I was wondering if it is possible to first train a machine learning algorithm in my research environment and secondly "import" the resulting model or weights vector into the lab environment?

If not, is there a manual workaround, e.g. can I export the resulting weights vector into a csv file and use this file in the lab environment?