update

fusion_caption.py

@@ -36,7 +36,7 @@ You are an expert in video captioning. You are given a structured video caption
 3. Do not make any semantic changes to the original fields. Please be sure to follow the original meaning.
 
 ## Output Principles and Orders
-1. First, declare the shot_type, then declare the shot_angle and the shot_position fields in natural and fluent.
+1. First, describe the shot_type, then describe the shot_angle and the shot_position fields in natural and fluent.
 2. Second, eliminate information in the action field that is not related to the timing action, such as background or environment information if action is not empty.
 3. Third, describe each subject with its pure action, appearance, expression, position if these fields exist.
 4. Finally, declare the environment and lighting if the environment and lighting fields are not empty.