[Torch Dialect] Add split.tensor support + recompose rules (#2102)

* add split.tensor support + recompose rules

* add e2e test

* address comments

* address comments

* erase op in recomposeOp

---------

Co-authored-by: zhekun.zhang <zhekun.zhang@bytedance.com>

build_tools/autogen_ltc_backend.yaml

@@ -8,6 +8,7 @@ blacklist:
 - index_put_  # Error: TODO not sure if there are other valid types to handle here
 
 # Ops with list of tensors output
+- split.Tensor
 - unbind.int
 
 # Additional ops which autogen is supported for but don't compile yet

e2e_testing/xfail_sets.py

@@ -726,6 +726,7 @@ STABLEHLO_PASS_SET = {
     "PrimsViewOfModule_basic",
     "PrimsViewOfZeroRankModule_basic",
     "AtenComplex64Module_basic",
+    "SplitTensorGetItem_Module_basic",
     "UnbindIntListUnpack_Module_basic",
     "UnbindIntGetItem_Module_basic",
 }
@@ -1012,6 +1013,7 @@ TOSA_PASS_SET = {
     "TensorsConcatStaticModule_basic",
     "TensorsConcatNegativeDimStaticModule_basic",
     "AtenComplex64Module_basic",
+    "SplitTensorGetItem_Module_basic",
 }
 
 LTC_XFAIL_SET = {
@@ -1191,6 +1193,7 @@ LTC_XFAIL_SET = {
     "AtenComplexImagModule_basic",
     "AtenComplexRealModule_basic",
     "AtenComplexViewModule_basic",
+    "SplitTensorGetItem_Module_basic",
     "UnbindIntListUnpack_Module_basic",
     "UnbindIntGetItem_Module_basic",
 }

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

@@ -9566,6 +9566,30 @@ def Torch_AtenSortOp : Torch_Op<"aten.sort", [
   }];
 }
 
+def Torch_AtenSplitTensorOp : Torch_Op<"aten.split.Tensor", [
+    AllowsTypeRefinement,
+    ReadOnly
+  ]> {
+  let summary = "Generated op for `aten::split.Tensor : (Tensor, int, int) -> (Tensor[])`";
+  let arguments = (ins
+    AnyTorchTensorType:$self,
+    Torch_IntType:$split_size,
+    Torch_IntType:$dim
+  );
+  let results = (outs
+    AnyTorchListOfTensorType:$result
+  );
+  let hasCustomAssemblyFormat = 1;
+  let extraClassDefinition = [{
+    ParseResult AtenSplitTensorOp::parse(OpAsmParser &parser, OperationState &result) {
+      return parseDefaultTorchOp(parser, result, 3, 1);
+    }
+    void AtenSplitTensorOp::print(OpAsmPrinter &printer) {
+      printDefaultTorchOp(printer, *this, 3, 1);
+    }
+  }];
+}
+
 def Torch_AtenUnbindIntOp : Torch_Op<"aten.unbind.int", [
     AllowsTypeRefinement,
     ReadOnly

lib/Dialect/Torch/Transforms/RecomposeComplexOps.cpp

@@ -181,6 +181,48 @@ public:
     return success();
   }
 };
+
+class RecomposeSplitTensorGetItemOp
+    : public OpRewritePattern<Aten__Getitem__TOp> {
+public:
+  using OpRewritePattern::OpRewritePattern;
+  LogicalResult matchAndRewrite(Aten__Getitem__TOp op,
+                                PatternRewriter &rewriter) const override {
+    // recompose AtenSplitTensorOp + __getitem__t to AtenSliceTensorOp
+    auto splitTensorOp =
+        dyn_cast<AtenSplitTensorOp>(op.getList().getDefiningOp());
+    if (!splitTensorOp)
+      return failure();
+    if (isListPotentiallyMutated(splitTensorOp.getResult()))
+      return failure();
+    int64_t index;
+    if (!matchPattern(op.getIdx(), m_TorchConstantInt(&index)))
+      return rewriter.notifyMatchFailure(
+          op, "Expected `idx` of `Aten__Getitem__TOp` to be a constant int");
+
+    int64_t splitSize;
+    if (!matchPattern(splitTensorOp.getSplitSize(),
+                      m_TorchConstantInt(&splitSize)))
+      return rewriter.notifyMatchFailure(
+          op,
+          "Expected `SplitSize` of `AtenSplitTensorOp` to be a constant int");
+
+    Location loc = op.getLoc();
+    Value step =
+        rewriter.create<ConstantIntOp>(loc, rewriter.getI64IntegerAttr(1));
+    Value start = rewriter.create<ConstantIntOp>(
+        loc, rewriter.getI64IntegerAttr(index * splitSize));
+    Value end = rewriter.create<ConstantIntOp>(
+        loc, rewriter.getI64IntegerAttr(index * splitSize + splitSize));
+    Value sliceTensorOp = rewriter.create<AtenSliceTensorOp>(
+        loc, op.getResult().getType(), splitTensorOp.getSelf(),
+        splitTensorOp.getDim(), start, end, step);
+    rewriter.replaceOp(op, sliceTensorOp);
+    if (splitTensorOp.getResult().use_empty())
+      rewriter.eraseOp(splitTensorOp);
+    return success();
+  }
+};
 } // namespace
 
 namespace {
@@ -194,6 +236,7 @@ public:
     // pattern.add calls go here
     patterns.add<RecomposeSliceCopy_>(context);
     patterns.add<RecomposeSelectFill_>(context);
+    patterns.add<RecomposeSplitTensorGetItemOp>(context);
     patterns.add<RecomposeUnbindListUnpack>(context);
     patterns.add<RecomposeUnbindGetItem>(context);
 

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

@@ -590,6 +590,7 @@ def emit_ops(emitter_td: TextEmitter, registry: Registry):
     emit("aten::any.bool : (bool[]) -> (bool)")
     emit("aten::sort.int : (int[], bool) -> ()", has_canonicalizer=True)
     emit("aten::sort : (Tensor, int, bool) -> (Tensor, Tensor)")
+    emit("aten::split.Tensor : (Tensor, int, int) -> (Tensor[])")
     emit("aten::unbind.int : (Tensor, int) -> (Tensor[])")
 
     # Str ops.

python/torch_mlir_e2e_test/test_suite/slice_like.py

@@ -581,3 +581,24 @@ class UnbindIntGetItem_Module(torch.nn.Module):
 @register_test_case(module_factory=lambda: UnbindIntGetItem_Module())
 def UnbindIntGetItem_Module_basic(module, tu: TestUtils):
     module.forward(tu.rand(2, 3, 4))
+
+
+# ==============================================================================
+
+class SplitTensorGetItem_Module(torch.nn.Module):
+    def __init__(self):
+        super().__init__()
+
+    @export
+    @annotate_args([
+        None,
+        ([2, 3, 4], torch.float32, True),
+    ])
+    def forward(self, x):
+        splits = torch.ops.aten.split(x, 1, 0)
+        return torch.ops.aten.sub(splits[0], splits[1])
+
+@register_test_case(module_factory=lambda: SplitTensorGetItem_Module())
+def SplitTensorGetItem_Module_basic(module, tu: TestUtils):
+    module.forward(tu.rand(2, 3, 4))
+