set_repeat_penalty sets the repeat penalty related to the model in the workflow.
Usage
set_repeat_penalty(workflow_obj, repeat_penalty)
Arguments
- workflow_obj
an ai_workflow object created by ai_workflow() in the first place.
- repeat_penalty
the repeat penalty value used by the model. A value of 1 means there is no penalty. A value higher than 1 means there is increased penalty on repetition of tokens.
Details
Set the repeat_penalty to be used by the model in the workflow.
Examples
my_workflow <- ai_workflow() |>
set_model(model_name="llama3:8b-instruct-q5_0") |>
set_repeat_penalty(1.3)