Merge pull request #336 from marrlab/lb_doc_mk_exp

domainlab/mk_exp.py

@@ -10,10 +10,24 @@
 
 
 
-def mk_exp(task, model, trainer, test_domain, batchsize):
+def mk_exp(task, model, trainer:str, test_domain:str, batchsize:int):
     """
-    test_domain must be a single string
+    Creates a custom experiment. The user can specify the input parameters.
+
+    Input Parameters:
+        - task: create a task to a custom dataset by importing "mk_task_dset" function from
+        "domainlab.tasks.task_dset". For more explanation on the input params refer to the
+        documentation found in "domainlab.tasks.task_dset.py".
+        - model: create a model [NameOfModel] by importing "mk_[NameOfModel]" function from
+        "domainlab.models.model_[NameOfModel]". For a concrete example and explanation of the input
+        params refer to the documentation found in "domainlab.models.model_[NameOfModel].py"
+        - trainer: string,
+        - test_domain: string,
+        - batch size: int
+
+    Returns: experiment
     """
+
     str_arg = f"--aname=apimodel --trainer={trainer} --te_d={test_domain} --bs={batchsize}"
     parser = mk_parser_main()
     conf = parser.parse_args(str_arg.split())