Set dinamically input for pipeline

[JoseEspinosa]

Inside the yml describing the pipeline include a field such:

- input_param: 'input_sequences'

This field should be read by a function which gets the name of it and uses it to populate with the target data set:

param_name = set_input_param (yml_file)
params[param_name] = functionToGetTargetDataSet()

This should be also implemented by the benchmark module in a similar manner. Actually, if this is implemented probably the best way to go will be to treat exactly in the same the benchmark module as the pipeline module and thus, the benchmark should be described in the same manner.