apache-tvm-ffi==0.1.12
numpy
packaging
PyYAML>=6.0
tokenspeed-proton>=3.8.10.post20260721
tokenspeed-triton>=3.8.10.post20260721
wheel
psutil
nvidia-cutlass-dsl[cu13]==4.6.0
nvidia-cutlass-dsl-libs-cu13==4.6.0
torch==2.11.0
flashinfer-python==0.6.15.post1
nvidia-ml-py==13.580.82
nvtx
tokenspeed-deepep==1.2.1.post20251110
tokenspeed-deepgemm==2.5.0.post20260629
tokenspeed-flashmla==1.0.0.post20260615
tokenspeed-mla==0.2.1
tokenspeed-fast-hadamard-transform==1.1.0.post20251110
tokenspeed-fa3==3.0.0.post20260408
tokenspeed-fa4==4.0.0.post20260710
quack-kernels>=0.6.1
tokenspeed-triton-kernels==1.0.0.post20260510
tokenspeed-trtllm-kernel==1.2.1.post20260427
