NEC SX-Aurora TSUBASA: ICM User's Guide
Last revision: 2023-10-27

3. SOL: Transparent Neural Network Acceleration

The SOL project aims at accelerating neural network tasks
[17]. Working as middleware, it integrates with PyTorch, TensorFlow
and MxNet, and supports NEC SX-Aurora TSUBASA as well as standard CPU
and GPU architectures (x86, ARM64, NVIDIA) [18].

The software, including offline documentation, is located in the
/apps/nec/sol directory. SOL is provided as a Python Wheel package (no
system module available) and needs to be installed locally (via pip)
by each user -- see Listing 7. Consult the project website for use
cases, papers, and presentations [17].

[Listing 7] SOL installation
$ pip3 install --user /apps/nec/sol/sol-0.1.8-py3-none-any.whl

References:

[17] SOL: Transparent Neural Network Acceleration
    http://sysml.neclab.eu/projects/sol
[18] SOL: Talks/Publications
    http://sysml.neclab.eu/projects/sol/talks

You are viewing proxied material from sdf.org. The copyright of proxied material belongs to its original authors. Any comments or complaints in relation to proxied material should be directed to the original authors of the content concerned. Please see the disclaimer for more details.