BT-Adapter: Video Conversation is Feasible Without Video Instruction Tuning

Paper	Weights	Video-text Pretraining	Downstream Evaluation	Instruction Tuning	VideoChatGPT Evaluation
		Video-text Pretraining	Downstream Evaluation	Instruction Tuning	VideoChatGPT Evaluation

Overview and Highlights

💡 Plug-and-use, parameter-efficient, multimodal-friendly, and temporal-sensitive structure

💡 State-of-the-art zero-shot results on various video tasks using thousands of fewer GPU hours

💡 State-of-the-art video conversation results with and without video instruction tuning

Qualitative Results

The Evaluation of BT-Adapter's Performance across Different Situations.

👀 The Sequence of Actions

👀 Unusual Actions

👀 Complex Actions and Scenes In A Long Video

Citation

If you find the code useful for your research, please consider citing our paper:

@article{liu2023one,
  title={One for all: Video conversation is feasible without video instruction tuning},
  author={Liu, Ruyang and Li, Chen and Ge, Yixiao and Shan, Ying and Li, Thomas H and Li, Ge},
  journal={arXiv preprint arXiv:2309.15785},
  year={2023}
}