https://github.com/cambridge-iccs/ftorch-benchmarks

Benchmarking for the fortran-pytorch-lib against forpy (disconnected form MiMA)

This score indicates how likely this project is to be science-related based on various indicators:

○
CITATION.cff file
✓
codemeta.json file
Found codemeta.json file
✓
.zenodo.json file
Found .zenodo.json file
○
DOI references
○
Academic publication links
○
Academic email domains
○
Institutional organization owner
○
JOSS paper metadata
○
Scientific vocabulary similarity
Low similarity (5.7%) to scientific vocabulary

Last synced: 10 months ago · JSON representation

Repository

Benchmarking for the fortran-pytorch-lib against forpy (disconnected form MiMA)

Created about 3 years ago · Last pushed about 1 year ago

Readme License

The Python module to use should be: module use /usr/local/software/spack/spack-modules/icelake-python3_10/linux-rocky8-icelake; module load python/3.10.4/gcc/4gjyuriy
Construct a virtual environment with this Python, and pip install torch numpy.
Build the code with the Python-3.10 module above loaded but without the virtual environment activated (it will contain a cmake that doesn't work).
Run the code with both the Python-3.10 module and the virtual environment loaded.
The code should be run so that the input files (uuu.txt, vvv.txt, etc) are located ../cgdrag_model/uuu.txt from the current directory; and that the network_wst.pkl file is in the current directory.
So for example I run it from the pytorch directory with the command ../build/benchmarker_forpy ../pytorch run_emulator_davenet 5

Institute of Computing for Climate Science at the University of Cambridge