Markus Goetz

Implement sparse matrices

1

**Feature functionality** Design and implement a distributed sparse DNDarray class. Come up with one or more reasonable storage formats (e.g. COO, CRS, CSC, ...) **Additional context** PyTorch implements COO only...

Replace scattered scikit-learn references with central BSD-3 clause reference

10

**Feature functionality** Remove the scattered references to scikit learn (in the docstrings) and replace it with a BSD-3 clause reference.

Documentation - add programming model page

5

**Feature functionality** Add a page describing HeAT's programming model.

Documentation - add 'Hello World' page

2

**Feature functionality** Add a page in the documentation showing a small 'Hello World' usage example.

Allow HeAT to use the OpenCL backend

2

**Feature functionality** Since torch 1.7 other backends than CPU and CUDA are supported. Among them is OpenCL. HeAT should be able to use it. Pay close attention whether torch's OpenCL...

**Feature functionality** Since torch 1.7 other backends than CPU and CUDA are supported. Among them is the HIP, aka AMD's CUDA, backend. HeAT should be able to use it. Moreover,...

**Feature functionality** Allow for DNDarrays to not have all necessary data directly in RAM. Instead, the data is only partially loaded from disk and swapped in and out adaptively. **Additional...

Rework unittest to be more atomic

1

Break down existing unittest cases into singular, more atomic test cases instead of multiple in a single method