LICENSE
README.md
THIRDPARTYNOTICES
pyproject.toml
python/tokenspeed_kernel_amd/__init__.py
python/tokenspeed_kernel_amd/_triton.py
python/tokenspeed_kernel_amd.egg-info/PKG-INFO
python/tokenspeed_kernel_amd.egg-info/SOURCES.txt
python/tokenspeed_kernel_amd.egg-info/dependency_links.txt
python/tokenspeed_kernel_amd.egg-info/requires.txt
python/tokenspeed_kernel_amd.egg-info/top_level.txt
python/tokenspeed_kernel_amd/ops/__init__.py
python/tokenspeed_kernel_amd/ops/attention/__init__.py
python/tokenspeed_kernel_amd/ops/attention/gluon/__init__.py
python/tokenspeed_kernel_amd/ops/attention/gluon/dsa_gfx950.py
python/tokenspeed_kernel_amd/ops/attention/gluon/dsa_score_gfx950.py
python/tokenspeed_kernel_amd/ops/attention/gluon/dsa_topk_gfx950.py
python/tokenspeed_kernel_amd/ops/attention/gluon/mha_decode_gfx950.py
python/tokenspeed_kernel_amd/ops/attention/gluon/mha_extend_gfx950.py
python/tokenspeed_kernel_amd/ops/attention/gluon/mha_prefill_gfx1250.py
python/tokenspeed_kernel_amd/ops/attention/gluon/mha_prefill_gfx950.py
python/tokenspeed_kernel_amd/ops/attention/gluon/mla_decode_bf16_gfx950.py
python/tokenspeed_kernel_amd/ops/attention/gluon/mla_prefill_bf16_gfx950.py
python/tokenspeed_kernel_amd/ops/attention/gluon/rel_mha_decode_gfx950.py
python/tokenspeed_kernel_amd/ops/attention/gluon/rel_mha_extend_gfx950.py
python/tokenspeed_kernel_amd/ops/attention/gluon/rel_mha_prefill_gfx950.py
python/tokenspeed_kernel_amd/ops/attention/gluon/utils.py
python/tokenspeed_kernel_amd/ops/gemm/__init__.py
python/tokenspeed_kernel_amd/ops/gemm/mm_a16w16_gfx950.py
python/tokenspeed_kernel_amd/ops/gemm/mm_a16w16_largem_gfx950.py
python/tokenspeed_kernel_amd/ops/moe/__init__.py
python/tokenspeed_kernel_amd/ops/moe/fused_mxfp_gfx1250.py
python/tokenspeed_kernel_amd/ops/moe/fused_mxfp_gfx950.py
python/tokenspeed_kernel_amd/ops/moe/mxfp4_cdna4_scale_layout.py
python/tokenspeed_kernel_amd/ops/moe/mxfp4_gfx1250_preprocess.py
python/tokenspeed_kernel_amd/ops/moe/mxfp4_gfx950_preprocess.py
python/tokenspeed_kernel_amd/ops/moe/utils.py
python/tokenspeed_kernel_amd/ops/moe/gluon_a4w4_gfx950/__init__.py
python/tokenspeed_kernel_amd/ops/moe/gluon_a4w4_gfx950/decode_kernels.py
python/tokenspeed_kernel_amd/ops/moe/gluon_a4w4_gfx950/decode_stage1.py
python/tokenspeed_kernel_amd/ops/moe/gluon_a4w4_gfx950/decode_stage2.py
python/tokenspeed_kernel_amd/ops/moe/gluon_a4w4_gfx950/moe.py
python/tokenspeed_kernel_amd/ops/moe/gluon_a4w4_gfx950/moe_sorting.py
python/tokenspeed_kernel_amd/ops/moe/gluon_a4w4_gfx950/prefill_stage1.py
python/tokenspeed_kernel_amd/ops/moe/gluon_a4w4_gfx950/prefill_stage2.py
python/tokenspeed_kernel_amd/ops/moe/gluon_a4w4_gfx950/preprocess.py
python/tokenspeed_kernel_amd/ops/moe/gluon_a4w4_gfx950/routing.py
python/tokenspeed_kernel_amd/ops/moe/gluon_a4w4_gfx950/scale.py
python/tokenspeed_kernel_amd/ops/moe/gluon_bf16_moe/__init__.py
python/tokenspeed_kernel_amd/ops/moe/gluon_bf16_moe/_grid.py
python/tokenspeed_kernel_amd/ops/moe/gluon_bf16_moe/moe.py
python/tokenspeed_kernel_amd/ops/moe/gluon_bf16_moe/moe_align_device.py
python/tokenspeed_kernel_amd/ops/moe/gluon_bf16_moe/moe_align_fused.py
python/tokenspeed_kernel_amd/ops/moe/gluon_bf16_moe/stage1_kernel.py
python/tokenspeed_kernel_amd/ops/moe/gluon_bf16_moe/stage1_splitk_kernel.py
python/tokenspeed_kernel_amd/ops/moe/gluon_bf16_moe/stage2_decode_kernel.py
python/tokenspeed_kernel_amd/ops/moe/gluon_bf16_moe/stage2_kernel.py
python/tokenspeed_kernel_amd/ops/moe/gluon_bf16_moe/warp_decode_gluon_kernel.py
python/tokenspeed_kernel_amd/ops/sampling/__init__.py
python/tokenspeed_kernel_amd/ops/sampling/gluon/__init__.py
python/tokenspeed_kernel_amd/ops/sampling/gluon/argmax_gfx950.py
test/test_no_triton_kernels_dependency.py