https://github.com/berkeleyautomation/rlqp_train
Environment and training for RQLP (Accelerating Quadratic Optimization with RL)
https://github.com/berkeleyautomation/rlqp_train
Last synced: 2 months ago
JSON representation
Environment and training for RQLP (Accelerating Quadratic Optimization with RL)
- Host: GitHub
- URL: https://github.com/berkeleyautomation/rlqp_train
- Owner: BerkeleyAutomation
- Created: 2021-06-21T15:31:54.000Z (almost 4 years ago)
- Default Branch: master
- Last Pushed: 2022-12-07T21:28:44.000Z (over 2 years ago)
- Last Synced: 2025-01-25T21:26:33.653Z (4 months ago)
- Language: Python
- Homepage: https://berkeleyautomation.github.io/rlqp/
- Size: 61.5 KB
- Stars: 3
- Watchers: 7
- Forks: 0
- Open Issues: 1
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# IMPORTANT! INCOMPLETE
We are in the process of cleaning up and migrating research-quality code to this repo. This repo currently only has an implementation of DDPG, and TD3 is not yet tested. DDPG, however seems to work well in practice.
Feel free to file an issue with any questions or concerns.
# Submodules
rlqp_train (this repo, not a fork)
+- rlqp_benchmarks (fork of https://github.com/osqp/osqp_benchmarks)
# Requirements:
The RLQP python solver must be separately installed.
rlqp-python (fork of https://github.com/osqp/osqp-python)
+- rlqp_solver (fork of https://github.com/osqp/osqp.git)
+- qdldl (https://github.com/osqp/qdldl, not forked)
# Setup instructions
% git submodule update --init --recursive
# Virtualenv Instructions
For all these instructions, we assume the environment variable `RLQP` is set to the directory in which this README.md file is contained.
Create a virtual environment (1 time)
% export RLQP=`pwd`
% cd $RLQP
% virtualenv venv/rlqp
Activate the virtual environment (run once in each new terminal):
% . $RLQP/venv/rlqp/bin/activate
Compile and install the RLQP Python wrapper (1 time per git update)
% pip uninstall -y rlqp
% cd $RLQP/rlqp-python
% rm -rf build extension/src/*.a osqp_sources/build
% python setup.py install
Install rlqp_benchmarks to the venv:
% cd $RLQP/rlqp_benchmarks
% python setup.py develop
Place this project in development mode (1 time)
% cd $RLQP
% python setup.py develop
# Example Training:
% python scripts/train.py
# Monitoring Training with Tensorboard
Assuming logs are being written to a subdirectory in experiments, run (with the virtualenv active):
% tensorboard --logdir experiments