Tensor library & inference framework for machine learning

cuda gpu large-language-models machine-learning pytorch tensor-algebra
2 Open Issues Need Help Last updated: Jun 25, 2025

Open Issues Need Help

View All on GitHub

AI Summary: Implement a Mixture-of-Experts (MoE) model using the DSC library, leveraging its PyTorch-compatible API. This involves creating the MoE model in Python using DSC's `nn.Module`, potentially extending the DSC API if necessary, and ensuring compatibility without requiring new C++ kernels.

Complexity: 4/5
good first issue model

Tensor library & inference framework for machine learning

C++
#cuda#gpu#large-language-models#machine-learning#pytorch#tensor-algebra

AI Summary: Benchmark the performance of calling a C++ function from Python using three different methods: ctypes, CFFI, and nanobind. This involves creating a simple C++ function, compiling it into a shared library, and then writing Python code to call the function using each of the three methods. The results will be compared to determine which method offers the best performance.

Complexity: 4/5
enhancement good first issue

Tensor library & inference framework for machine learning

C++
#cuda#gpu#large-language-models#machine-learning#pytorch#tensor-algebra