-
I'm not sure if it is proper to open up an issue for documenting my own experiences installing fastFM on a Linux machine without root privilege. Administrator, please move it to some more suitable pla…
-
# Summary
I found on MTL iGPU, if I call FP16 gemm of onemkl (no matter using OneAPI 2024.0 or 2024.2), the program will crash, and if I call it many times, it will cause my machine to freeze direc…
-
Unlike forward-backward, Viterbi doesn't leverage BLAS, so it scales badly with the number of states
-
Hi,
Is it possible to build rchol for MATLAB on MacOSX? In `/rchol/matlab/` I run "make," and I get the following errors:
```
rchol/matlab/rchol_lap/rchol_lap.cpp:97:5: error: unknown type name 'c…
-
This issue is the tracker for benchmarks on pytorch ops with different axes order.
All time counted in $\mu s$.
Notations:
- `B`: batch
- `K` (also `I`, `J`): unit
- `F`: folding
- `R`: rank…
lkct updated
2 months ago
-
Investigate whether using `Tensor` in backend operations can improve speed.
Relevant materials:
https://github.com/dotnet/core/blob/main/release-notes/9.0/preview/preview4/libraries.md#new-tensort…
-
Hi,
I'm trying to build IlluminaHumanMethylationEPICanno.ilm10b5.hg38 package, but package downloading fails constantly with this error:
> devtools::install_github("achilleasNP/IlluminaHumanMet…
-
In working on mamtul, it became apparent that the `numpy/core/src/common/python_xerbla.c` error handler, which should be registered when linking with OpenBLAS, is not being called. This should be used…
-
InternalError: Blas GEMM launch failed : a.shape=(1024, 64), b.shape=(64, 50), m=1024, n=50, k=64
-
Getting the following output when trying to execute the demo:
```
Traceback (most recent call last):
File "./tools/demo.py", line 18, in
from fast_rcnn.test import im_detect
File "/hom…