3 people e83222396f [ROCm][Perf] W4A16: speed up skinny GEMM decode, including speculative batches (#57199) ...

Signed-off-by: Robert Esclapez Garcia <robert.garcia@amd.com>
Co-authored-by: Claude Opus 4.7 <noreply@anthropic.com>
Co-authored-by: Douglas Lehr <91553416+dllehr-amd@users.noreply.github.com>

2026-10-13 17:21:30 +00:00

..

__init__.py

[Kernel] Remove AllSpark INT8 W8A16 GEMM backend (#58001)

2026-09-22 14:07:34 -07:00

conch.py

[Quant][Kernel] Remove GPTQ Group/Dynamic Activation Ordering (#54809)

2026-09-08 12:45:12 -07:00

cpu.py

[Hardware][Power] Enable W4A16 (AWQ & GPTQ) quantization on POWER10 using VSX (#59149)

2026-09-30 19:44:48 +08:00

cutlass.py

[Quant][Kernel] Remove GPTQ Group/Dynamic Activation Ordering (#54809)

2026-09-08 12:45:12 -07:00

dynamic_4bit.py

[CPU][Feat] Enable KleidiAI INT8_W4A8 for all input dtypes (#34890)

2026-02-26 05:00:10 +00:00

exllama.py

[Quant][Kernel] Remove GPTQ Group/Dynamic Activation Ordering (#54809)

2026-09-08 12:45:12 -07:00

humming.py

[Quantization] Prefer Humming before Marlin backends on SM90 (#56997)

2026-10-06 12:31:35 -07:00

machete.py

[Quant][Kernel] Remove GPTQ Group/Dynamic Activation Ordering (#54809)

2026-09-08 12:45:12 -07:00

marlin.py

[Core][Kernel] Share persistent workspaces for Marlin and Humming (#57421)

2026-09-20 17:28:33 +08:00

MPLinearKernel.py

[Quant][Kernel] Remove GPTQ Group/Dynamic Activation Ordering (#54809)

2026-09-08 12:45:12 -07:00

rdna3_w4a16.py

[Quant][Kernel] Remove GPTQ Group/Dynamic Activation Ordering (#54809)

2026-09-08 12:45:12 -07:00

rdna_hybrid_w4a16.py

[ROCm][Perf] W4A16: speed up skinny GEMM decode, including speculative batches (#57199)

2026-10-13 17:21:30 +00:00

triton_w4a16.py

Add pydocstyle to the ruff rules (#52136)

2026-09-16 16:44:29 +00:00

xpu.py

[Bugfix][XPU] Remove stale W4A8 weight after repacking (#57511)

2026-10-10 04:27:35 +00:00

zentorch.py

[CPU][Zen] Add DA8W4 (W4A8) int4 support for dense and MoE layers (#54024)

2026-10-01 15:18:25 +08:00