4 people 9bd831f501 [Model] New model support for Motif-1-Tiny (#23414) ...

Signed-off-by: ca1207 <ca1207zzz@gmail.com>
Signed-off-by: TaehyunKim <73943231+ca1207@users.noreply.github.com>
Co-authored-by: WyldeCat <skan1543@gmail.com>
Co-authored-by: Jee Jee Li <pandaleefree@gmail.com>

2025-09-10 23:29:40 -07:00

..

deepgemm

[Refactor] Remove Duplicate per_block_cast_to_fp8, Remove Dependencies of DeepGEMM (#21787)

2025-08-01 01:13:27 +00:00

bench_block_fp8_gemm.py

Tuned H100/H200 triton fp8 block configs for fused_qkv_a_proj (#23939)

2025-08-29 10:28:35 -07:00

bench_fp8_gemm.py

[Misc] Add SPDX-FileCopyrightText (#20428)

2025-07-04 07:40:42 +00:00

bench_int8_gemm.py

[Benchmark] Refactor benchmark script for fp8 & int8 (#19627)

2025-06-15 15:15:37 +08:00

bench_nvfp4_gemm.py

[Bench] Add NVFP4 GEMM benchmark script (#20578)

2025-07-09 13:23:48 -04:00

bench_per_token_quant_fp8.py

[Perf][fp8] Use CustomOp abstraction for fp8 quant for better perf (#19830)

2025-07-11 04:56:28 +00:00

benchmark_activation.py

[Benchmark] add benchmark for custom activation op (#23908)

2025-09-06 20:12:05 -07:00

benchmark_bitblas.py

[Bugfix] Add proper comparison for package versions (#22314)

2025-08-06 20:31:03 -07:00

benchmark_cutlass_fp4_moe.py

[Hardware][NVIDIA] FP4 MoE kernel optimization (#19110)

2025-06-05 09:48:26 -07:00

benchmark_grouped_gemm_cutlass.py

[Kernel] CUTLASS MoE FP8: Integrate cuda moe permute/unpermute (#23045)

2025-08-20 10:35:26 -04:00

benchmark_layernorm.py

[Misc] Add SPDX-FileCopyrightText (#19100)

2025-06-03 11:20:17 -07:00

benchmark_lora.py

[Doc]: fix typos in Python comments (#24173)

2025-09-04 08:52:17 -07:00

benchmark_machete.py

[kernel] Support W4A8 on Hopper (#23198)

2025-08-24 06:18:04 +00:00

benchmark_marlin.py

[Bugfix][Benchmark] Fix Marlin benchmark (#19929)

2025-06-24 07:25:12 +09:00

benchmark_moe_align_block_size.py

[Refactor] Remove moe_align_block_size_triton (#21335)

2025-07-26 07:08:29 -07:00

benchmark_moe_permute_unpermute.py

Fix CUDA permute/unpermute for use with DeepGemm Moe (#17934)

2025-07-27 07:08:00 -07:00

benchmark_moe.py

[Bugfix] Fix Qwen3-coder moe tuned config (#24072)

2025-09-07 05:19:46 +00:00

benchmark_mrope.py

[FEAT] [Performance] Add triton mrope to replace the torch code path (#22375)

2025-08-09 11:50:03 -07:00

benchmark_paged_attention.py

[Misc] Add SPDX-FileCopyrightText (#19100)

2025-06-03 11:20:17 -07:00

benchmark_per_token_group_quant.py

[Test] Add Benchmark and Unit Test for per_token_group_quant (#21860)

2025-07-30 07:15:02 -07:00

benchmark_polynorm.py

[Model] New model support for Motif-1-Tiny (#23414)

2025-09-10 23:29:40 -07:00

benchmark_quant.py

[Misc] Add SPDX-FileCopyrightText (#19100)

2025-06-03 11:20:17 -07:00

benchmark_reshape_and_cache_flash.py

[Perf] Optimize reshape_and_cache_flash CUDA Kernel (#22036)

2025-08-01 19:18:51 -04:00

benchmark_rmsnorm.py

[Misc] Add SPDX-FileCopyrightText (#19100)

2025-06-03 11:20:17 -07:00

benchmark_rope.py

[Misc] Add SPDX-FileCopyrightText (#19100)

2025-06-03 11:20:17 -07:00

benchmark_shapes.py

[Misc] Add SPDX-FileCopyrightText (#19100)

2025-06-03 11:20:17 -07:00

benchmark_silu_mul_fp8_quant.py

[Perf] Small optimizations for silu_mul_fp8_quant_deep_gemm (#23265)

2025-08-21 17:56:15 -04:00

benchmark_trtllm_decode_attention.py

[Flashinfer] Support Flashinfer TRTLLM FP8-qkv BF16/FP16-out Attention Kernel (#23647)

2025-09-08 20:53:07 -07:00

benchmark_trtllm_prefill_attention.py

[Flashinfer] Support Flashinfer TRTLLM FP8-qkv BF16/FP16-out Attention Kernel (#23647)

2025-09-08 20:53:07 -07:00

benchmark_w8a8_block_fp8.py

Tuned H100/H200 triton fp8 block configs for fused_qkv_a_proj (#23939)

2025-08-29 10:28:35 -07:00

graph_machete_bench.py

[Misc] Add SPDX-FileCopyrightText (#19100)

2025-06-03 11:20:17 -07:00

requirements.txt

[Kernel] (2/N) Machete - Integrate into CompressedTensorsWNA16 and GPTQMarlin (#7701)

2024-09-23 13:46:26 -04:00

utils.py

[Misc] Add SPDX-FileCopyrightText (#19100)

2025-06-03 11:20:17 -07:00

weight_shapes.py

[kernel] Support W4A8 on Hopper (#23198)

2025-08-24 06:18:04 +00:00