Description
Bug Description
It is impossible to create an environment or install the proper dependencies for Flow to work, with the current environment.yml
and requirements.txt
. This is because the dependencies are mostly outdated and have conflicts, which need to be resolved during installation.
I personally had a hard time before I could actually get Flow to work on my Ubuntu 18.04 machine, so I'm posting the changes that MIGHT work for you as well (this totally depends on your machine as I have not tested it on any other OS).
environment.yml:
name: flow
dependencies:
- python==3.7.3
- pip:
- scipy==1.1.0
- lxml==4.4.1
- six==1.11.0
- path.py
- python-dateutil==2.7.3
- pip>=18.0
- tensorflow==1.15.2
- setuptools==41.0.0
- plotly==2.4.0
- gym==0.14.0
- pyprind==2.11.2
- nose2==0.8.0
- cached_property
- joblib==0.10.3
- matplotlib==3.0.0
- dill
- lz4
- ray==0.8.0
- setproctitle
- psutil
- opencv-python
- boto3==1.10.45
- redis
requirements.txt
gym==0.14.0
numpy==1.18.4
scipy==1.1.0
lxml==4.4.1
pyprind==2.11.2
nose2==0.8.0
six==1.11.0
path.py
joblib==0.10.3
python-dateutil==2.7.3
cached_property
pyglet==1.3.2
matplotlib==3.1.0
imutils==0.5.1
numpydoc
ray==0.8.0
opencv-python
dill
lz4
setproctitle
psutil
opencv-python
boto3==1.10.45
redis
pandas==0.24.2
plotly==2.4.0
tabulate
tensorflow==1.15.2
awscli==1.16.309
torch==1.4.0
pytz
tensorboardX
If you still have some dependency issues, just make sure that you remove the versions from the above files and allow pip to resolve any conflicts.
Another tip is to uninstall and reinstall packages if they throw ImportError, ModuleNotFoundError etc.
Also, any help regarding how to install Flow on Windows would be greatly appreciated.