Short-range clip captioning on Ego4D HCap (test)
98.35CIDErVideo ReCap
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| Video ReCapProtocol=Finetuned, Visual Encoder=TSF-B, Text Decoder=GPT2, Train Params=339M2024.02 | 98.35 | 48.77 | 28.28 | |
| Video ReCap-UProtocol=Finetuned, Visual Encoder=TSF-B, Text Decoder=GPT2, Train Params=113M2024.02 | 92.67 | 47.9 | 28.08 | |
| LaViLaProtocol=Finetuned, Visual Encoder=TSF-B, Text Decoder=GPT2, Train Params=258M2024.02 | 88.56 | 47.64 | 28.03 | |
| BLIP2Protocol=Zero-Shot, Visual Encoder=ViT-G, Text Decoder=FT5-XL, Train Params=02024.02 | 8.1 | 7.4 | 12.7 |