17 improvements for compress function using svd

tensor4all / TensorCrossInterpolation.jl

MIT License

13 stars 0 forks source link

Open shinaoka opened 1 month ago

shinaoka commented 1 month ago

This is still a complete version. @rittermarc

More systematic truncation for SVD. We estimate the truncation error from all the truncated singular values (like truncate in ITensors.jl).
We could either normalize the error or not.
Bug fixes in #17

shinaoka commented 1 month ago

@rittermarc Shall we merge it now or do we need more tests?