https://pypi.org/project/sgemm-bi/ Recent updates to the Python Package Index for sgemm-bi en https://pypi.org/project/sgemm-bi/0.1.1.post2/ Deterministic, batch-invariant CUDA GEMM for PyTorch: bit-identical training matmuls (forward / dW / dX) in f32, bf16, f16, with an opt-in tensor-core tier. Requires Linux x86_64 + NVIDIA Ampere or newer. Fri, 12 Jun 2026 21:54:30 GMT https://pypi.org/project/sgemm-bi/0.1.1.post1/ Deterministic, batch-invariant CUDA GEMM for PyTorch: bit-identical training matmuls (forward / dW / dX) in f32, bf16, f16, with an opt-in tensor-core tier. Requires Linux x86_64 + NVIDIA Ampere or newer. Fri, 12 Jun 2026 21:35:37 GMT https://pypi.org/project/sgemm-bi/0.1.1/ Deterministic, batch-invariant CUDA GEMM for PyTorch: bit-identical training matmuls (forward / dW / dX) in f32, bf16, f16, with an opt-in tensor-core tier. Fri, 12 Jun 2026 19:00:12 GMT