Datasets:

Modalities:
Text
Formats:
csv
Languages:
English
Tags:
math
License:
Gugu8 commited on
Commit
f2df2a7
·
verified ·
1 Parent(s): 1f7a75f

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +83 -0
README.md CHANGED
@@ -2,4 +2,87 @@
2
  license: other
3
  license_name: open-data-attribution-training-disclosure-license-odatl-1.0
4
  license_link: LICENSE
 
 
 
 
5
  ---
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
2
  license: other
3
  license_name: open-data-attribution-training-disclosure-license-odatl-1.0
4
  license_link: LICENSE
5
+ language:
6
+ - en
7
+ tags:
8
+ - math
9
  ---
10
+ # Math Teaching Dataset - 100MB / 735k Rows
11
+
12
+ A large math dataset designed for training and fine-tuning AI models with step-by-step reasoning.
13
+
14
+ ## Files
15
+
16
+ - `math_dataset_100MB.csv` - Full dataset, 100.33 MB, 735,000 rows
17
+ - `math_dataset_100MB.csv.gz` - Compressed version, 18.95 MB
18
+ - `splits/math_part_1.csv` to `math_part_10.csv` - 10 x ~10MB chunks
19
+
20
+ ## Dataset Stats
21
+
22
+ - **Total rows:** 735,000
23
+ - **Size:** 100.33 MB raw, 18.95 MB gzipped
24
+ - **Format:** CSV, UTF-8, with header
25
+
26
+ ### Topic Distribution
27
+
28
+ | Topic | Count | Description |
29
+ |-------|-------|-------------|
30
+ | arithmetic | 184,304 | Addition, subtraction, multiplication, division, order of operations |
31
+ | algebra | 91,880 | Linear equations and quadratics |
32
+ | calculus | 91,675 | Basic derivatives and integrals |
33
+ | geometry | 46,373 | Area and perimeter of rectangles, circles, triangles |
34
+ | exponents | 46,303 | Powers |
35
+ | word_problems | 45,929 | Applied addition problems |
36
+ | statistics | 45,920 | Mean / average |
37
+ | percentages | 45,857 | Percentage calculations |
38
+ | number_theory | 45,789 | GCD and LCM |
39
+ | roots | 45,567 | Square roots |
40
+ | fractions | 45,403 | Fraction addition, subtraction, multiplication |
41
+
42
+ ### Difficulty
43
+ - `easy` - single-step problems
44
+ - `medium` - multi-step problems
45
+
46
+ ## Schema
47
+
48
+ | Column | Type | Description |
49
+ |--------|------|-------------|
50
+ | `id` | int | Unique row ID |
51
+ | `topic` | string | High-level topic |
52
+ | `difficulty` | string | easy or medium |
53
+ | `question` | string | The math problem |
54
+ | `answer` | string | Final answer |
55
+ | `reasoning` | string | Step-by-step explanation |
56
+ | `category` | string | Specific sub-type of problem |
57
+
58
+ ## Example Rows
59
+
60
+ "What is 1900 / 5?" -> Answer: 380
61
+ Reasoning: "Divide: 1900 ÷ 5 = 380 because 5*380=1900."
62
+
63
+ "Solve for x: 12x + 15 = 87" -> Answer: 6
64
+ Reasoning: "Subtract 15 from both sides: 12x = 72. Then divide by 12: x = 6."
65
+
66
+ "What is 3/4 + 2/5?" -> Answer: "23/20"
67
+ Reasoning: "Common denominator 20. 3/4=15/20, 2/5=8/20. Sum = 23/20."
68
+
69
+ "What is the derivative of 7x^5 with respect to x?" -> Answer: "35x^4"
70
+ Reasoning: "Power rule: d/dx [x^n] = n*x^(n-1). So 7*5*x^4 = 35x^4."
71
+
72
+ ## Intended Use
73
+
74
+ Built for instruction tuning and chain-of-thought training.
75
+
76
+ Recommended training format:
77
+ Input: {question}
78
+ Target: {reasoning} Answer: {answer}
79
+
80
+ Works for:
81
+ - Fine-tuning language models for math reasoning
82
+ - Teaching step-by-step problem solving
83
+ - Evaluating math capabilities
84
+ - Data augmentation alongside other math datasets
85
+
86
+ ## License
87
+
88
+ open-data-attribution-training-disclosure-license-odatl-1.0