ibm / vsrl-framework Goto Github PK
View Code? Open in Web Editor NEWThe Verifiably Safe Reinforcement Learning Framework
License: MIT License
The Verifiably Safe Reinforcement Learning Framework
License: MIT License
I am trying to use the framework, but unfortunately the provided scripts do not work. I'm not sure if this is an issue on my end, or if it may be because of unspecified dependency versions. I would be really happy for some help.
Running the scripts/train_rl.py
should work. The error I get is the following.
Add a requirements.txt
file
As stated in the README.
python scripts/train_rl.py --env_name ACC --pbar true
P.S. it is required to specify the pbar parameter as true/false. The README does not state that.
Branch - master
Commit hash - 93d85dd
Environment - macOS
Move the java path into the configuration file, using java as a default. Check that the java version comports with what KeYmaera X expects.
vsrl-framework/vsrl/verifier/keymaerax.py
Line 37 in ded48a7
System Information
Describe the current behavior
Running python setup.py install
in a virtual environment fails to install. This is due to errors about install pyrl and version problems with pillow and chardet. After fixing the aforementioned issues, tests fail because of missing pytest dependency.
Describe the expected behavior
Creating a virtual environment and running python setup.py install
should work out of the box with no errors. Tests should also work without getting an issue about missing dependency.
Standalone code to reproduce the issue
virtualenv --clear --python=python3.7 venv
source venv/bin/activate
python setup.py install
Below is my fix for the setup.py
from distutils.core import setup
from setuptools import find_packages
opencv_pkg = ""
if "DISPLAY" not in os.environ.keys():
opencv_pkg = "opencv-python-headless"
else:
opencv_pkg = "opencv-python"
setup(
name="vsrl",
version="0.0.1",
description="Visceral: A Framework for Verifiably Safe Reinforcement Learning",
author="IBM Research",
author_email="[email protected]",
url="https://visceral.safelearning.ai",
packages=find_packages(),
install_requires=[
"scipy",
"numpy",
"torch",
"pillow==7.2.0",
"chardet==3.0.4",
opencv_pkg,
"pytorch_lightning",
"comet_ml",
"psutil",
"torchvision",
"parsimonious",
"matplotlib",
"portion",
"toml",
"auto-argparse",
"gym",
"pytest",
],
extras_require={"dev": ["pytest", "pytest-cov"]},
dependency_links=["http://github.com/astooke/rlpyt/tarball/master"]
)
Maintain the option to specify a local KeYmaera X installation, but use web.keymaerax.org by default to make installation easier.
We should have tests covering the following:
A declarative, efficient, and flexible JavaScript library for building user interfaces.
๐ Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
The Web framework for perfectionists with deadlines.
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. ๐๐๐
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
We are working to build community through open source technology. NB: members must have two-factor auth.
Open source projects and samples from Microsoft.
Google โค๏ธ Open Source for everyone.
Alibaba Open Source for everyone
Data-Driven Documents codes.
China tencent open source team.