Datasets:
Update README.md
Browse files
README.md
CHANGED
@@ -7,63 +7,6 @@ size_categories:
|
|
7 |
task_categories:
|
8 |
- text-generation
|
9 |
- question-answering
|
10 |
-
configs:
|
11 |
-
- config_name: calcx-splits
|
12 |
-
data_files:
|
13 |
-
- split: train
|
14 |
-
path: calcx-splits/train-*
|
15 |
-
- split: validation
|
16 |
-
path: calcx-splits/validation-*
|
17 |
-
- split: test
|
18 |
-
path: calcx-splits/test-*
|
19 |
-
- config_name: original-splits
|
20 |
-
data_files:
|
21 |
-
- split: train
|
22 |
-
path: original-splits/train-*
|
23 |
-
- split: test
|
24 |
-
path: original-splits/test-*
|
25 |
-
dataset_info:
|
26 |
-
- config_name: calcx-splits
|
27 |
-
features:
|
28 |
-
- name: question
|
29 |
-
dtype: string
|
30 |
-
- name: chain
|
31 |
-
dtype: string
|
32 |
-
- name: result
|
33 |
-
dtype: string
|
34 |
-
- name: result_float
|
35 |
-
dtype: float64
|
36 |
-
splits:
|
37 |
-
- name: train
|
38 |
-
num_bytes: 5176237.798340693
|
39 |
-
num_examples: 7273
|
40 |
-
- name: validation
|
41 |
-
num_bytes: 142341.20165930683
|
42 |
-
num_examples: 200
|
43 |
-
- name: test
|
44 |
-
num_bytes: 957406
|
45 |
-
num_examples: 1319
|
46 |
-
download_size: 2958009
|
47 |
-
dataset_size: 6275985.0
|
48 |
-
- config_name: original-splits
|
49 |
-
features:
|
50 |
-
- name: question
|
51 |
-
dtype: string
|
52 |
-
- name: chain
|
53 |
-
dtype: string
|
54 |
-
- name: result
|
55 |
-
dtype: string
|
56 |
-
- name: result_float
|
57 |
-
dtype: float64
|
58 |
-
splits:
|
59 |
-
- name: train
|
60 |
-
num_bytes: 5318579
|
61 |
-
num_examples: 7473
|
62 |
-
- name: test
|
63 |
-
num_bytes: 957406
|
64 |
-
num_examples: 1319
|
65 |
-
download_size: 2949137
|
66 |
-
dataset_size: 6275985
|
67 |
---
|
68 |
# Dataset Card for "Calc-gsm8k"
|
69 |
|
@@ -93,7 +36,11 @@ into a simple html-like language that BeautifulSoup can parse.
|
|
93 |
|
94 |
Content corresponds to the original gsm8k dataset.
|
95 |
|
96 |
-
In this version, we created validation set by sampling 200 random examples from the original train split.
|
|
|
|
|
|
|
|
|
97 |
|
98 |
See [gsm8k HF dataset](https://huggingface.co/datasets/gsm8k) and [official repository](https://github.com/openai/grade-school-math) for more info.
|
99 |
|
|
|
7 |
task_categories:
|
8 |
- text-generation
|
9 |
- question-answering
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
10 |
---
|
11 |
# Dataset Card for "Calc-gsm8k"
|
12 |
|
|
|
36 |
|
37 |
Content corresponds to the original gsm8k dataset.
|
38 |
|
39 |
+
In this version, we created validation set by sampling 200 random examples from the original train split. The original data splits can be downloaded using:
|
40 |
+
|
41 |
+
```
|
42 |
+
datasets.load_dataset("MU-NLPC/Calc-gsm8k", "original-splits")
|
43 |
+
```
|
44 |
|
45 |
See [gsm8k HF dataset](https://huggingface.co/datasets/gsm8k) and [official repository](https://github.com/openai/grade-school-math) for more info.
|
46 |
|