v0.1.0

What's Changed

• feat: integrate faer for Standard GEMM and migrate to column-major layout by @GiggleLiu in #1
• Fix unary einsum operations with naive loop approach by @GiggleLiu in #11
• feat: showcase examples and complex number support by @GiggleLiu in #12
• feat(einsum): support unary gradients via index-exchange trick by @GiggleLiu in #13
• feat(einsum): add tropical unary gradients with argmax routing by @GiggleLiu in #14
• feat: add MPS Heisenberg ground state showcase example by @GiggleLiu in #15
• feat: unified Backend trait with contract API by @GiggleLiu in #18
• test: add comprehensive test coverage ported from OMEinsum.jl by @GiggleLiu in #21
• test: add comprehensive test coverage ported from OMEinsum.jl by @GiggleLiu in #24
• feat(cuda): add cuTENSOR version hints and high-level einsum tests by @GiggleLiu in #20
• test(cuda): add comprehensive GPU test coverage (107 tests) by @GiggleLiu in #26
• feat(einsum): add cost_and_gradient API for optimized backward pass by @GiggleLiu in #27
• Unify gradient tape and test harness by @GiggleLiu in #35
• Fix contraction to handle trace modes by @shinaoka in #30
• feat: add omeinsum CLI workspace tool by @GiggleLiu in #36
• fix(einsum): support scalar-output backward passes by @GiggleLiu in #37
• feat(cli): add autodiff subcommand by @GiggleLiu in #38
• perf: reduce contraction materialization overhead by @GiggleLiu in #39
• Add benchmarks and optimize CPU GEMM output writes by @GiggleLiu in #29
• Clean up archived plan docs by @GiggleLiu in #42

New Contributors

• @GiggleLiu made their first contribution in #1
• @shinaoka made their first contribution in #30

Full Changelog: https://github.com/tensor4all/omeinsum-rs/commits/v0.1.0