fix: fix import of ttx npu impl for quantgemm

[Neuromancer42]

No description provided.

[gemini-code-assist]

Code Review

This pull request exposes int8_gemm_dequant_impl and prepare_b_impl in the NPU kernels initialization file and renames prepare_b to prepare_b_impl in int8_gemm.py. The reviewer noted that renaming prepare_b is a breaking change and pointed out an outdated docstring reference that still uses the old function name.

Important

The consumer version of Gemini Code Assist on GitHub is being sunset. Starting June 18, 2026, new organization installations will be blocked, and all code review activity will officially cease on July 17, 2026.
For more details on the timeline and next steps, please review the Help Documentation.

[github-actions]

Claude Code Review

Verdict: Comment -- Renaming exports look fine but there is a trailing-whitespace/style nit and a potential API-break to double-check.

Summary

Exposes int8_gemm_dequant_impl and prepare_b_impl from the NPU kernels package and renames prepare_b to prepare_b_impl in the module.

Must fix

None.

Suggestions

Suggestions (1)

• [MAJOR] Function rename may break existing callers -- mojo_opset/backends/ttx/kernels/npu/int8_gemm.py:113 -- prepare_b was renamed to prepare_b_impl with no backward-compatible alias; verify no other module imports the old name, or keep prepare_b = prepare_b_impl as a shim.

Nits

Nits (1)

• [NIT] Trailing whitespace after "int8_gemm_dequant_impl", -- mojo_opset/backends/ttx/kernels/npu/__init__.py:135.

Notes

[CHECK] Confirm no callers (tests, higher-level dispatch, other backends) still reference the previous prepare_b symbol.