Lu Fang

Results 87 comments of Lu Fang

[Fix][ROCm] Remove unused variables to fix build error on GFX11/12

Retry the failed tests. Curious how the breakage was landed.

[Fix][ROCm] Remove unused variables to fix build error on GFX11/12

Oh, my PR didn't cover the old arches... sorry.

[Fix][ROCm] Remove unused variables to fix build error on GFX11/12

Retrying all the failed tests now.

[Kernel] Integrate DeepGEMM dense block fp8

Curious, do we see perf wins here?

[Kernel] Integrate DeepGEMM dense block fp8

fi we do see perf wins, we can pick this PR up. cc: @chenyang78

Fix PiecewiseCompileInterpreter

Maybe just a case which fails with pt nightly, but runs after your fix.

Fix PiecewiseCompileInterpreter

Retried the tests, let's see the results.

Fix PiecewiseCompileInterpreter

Entrypoints Test failed due to OOM, can we take a look?

LoRA support on llama4

Let's hold on for a bit.

[NVIDIA] Add Cutlass MLA backend

I think we have fixed the CI, could you rebase the PR again?