Datasets:

Modalities:
Image
Size:
< 1K
Libraries:
Datasets
License:
Weiyun1025 commited on
Commit
b6a90b5
1 Parent(s): b68bfcb

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +28 -1
README.md CHANGED
@@ -24,4 +24,31 @@ For a robust evaluation, we adopt CircularEval as our evaluation strategy.
24
  Under this setting, a question is considered as correctly answered only when the model consistently predicts the correct answer in each of the N iterations, with N corresponding to the number of choices.
25
  In each iteration, a circular shift is applied to both the choices and the answer to form a new query for the model.
26
 
27
- See our [project](https://github.com/OpenGVLab/all-seeing/all-seeing-v2) to learn more details!
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
24
  Under this setting, a question is considered as correctly answered only when the model consistently predicts the correct answer in each of the N iterations, with N corresponding to the number of choices.
25
  In each iteration, a circular shift is applied to both the choices and the answer to form a new query for the model.
26
 
27
+ CRPE contains the following files:
28
+ - `crpe_exist.jsonl`: the evaluation data of **Existence** split.
29
+ - `crpe_exist_meta.jsonl`: the evaluation data of **Existence** split without CircularEval.
30
+ - `crpe_relation.jsonl`: the evaluation data of **Subject**, **Predicate**, and **Object** split.
31
+ - `crpe_relation_meta.jsonl`: the evaluation data of **Subject**, **Predicate**, and **Object** split without CircularEval.
32
+
33
+ **NOTE**: You should use `crpe_exist.jsonl` and `crpe_relation.jsonl` for evaluation. The evaluation script is presented [here](https://github.com/OpenGVLab/all-seeing/blob/main/all-seeing-v2/llava/eval/eval_crpe.py).
34
+
35
+ See our [project](https://github.com/OpenGVLab/all-seeing/all-seeing-v2) to learn more details!
36
+
37
+ # Citation
38
+
39
+ If you find our work useful in your research, please consider cite:
40
+
41
+ ```BibTeX
42
+ @article{wang2023allseeing,
43
+ title={The All-Seeing Project: Towards Panoptic Visual Recognition and Understanding of the Open World},
44
+ author={Wang, Weiyun and Shi, Min and Li, Qingyun and Wang, Wenhai and Huang, Zhenhang and Xing, Linjie and Chen, Zhe and Li, Hao and Zhu, Xizhou and Cao, Zhiguo and others},
45
+ journal={arXiv preprint arXiv:2308.01907},
46
+ year={2023}
47
+ }
48
+ @article{wang2024allseeing_v2,
49
+ title={The All-Seeing Project V2: Towards General Relation Comprehension of the Open World},
50
+ author={Wang, Weiyun and Ren, Yiming and Luo, Haowen and Li, Tiantong and Yan, Chenxiang and Chen, Zhe and Wang, Wenhai and Li, Qingyun and Lu, Lewei and Zhu, Xizhou and others},
51
+ journal={arXiv preprint arXiv:2402.19474},
52
+ year={2024}
53
+ }
54
+ ```