add moondream step

app.py

@@ -95,10 +95,10 @@ def infer(image_in, camera_shot, conditional_pose, prompt, style, chosen_model):
     
     iid_img = get_instantID(image_in, conditional_pose, prompt, style)
 
-    #short_cap = get_short_caption(iid_img)
+    short_cap = get_short_caption(iid_img)
     
     if chosen_model == "i2vgen-xl" :
-        video_res = get_video_i2vgen(iid_img, prompt)
+        video_res = get_video_i2vgen(iid_img, short_cap)
     elif chosen_model == "stable-video" :
         video_res = get_video_svd(image_in)