owl
owl copied to clipboard
owlbarn
Owl - OCaml Scientific Computing @ https://ocaml.xyz
In the Linalg module in `owl-base`, several functions can only be applied to float input type. They should be extended to accept complex ndarrays. The unittest should also be updated...
Einstein summation generalizes most tensor operations (product, matrix multiplication, etc).
In constructing a neural network in Tensorflow, user can choose to initialise the weight value as a `tf.varialble` or `tf.const`. The latter means that the weight will not be updated...
Hello, It would be nice to have them in owl. ``` References : 1- Tipping, M. E. and A. C. Faul (2003). Fast marginal likelihood maximisation for sparse Bayesian models....
After switching from `lacaml` to `owl` on one of my projects, I observed a significant (x2) performance regression. I managed to track it down to the fact that I'm performing...
According to the tutorial, `Mat.uniform_int` should be there. However it seems to be availabe only for sparse matrices and not for dense.
Hi Liang, I am amazed at how much things have been added to this library. I just saw that owl has a LASSO implementation. Which might win owl a new...
of_array on sparse matrices currently calls insert on a loop: https://github.com/owlbarn/owl/blob/c862c0741ceda37e160894254d93501fd50b5acc/src/owl/sparse/owl_sparse_matrix_generic.ml#L534-L538 this results in a lot of memory allocation and memcpy-ing if you're constructing a matrix with a lot of...
Currently, View is a module that's functorized over an ndarray. This implementation is relatively cumbersome, because of the following points: 1. Many of the functions over ndarray should apply to...
The determinant function `Linalg.?.det` fails when supplied a matrix which is singular, but should return zero instead according to the mathematical definition. After looking into the code, it looks like...
