Andreas Karatzas 275a131233

Merge remote-tracking branch 'origin/main' into fix/gptq-rocm

2025-12-24 17:39:50 +00:00

..

awq

…

cutlass_w4a8

[Kernel]Support W4A8 Grouped GEMM on Hopper (#29691)

2025-12-08 19:29:06 -08:00

fp4

[FIX] FP4 quantization kernel padding initialization bug (#31097)

2025-12-23 08:45:18 -08:00

fused_kernels

[Performance] Fused blockwise quant RMS norm (#27883)

2025-12-07 16:38:04 +00:00

gguf

…

gptq

fix output zeroing race condition in GPTQ GEMM kernels

2025-12-15 21:01:15 +00:00

gptq_allspark

[Kernel][Quantization] add w4a8 support for marlin kernel (#24722)

2025-11-29 07:19:33 -08:00

gptq_marlin

[Kernel][Quantization][MoE] add marlin kernel support for turing (sm75) (#29901)

2025-12-16 14:35:28 -08:00

hadamard/hadacore

Use narrow over indexing in hadacore_transform to prep for ABI stable (#28756)

2025-11-15 01:10:15 -08:00

machete

Fix typos in comments across multiple files (#30345)

2025-12-09 20:05:28 -08:00

marlin/sparse

…

w8a8

[MoE Refactor][3/N] Deprecate cutlass block quant fp8 (b200) (#30990)

2025-12-19 13:09:54 -08:00

activation_kernels.cu

[Performance][B200] silu_mul_quant: pack scales in int32 (#28358)

2025-11-13 10:16:55 -08:00

utils.cuh

…

vectorization_utils.cuh

…

vectorization.cuh

…