xxjx25000 commited on
Commit
0c4e64d
·
verified ·
1 Parent(s): 435f39e

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +23 -26
README.md CHANGED
@@ -32,15 +32,11 @@ dataset_info:
32
  sequence: float64
33
  splits:
34
  - name: test
35
- num_bytes: 1623762
36
- num_examples: 763
37
- download_size: 1278082
38
- dataset_size: 1623762
39
  configs:
40
  - config_name: default
41
  data_files:
42
  - split: test
43
- path: data/test-*
44
  task_categories:
45
  - question-answering
46
  language:
@@ -58,27 +54,6 @@ size_categories:
58
  - n<1K
59
  ---
60
 
61
- ## Changelog
62
- ### [v1.1] - 2025-03-12
63
-
64
- **Enhancements:**
65
- - Adjusted generation hyperparameters and templates to eliminate scenarios that might lead to ambiguous or incorrect responses.
66
- - Improved data formatting for consistency.
67
- - Updated time-series sample length to 1024 to capture more diverse and complex features.
68
-
69
- **Updated Model Evaluations:**
70
- - The following table shows the updated evaluation on models (tokenization method):
71
- | Model | Tokenization Method | Accuracy |
72
- |----------|---------------------|----------|
73
- | gpt-4o | image | 75.2% |
74
- | gpt-4o | plain_text | 51.7% |
75
- | 4o-mini | plain_text | 46.6% |
76
- | o3-mini | plain_text | 59.0% |
77
-
78
- **Additional Information:**
79
- - **Note:** The previous version ([v1.0](https://huggingface.co/datasets/AutonLab/TimeSeriesExam1/resolve/9f23771ca10d66607ee7abca0c5dcbae57349ac2/qa_dataset.json)) is still available for reference.
80
- - **Research code for exam generation via templates is available on [GitHub](https://github.com/moment-timeseries-foundation-model/TimeSeriesExam/tree/exam_generation).**
81
-
82
  # Dataset Card for TimeSeriesExam-1
83
 
84
  This dataset provides Question-Answer (QA) pairs for the paper [TimeSeriesExam: A Time Series Understanding Exam](https://arxiv.org/pdf/2410.14752). Example inference code can be found [here](https://github.com/moment-timeseries-foundation-model/TimeSeriesExam).
@@ -114,6 +89,28 @@ If you find this work helpful, please consider citing our paper:
114
  }
115
  ```
116
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
117
  ## Liscense
118
 
119
  MIT License
 
32
  sequence: float64
33
  splits:
34
  - name: test
 
 
 
 
35
  configs:
36
  - config_name: default
37
  data_files:
38
  - split: test
39
+ path: qa_dataset.json
40
  task_categories:
41
  - question-answering
42
  language:
 
54
  - n<1K
55
  ---
56
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
57
  # Dataset Card for TimeSeriesExam-1
58
 
59
  This dataset provides Question-Answer (QA) pairs for the paper [TimeSeriesExam: A Time Series Understanding Exam](https://arxiv.org/pdf/2410.14752). Example inference code can be found [here](https://github.com/moment-timeseries-foundation-model/TimeSeriesExam).
 
89
  }
90
  ```
91
 
92
+ ## Changelog
93
+ ### [v1.1] - 2025-03-12
94
+
95
+ **Enhancements:**
96
+ - Adjusted generation hyperparameters and templates to eliminate scenarios that might lead to ambiguous or incorrect responses.
97
+ - Improved data formatting for consistency.
98
+ - Updated time-series sample length to 1024 to capture more diverse and complex features.
99
+
100
+ **Updated Model Evaluations:**
101
+ - The following table shows the updated evaluation on models (tokenization method):
102
+ | Model | Tokenization Method | Accuracy |
103
+ |----------|---------------------|----------|
104
+ | gpt-4o | image | 75.2% |
105
+ | gpt-4o | plain_text | 51.7% |
106
+ | 4o-mini | plain_text | 46.6% |
107
+ | o3-mini | plain_text | 59.0% |
108
+
109
+ **Additional Information:**
110
+ - **Note:** The previous version ([v1.0](https://huggingface.co/datasets/AutonLab/TimeSeriesExam1/resolve/9f23771ca10d66607ee7abca0c5dcbae57349ac2/qa_dataset.json)) is still available for reference.
111
+ - **Research code for exam generation via templates is available on [GitHub](https://github.com/moment-timeseries-foundation-model/TimeSeriesExam/tree/exam_generation).**
112
+
113
+
114
  ## Liscense
115
 
116
  MIT License