drop_pose_ratio is not defined (temporal module finetuning)

In /config/disco_w_tm/tem_disco_temp_attn/net.py (that is activated by Temporal Module finetuning),
    if self.args.drop_pose_ratio > 0:

But, drop_pose_ratio is defined nowhere, and raises the error:
"AttributeError: 'EasyDict' object has no attribute 'drop_pose_ratio'".

What value for drop_pose_ratio did you use for temporal module finetuning training, 0.0 or 0.05?