Datasets:

Modalities:
Image
Text
Formats:
parquet
Languages:
English
Size:
< 1K
ArXiv:
Libraries:
Datasets
pandas
License:
leonardPKU commited on
Commit
b3006e7
·
verified ·
1 Parent(s): 9bb7d29

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +64 -9
README.md CHANGED
@@ -24,13 +24,13 @@ dataset_info:
24
  dtype: string
25
  splits:
26
  - name: test_open
27
- num_bytes: 134659773.0
28
  num_examples: 100
29
  - name: test_closed
30
- num_bytes: 67549223.0
31
  num_examples: 150
32
  download_size: 270416985
33
- dataset_size: 202208996.0
34
  - config_name: Domestic Robot
35
  features:
36
  - name: domain
@@ -55,13 +55,13 @@ dataset_info:
55
  dtype: string
56
  splits:
57
  - name: test_open
58
- num_bytes: 91702060.0
59
  num_examples: 100
60
  - name: test_closed
61
- num_bytes: 177827577.0
62
  num_examples: 200
63
  download_size: 105390299
64
- dataset_size: 269529637.0
65
  - config_name: Open-World Game
66
  features:
67
  - name: domain
@@ -86,13 +86,13 @@ dataset_info:
86
  dtype: string
87
  splits:
88
  - name: test_open
89
- num_bytes: 16139511.0
90
  num_examples: 117
91
  - name: test_closed
92
- num_bytes: 19069366.0
93
  num_examples: 141
94
  download_size: 34988721
95
- dataset_size: 35208877.0
96
  configs:
97
  - config_name: Autonomous Driving
98
  data_files:
@@ -112,4 +112,59 @@ configs:
112
  path: Open-World Game/test_open-*
113
  - split: test_closed
114
  path: Open-World Game/test_closed-*
 
 
 
 
 
 
 
115
  ---
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
24
  dtype: string
25
  splits:
26
  - name: test_open
27
+ num_bytes: 134659773
28
  num_examples: 100
29
  - name: test_closed
30
+ num_bytes: 67549223
31
  num_examples: 150
32
  download_size: 270416985
33
+ dataset_size: 202208996
34
  - config_name: Domestic Robot
35
  features:
36
  - name: domain
 
55
  dtype: string
56
  splits:
57
  - name: test_open
58
+ num_bytes: 91702060
59
  num_examples: 100
60
  - name: test_closed
61
+ num_bytes: 177827577
62
  num_examples: 200
63
  download_size: 105390299
64
+ dataset_size: 269529637
65
  - config_name: Open-World Game
66
  features:
67
  - name: domain
 
86
  dtype: string
87
  splits:
88
  - name: test_open
89
+ num_bytes: 16139511
90
  num_examples: 117
91
  - name: test_closed
92
+ num_bytes: 19069366
93
  num_examples: 141
94
  download_size: 34988721
95
+ dataset_size: 35208877
96
  configs:
97
  - config_name: Autonomous Driving
98
  data_files:
 
112
  path: Open-World Game/test_open-*
113
  - split: test_closed
114
  path: Open-World Game/test_closed-*
115
+ license: apache-2.0
116
+ task_categories:
117
+ - multiple-choice
118
+ - visual-question-answering
119
+ language:
120
+ - en
121
+ pretty_name: PCA-Bench
122
  ---
123
+
124
+ <h1 align="center">PCA-Bench</h1>
125
+
126
+ <p align="center">
127
+
128
+ <a href="https://github.com/pkunlp-icler/PCA-EVAL">
129
+ <img alt="Static Badge" src="https://img.shields.io/badge/Github-Online-white">
130
+ </a>
131
+
132
+ <a href="https://arxiv.org/abs/2310.02071">
133
+ <img alt="Static Badge" src="https://img.shields.io/badge/Paper-PCAEVAL-red">
134
+
135
+ <a href="https://huggingface.co/datasets/PCA-Bench/PCA-Bench-V1">
136
+ <img alt="Static Badge" src="https://img.shields.io/badge/Datasets-HuggingFace-yellow">
137
+ </a>
138
+ </p>
139
+
140
+
141
+
142
+
143
+ *PCA-Bench is an innovative benchmark for evaluating and locating errors in Multimodal LLMs when conducting embodied decision making tasks, specifically focusing on perception, cognition, and action.*
144
+
145
+
146
+ ## News
147
+ - PCA-Bench-V1 is released in HuggingFace Datasets (Leaderboard Coming Soon).
148
+
149
+
150
+
151
+ ## Run Evaluation on Accuracy
152
+
153
+ ```python
154
+ #pip install datasets
155
+ from datasets import load_dataset
156
+ dataset_ad = load_dataset("PCA-Bench/PCA-Bench-V1","Autonomous Driving")
157
+ dataset_dr = load_dataset("PCA-Bench/PCA-Bench-V1","Domestic Robot")
158
+ dataset_og = load_dataset("PCA-Bench/PCA-Bench-V1","Open-World Game")
159
+
160
+ # use your model to inference on the test_open/close split with "question_prompt" and "image" given in the datasets and extract the answers.
161
+ # compute the acc regarding the action groundtruth of open track.
162
+
163
+ ```
164
+
165
+ 📢 For close track data, please follow [this file](https://github.com/pkunlp-icler/PCA-EVAL/blob/main/pca-eval/results/chatgpt_holmes_outputs/Autonomous%20Driving.json) to organize your model output. Submit **three JSON files** from different domains, along with your **model name** and **organization** to us via [email](mailto:[email protected]). Ensure you use the dataset's provided prompt as the default input for fair comparison (You should mention in the email if custom prompts were used).
166
+
167
+ We will send the closed track PCA-Eval results of your model to you.
168
+
169
+ ## PCA Evaluation
170
+ To run PCA-Evaluation yourself, please follow the guidelines in the github repo [PCA-EVAL](https://github.com/pkunlp-icler/PCA-EVAL).