-
We should use the `COO` sparse format when we construct a sparse matrix in several steps (add each transition at a time). In the `COO` format (see `SparseMatrixCOO` from [LuxurySparse.jl](https://gith…
-
I am trying to use this library with a third party C library, but I need to carry some additional context to perform the computations.
My issue boils down to how do I integrate library calls like:
…
-
### Description
My original aim is to compute a huge amount of determinants that can't fit into the memory of a single GPU. I always get memory overflows when I run it on multiple GPUs, and the probl…
-
I've been doing some experimentation with Flash floating point values, and it seems that AVM1 has no way of actually creating a `-0.0` value.
For example, running this code in Flash Player prints `…
-
Hello,
When searching more neighbours than there are samples in the population, one expects `-1` returned indexes and infinite (before `nan_to_num`) distances. I tested this on empty indexes and it…
-
I see you've added Gauss-Jordan. If you're looking for more computational linear algebra tools to add, I highly recommend Golub and Van Loan _Matrix Computations_ . This was *the* reference when I wor…
znmeb updated
2 years ago
-
The gaussian mixture model computations from mean and covariance matrix (obtained from feature's of model - nn.feature), is constantly throwing the error that covariance matrix is not positive defini…
-
Energy dispersion is a better format for storage, as it is less sparse, but for actual computations, one needs energy migration.
We should have a function that transforms the energy dispersion matr…
-
### 🐛 Describe the bug
For generate on llama3.1, I got 9.1 tok/s, but chat is much slower. I got around 1.4 tok/s.
Test laptop: MacBook Pro with M1 Max, 64 GB memory. Sonoma 14.5
Details for bot…
-
I noticed that `CooMatrix::push_matrix` does not threshold or detect zero elements in the input. Thus the resulting sparse matrix is more dense than it needs to be. By comparison, constructing a matri…