glopezas commited on
Commit
390c131
·
verified ·
1 Parent(s): 61aa552

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +75 -3
README.md CHANGED
@@ -1,3 +1,75 @@
1
- ---
2
- license: cc-by-sa-4.0
3
- ---
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: cc-by-sa-4.0
3
+ ---
4
+ ---
5
+ pretty_name: Math StackExchange Curated (Parquet, CC BY-SA 4.0)
6
+ dataset_type: text
7
+ license: cc-by-sa-4.0
8
+ language:
9
+ - en
10
+ tags:
11
+ - math
12
+ - question-answering
13
+ - reasoning
14
+ - stackexchange
15
+ - cc-by-sa
16
+ - parquet
17
+ task_categories:
18
+ - question-answering
19
+ task_ids:
20
+ - open-domain-qa
21
+ - abstractive-qa
22
+ configs:
23
+ - config_name: default
24
+ data_files:
25
+ train: math_stackexchange_train.parquet
26
+ validation: math_stackexchange_val.parquet
27
+ test: math_stackexchange_test.parquet
28
+ ---
29
+
30
+ # Math StackExchange Curated (Parquet, CC BY-SA 4.0)
31
+
32
+ This dataset is a curated collection of Math StackExchange (MSE) Q&A pairs packaged in **Parquet** format for efficient loading and training.
33
+ Each example contains a problem statement (`title`, `question_body`), a corresponding community answer (`answer_body`), a tag string (`tags`), and an indicator of whether the answer is the accepted solution (`accepted`).
34
+
35
+ > **Source & License**
36
+ > Content is derived from the **Stack Exchange Data Dump** for **Math StackExchange** (CC BY-SA 4.0, © Stack Exchange Inc.).
37
+ > This derived dataset is released under **CC BY-SA 4.0** to comply with the share-alike terms.
38
+ > Please attribute Math StackExchange and Stack Exchange Inc. when using this dataset.
39
+
40
+ ---
41
+
42
+ ## Splits & Files
43
+
44
+ - `math_stackexchange_train.parquet`
45
+ - `math_stackexchange_val.parquet`
46
+ - `math_stackexchange_test.parquet`
47
+
48
+ > All files are columnar **Parquet** for fast streaming and low storage overhead.
49
+
50
+ ---
51
+
52
+ ## Schema
53
+
54
+ | field | dtype | description |
55
+ |----------------|--------|-------------|
56
+ | `title` | string | Thread title (short problem description). |
57
+ | `question_body`| string | Full post body of the question (often includes LaTeX/MathJax). |
58
+ | `answer_body` | string | Answer text (the primary answer included in this dataset). |
59
+ | `tags` | string | Pipe-delimited MSE tags, e.g. `\|number-theory\|elementary-number-theory\|`. |
60
+ | `accepted` | int64 | `1` if the included answer is the accepted solution, otherwise `0`. |
61
+
62
+ > **Note on `tags`:** tags come as a single string with leading/trailing pipes. See the snippet below to normalize them into a Python list.
63
+
64
+ ---
65
+
66
+ ## Example
67
+
68
+ ```json
69
+ {
70
+ "title": "Numbers are too large to show $65^{64}+64^{65}$ is not a prime",
71
+ "question_body": "I tried to find cycles of powers, but they are too big. Also $65^{n} \\equiv 1(\\text{mod}64)$, so I dont know how to use that.",
72
+ "answer_body": "Hint ... x^4 + 64 y^4 = (x^2+8y^2)^2 - (4xy)^2 = ... [factorizations and references] ...",
73
+ "tags": "|elementary-number-theory|",
74
+ "accepted": 0
75
+ }