I would like to use tensorflow and use the model that I trained / saved in aws S3 buckets.
That being said, I would love if you could include the tensorflow and boto3 library in your backtesting / trading environment.
We would be happy to talk with you over the phone about our support of tensorflow. I have used boto3 before, but am interested in your intended use here. We are interested in supporting the various needs of our crowd research partners. I'll reach out to you via email. Best, Morgan
Hi, was searching if anyone has requested the tensorflow/keras yet. I am here to add my vote: it certainly expands substantially machine learning capabilities beyond sklearn.
Yes, I did request Keras a few months ago - it really is a great tool!
Just tried to import tensorflow, then keras and got an error for each: 'No module named tensorflow' , 'No module named keras'.
Does everyone have to make a seperate request to get access to these libraries or how does it work?
Our core environment from the start has been Python 2.7 and runs a distributed backtesting system using Iron Python, this has limited us regarding modern libraries.
We do have some internal facilities for Python 3 and modern libraries.
We have been working hard on switching the main backtesting system over to Python 3.6 and hope to have that available soon.
The system you use as an external user is the same used by the majority of our internal traders and has the ability to take a working script and swing it over to production with no changes. We have to tread very carefully to ensure that that capability is maintained for our traders and Quants.
In the meantime, we are always interested in speaking with users who are experienced in the use of ML libraries in and out of trading environments.
I you want to discuss further, either PM me or send an email to email@example.com.
Thanks for the answer. I'm quite the opposite of experienced, just starting to look into ML. But good to know that, I will first focus on sklearn then for now.
Please add libraries
(pip install tsfresh,
pip install xarray)