feat: align tensordot with PyTorch Python API

[voltjia]

Summary

• Add exactly two PyTorch-compatible tensordot(a, b, dims, out) forms: integer dims and a pair of dimension lists.
• Keep the previous split dims_self, dims_other interface as a [[deprecated]] compatibility overload.

Motivation

The previous interface only exposed the lower-level pair of dimension lists. PyTorch exposes one dims argument that accepts either an integer or a pair of lists. The earlier revision also added a no-dims convenience overload; this revision removes it to keep the overload set minimal.

No standalone issue is associated with this alignment follow-up.

Type of Change

• feat - new feature / new operator / new platform
• fix - bug fix
• perf - performance improvement (no behavioral change)
• refactor - code restructuring without behavior change
• test - adding or fixing tests only
• docs - documentation only
• build / ci - build system or CI configuration
• chore - tooling, formatting, or other non-code changes
• Breaking change

Platforms Affected

• CPU (WITH_CPU)
• NVIDIA (WITH_NVIDIA)
• Iluvatar (WITH_ILUVATAR)
• MetaX (WITH_METAX)
• Cambricon (WITH_CAMBRICON)
• Moore (WITH_MOORE)
• Ascend (WITH_ASCEND)
• PyTorch C++ bindings (WITH_TORCH)
• Build system / CMake / CI
• Python bindings / user-facing API

Smoke Test Result

Host: ssh nvidia
Image: accelerator-dev/nvidia:latest
PyTorch: 2.10.0a0+b4e4ee8

cmake ... -DWITH_CPU=ON -DWITH_NVIDIA=ON -DWITH_TORCH=ON \
  -DINFINI_OPS_OPS=tensordot -DINFINI_OPS_TORCH_OPS=tensordot
cmake --build build-nvidia --target ops -j 16
Build passed; codegen emitted 1 ATen implementation overload.

python -m pytest tests/test_generate_wrappers.py -q
13 passed

python -m pytest tests/test_tensordot.py --devices cpu nvidia -q
36 passed

Test Results on Supported Platforms

Platform	Affected	Build / Smoke Result	Full Result / Notes
CPU	Yes	targeted build passed	Integer, nested-list, and deprecated forms passed
NVIDIA	Yes	targeted build passed	Integer, nested-list, and deprecated forms passed
Iluvatar	No	N/A - not affected	N/A - not affected
MetaX	No	N/A - not affected	N/A - not affected
Cambricon	No	N/A - not affected	N/A - not affected
Moore	No	N/A - not affected	N/A - not affected
Ascend	No	N/A - not affected	N/A - not affected

Benchmark / Performance Impact

N/A. Integer dimensions are expanded during operator construction; the existing ATen execution path is unchanged.

Notes for Reviewers

API alignment

Interface	Signature	Alignment basis
InfiniOps integer form	tensordot(a, b, dims, out) where dims is int64_t	Matches one PyTorch public dims form; explicit out remains last per CONTRIBUTING.md.
InfiniOps list form	tensordot(a, b, dims, out) where dims is two dimension lists	Matches the other PyTorch public dims form.
PyTorch public API	torch.tensordot(a, b, dims=2, out=None)	Pinned PyTorch source
Deprecated compatibility	tensordot(input, other, dims_self, dims_other, out)	Preserves the pre-alignment split-list interface for migration.

• C++ callers pass dims=2 explicitly when they want the PyTorch default.
• No general wrapper-generator change is added by this PR.
• Ruff 0.15.22 and clang-format 21.1.8 checks passed.