
     emit("aten::sort : (Tensor, int, bool) -> (Tensor, Tensor)", has_folder=True)
     emit("aten::split.Tensor : (Tensor, int, int) -> (Tensor[])")
     emit("aten::split_with_sizes : (Tensor, int[], int) -> (Tensor[])")
+    emit("aten::split.sizes : (Tensor, int[], int) -> (Tensor[])")
     emit("aten::unbind.int : (Tensor, int) -> (Tensor[])")
     emit("aten::chunk : (Tensor, int, int) -> (Tensor[])")
 

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

@@ -897,3 +897,25 @@ def forward(self, x):
 @register_test_case(module_factory=lambda: ChunkListUnpackUnevenDynamic_Module())
 def ChunkListUnpackUnevenDynamic_Module_basic(module, tu: TestUtils):
     module.forward(tu.rand(2, 13, 2))
+
+# ==============================================================================
+
+class SplitWithSizes_Module(torch.nn.Module):
+    def __init__(self):
+        super().__init__()
+
+    @export
+    @annotate_args([
+        None,
+        ([5, -1, -1], torch.float32, True),
+    ])
+    def forward(self, x):
+        split = torch.split(x, [2, 1, 2], dim=0)
+        return split[0], split[1], split[2]
+
+@register_test_case(module_factory=lambda: SplitWithSizes_Module())
+def SplitWithSizes_Module_basic(module, tu: TestUtils):
+    module.forward(tu.rand(5, 2, 2))
+
+
+