torch>=2.12.0
cuda-python
nvidia-cutlass-dsl>=4.5.2
nvidia-cutlass-dsl-libs-base>=4.5.2
nvidia-cutlass-dsl-libs-cu13>=4.5.2
apache-tvm-ffi!=0.1.8,!=0.1.8.post0,<0.2,>=0.1.6
rich>=13
transformers>=4.51
safetensors>=0.6

[dev]
pytest
