Further reading

In times when graphics cards are hardly available on the market and often sold overpriced, there are also many GPU-acceleration libraries available. We haven’t covered all those in this chapter and thus, should at least provide a list for further reading.

• cucim, mimicing scikit-image using CUDA

• cudf, data frames similar to pandas, based on CUDA

• numba, a just-in-time python compiler

Deep learning became possible due to developments in GPU technology in the past decade. Hence, deep-learning libraries technially are also libraries that offer some form of GPU-acceleration. Many of them are also capable of processing images.

• Tensorflow

• Keras

• Pytorch