[pypto IR] Is there any option to dump the pypto IR or anywhere to get the example of pypto IR ?

[vfdff]

As we known, there is 3 level IR in the pypto, ie: Tensor IR → Tile IR → Pto IR, and the file docs/en/dev/ir/01-hierarchy.md description the IR AST, but is don't get a test file input by tensor ir （ all the test case are expressed by python），so the pypto can't use the tensor ir as a direct input, which done by llvm mlir ?

• The following give a example in the docs/en/dev/ir/01-hierarchy.md, but it still use python code

# with pl.cluster():
cluster = ir.ClusterScopeStmt(name_hint="", body=body, span=span)

# with pl.at(level=Level.HOST, role=Role.Worker):
hier = ir.HierarchyScopeStmt(level=ir.Level.HOST, role=ir.Role.Worker,
                             name_hint="", body=body, span=span)

# with pl.at(level=Level.CORE_GROUP,
#            optimizations=[pl.split(pl.SplitMode.UP_DOWN)]):
hier_core = ir.HierarchyScopeStmt(level=ir.Level.CORE_GROUP,
                                  split=ir.SplitMode.UP_DOWN,
                                  name_hint="", body=body, span=span)

# with pl.spmd(core_num=8):
spmd = ir.SpmdScopeStmt(core_num=8, sync_start=False,
                        name_hint="", body=body, span=span)

[vfdff]

In test case models/qwen3/14b/qwen3_14b_decode.py, I see dump_passes=True, which is used to control the pass dump, and the format of dumps are expressed by python code.

config=RunConfig(
            rtol=3e-3,
            atol=3e-3,
            compile=dict(dump_passes=True),
            runtime=dict(
                platform=args.platform,
                device_id=args.device,
                runtime_profiling=args.runtime_profiling,
            ),
        ),