Thanks for the amazing work! I have just working on it and i am wondering that prompts inside the json file for the nuprompt dataset are quite simple like { "scene_token": "0ac05652a4c44374998be876ba5cd6fd",
"prompt": [
"In motion",
"Moving",
"On the move"}. so is it a elementary annotation or the final GPT generated text description? if it is for the snippet, then isnt it short enough to describe the object movement?
@wudongming97
Thanks for the amazing work! I have just working on it and i am wondering that prompts inside the json file for the nuprompt dataset are quite simple like { "scene_token": "0ac05652a4c44374998be876ba5cd6fd",
"prompt": [
"In motion",
"Moving",
"On the move"}. so is it a elementary annotation or the final GPT generated text description? if it is for the snippet, then isnt it short enough to describe the object movement?
@wudongming97