Dependency issue

[sirbradflies]

Can the dependency requirements be loosened? Currently I have a deployment conflict I could not resolve with:

• cbpro 1.1.4 depends on requests==2.13.0
• tensorboard 2.5.0 depends on requests<3 and >=2.21.0

[javabiien]

As workaround, I did the following: create or update your requirement.txt by adding the cbpro dependencies and change the versions to match your requirements, for instance:

sortedcontainers>=1.5.9
requests>=2.13.0
six>=1.10.0
websocket-client>=0.40.0
pymongo>=3.5.1

Use the command pip install -r requirements.txt to install the deps. Then install cbpro without dependencies: pip install cbpro --no-deps

[roboserg]

Can the dependency requirements be loosened? Currently I have a deployment conflict I could not resolve with:

• cbpro 1.1.4 depends on requests==2.13.0
• tensorboard 2.5.0 depends on requests<3 and >=2.21.0

Did you manage to create a profitable ML bot? Did you use reinforcement learning or simple supervised learning?