-
I've built a test model with TorchSharp and it's chewing something between 0.5 and 1 GB of memory per second while training until OOM (all running on CPU).
Following the hints in https://github.com…
pkese updated
3 years ago
-
@gbaydin I'm just wondering what your expectation/desire is here for the programming model
I'm looking at the VAE.fsx sample and trying to run it on GPU. Is the intent that I do this by setting GPU…
dsyme updated
3 years ago
-
Torch has support for ComplexFloat and ComplexDouble tensors.
I took a look at lighting this up for ComplexDouble using System.Numerics.Complex as the .NET in-memory interchange type. However I …
dsyme updated
3 years ago
-
For a number of the functions on tensors, the doc strings are either incomplete or missing information. A systematic scrubbing needs to be done, consulting the libtorch source code and docs in order t…
-
I'd like to shift TorchSharp from CUDA 10.2 to 11.1 to keep things up-to-date
Please let me know if this would be a problem
dsyme updated
3 years ago
-
Hi, I've been trying to locate the libtorch tensor indexing/slicing API in TorchSharp:
https://pytorch.org/cppdocs/notes/tensor_indexing.html
If I haven't missed it, I think it needs to be added t…
-
This is a tracking issue for Linux binary comptibility issues for the native code component.
Currently we build this on Ubuntu 18.04
-----------------------
Original issue:
The `libLib…
dsyme updated
3 years ago
-
I'd like to shift TorchSharp from CUDA 10.2 to 11.1 to keep things up-to-date
Please let me know if this would be a problem
dsyme updated
3 years ago
-
There are problems using LibTorch packages from .NET Interactive and F# Interactive on Linux because the native libraries are not unified into a single directory.
As a workaround you can avoid usin…
dsyme updated
3 years ago
-
After merging, the MacOS Release build fails during tests. Same thing happens on my Mac if I run tests for the 'Release' build.