mdmitry1/python311-dev repository overview

⁠About SMLP

SMLP is a tool for optimization, synthesis and design space exploration. It is based on statistical and machine learning techniques combined with formal verification approaches that allows selection of optimal configurations with respect to given constraints on the inputs and outputs of the system under consideration.
SMLP has been applied at Intel for hardware design optimization. It is a general purpose optimization and verification tool applicable to any domain where ML models can be trained on data. The supported modes in SMLP include processing data and training models (neural networks, tree-based, and polynomial models), querying data and models, certifying and verifying assertions, synthesis and Pareto-optimization of configurations, design of experiments (to simulate systems and produce data), feature selection, rule learning/subgroup discovery, root-cause analysis, and more.

If you want to try out SMLP on your optimization problems and require support, please contact the developers through the discussion page⁠.

⁠Docker support for Intel and AMD processors

docker run -it mdmitry1/python311-dev:latest

Within docker container prepend SMLP Python script with xvfb-run. For example:

xvfb-run smlp/src/run_smlp.py -h

• Entering Docker container with optional VNC support

bin/enter

Starting VNC server within container:

./start_vnc

Recommended VNC client: remmina

• Entering Docker container with X11 support on native Linux

bin/enter_x11

Dependencies: socat

• Entering Docker container with X11 support on wslg

bin/enter_wslg

Dependencies: WSL2 with WSLG enabled

• Installation test:

bin/test_install

More details - see smlp/README.md