Keisuke Fujii
Keisuke Fujii
(from #529 ) I sometimes want to add a regularization term to the loss or clip its gradient to stabilize the learning process in particular for variational models with neural...
[Cupy](http://docs-cupy.chainer.org) provides [GPU enabled sparse matrix implementation](http://docs-cupy.chainer.org/en/latest/reference/generated/cupyx.scipy.sparse.coo_matrix.html#) that has similar APIs to scipy.sparse. I believe that only a slight modification makes *sparse* compatible to `cupyx.scipy.sparse`. Is it in the scope...
[docs](https://kwgoodman.github.io/bottleneck-doc/reference.html#bottleneck.nansum) says *The data type (dtype) of the output is the same as the input. On 64-bit operating systems, 32-bit input is NOT upcast to 64-bit accumulator and return values.*...
### from numpy [ ] fft [ ] histogram ### from scipy [ ] interpolation [ ] fftpack
### Describe the bug `decomposition.PCA(svd_solver='covariance_eigh')` is less stable with numpy==2.0 I noticed this issue as some tests started failing at the downstream [dask-ml/#997](https://github.com/dask/dask-ml/pull/997) For a certain data input, `pca.transform` gives...
It was pointed out by @mruijzendaal in #34 that > is there a specific reason that something like _to_string(fp.readline().split()) isn't used more often, instead of the _read_until(fp, ' ') ?...
Copied from an email from Soumen Basak I am unable to install the "PY3NJ" package in Thonny (https://thonny.org/). The installation error message had a single-version-externally-managed flag. The page https://pypi.org/project/py3nj/ says...
We have bunch of issues for installation (#10, #18, #20, #25, #26, #32). It would be nice if we could have a pure-python option, even if it is slower.
It looks like installation on windows fails. It says 'successfully installed', but while importing py3nj, ``` ImportError: DLL load failed while importing _wigner: The specified module could not be found....
Updating instruction for installing in macOS. closes #18