Hi,
Thank you for sharing your amazing work. I have a question on the dataset for training RM. Your script here shows the dataset_path is
dataset_path=${project_dir}/data/hh_rlhf/rm-not-formatted/hh_rlhf_rm_data.json
However, upon installing the data, there is only one folder .data/hh_rlhf/, and there is no rm-not-formatted.
Can you please share how to get the file, as the original data is not processed, thus can not be used for training?
Hi,
Thank you for sharing your amazing work. I have a question on the dataset for training RM. Your script here shows the dataset_path is
dataset_path=${project_dir}/data/hh_rlhf/rm-not-formatted/hh_rlhf_rm_data.jsonHowever, upon installing the data, there is only one folder .data/hh_rlhf/, and there is no rm-not-formatted.
Can you please share how to get the file, as the original data is not processed, thus can not be used for training?