Final eFlows4HPC Release

The last eFlows4HPC version of the dislib include the following features:

• New versions of the SVD for Pillar I (TSQR, Lanczos, R-SVD)
• New version of the RandomForest used in Pillar III
• Improvements in the GPU software cache to support PyTorch tensors
• Distributed NN training with EDDL and PyTorch
• Compressed saving/load capabilities.

2nd eFlows4HPC Release

This release includes some features developed in the dislib for the eFlows4HPC project

• New import load and save methods
• Support of GPUs for several algorithms
• Distributed training with Pytorch and EDDL models