Upload folder using huggingface_hub
Browse files- .gitattributes +4 -0
- README.md +88 -3
- Tele-Data.py +51 -0
- arxiv/arxiv.jsonl +3 -0
- standard/standard.jsonl +3 -0
- web/web.jsonl +3 -0
- wiki/wiki.jsonl +3 -0
.gitattributes
CHANGED
@@ -56,3 +56,7 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
|
|
56 |
# Video files - compressed
|
57 |
*.mp4 filter=lfs diff=lfs merge=lfs -text
|
58 |
*.webm filter=lfs diff=lfs merge=lfs -text
|
|
|
|
|
|
|
|
|
|
56 |
# Video files - compressed
|
57 |
*.mp4 filter=lfs diff=lfs merge=lfs -text
|
58 |
*.webm filter=lfs diff=lfs merge=lfs -text
|
59 |
+
arxiv/arxiv.jsonl filter=lfs diff=lfs merge=lfs -text
|
60 |
+
standard/standard.jsonl filter=lfs diff=lfs merge=lfs -text
|
61 |
+
web/web.jsonl filter=lfs diff=lfs merge=lfs -text
|
62 |
+
wiki/wiki.jsonl filter=lfs diff=lfs merge=lfs -text
|
README.md
CHANGED
@@ -1,3 +1,88 @@
|
|
1 |
-
---
|
2 |
-
license: mit
|
3 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
license: mit
|
3 |
+
language:
|
4 |
+
- en
|
5 |
+
tags:
|
6 |
+
- telecom
|
7 |
+
task_categories:
|
8 |
+
- text-generation
|
9 |
+
configs:
|
10 |
+
- config_name: default
|
11 |
+
data_files:
|
12 |
+
- split: data
|
13 |
+
path: Tele-Eval.jsonl
|
14 |
+
---
|
15 |
+
|
16 |
+
# Tele-Data
|
17 |
+
|
18 |
+
## Dataset Summary
|
19 |
+
|
20 |
+
Tele-Data is a comprehensive dataset of telecommunications material that revolves around four categories of sources: (1) scientific papers from arXiv, (2) 3GPP standards, (3) Wikipedia articles related to telecommunications, and (4) telecommunications-related websites extracted from Common Crawl dumps.
|
21 |
+
|
22 |
+
LLM-based filtering was used to identify the relevant material from these sources, which then underwent extensive cleaning, format unification, and equation material standardization. The dataset consists of approximately 2.5 billion tokens, making it ideal for continually pretraining language models to adapt them to the telecommunications domain.
|
23 |
+
|
24 |
+
|
25 |
+
## Dataset Structure
|
26 |
+
|
27 |
+
### Data Fields
|
28 |
+
|
29 |
+
The data fields are as follows:
|
30 |
+
|
31 |
+
* `ID`: Provides a unique identifier for each data sample.
|
32 |
+
* `Category`: Identifies the category of the sample.
|
33 |
+
* `Content`: Includes the full text of the data sample.
|
34 |
+
* `Metadata`: Includes a JSON object, cast as a string, with information relevant to each data sample, which varies depending on the category.
|
35 |
+
|
36 |
+
### Data Instances
|
37 |
+
|
38 |
+
An example of Tele-Data looks as follows:
|
39 |
+
|
40 |
+
|
41 |
+
```json
|
42 |
+
{
|
43 |
+
"ID": "standard_2413",
|
44 |
+
"Category": "standard",
|
45 |
+
"Content": "3rd Generation Partnership Project; \n Technical Specification Group Core Network and Terminals;\n Interworking between the Public Land Mobile Network (PLMN)\n supporting packet based services with\n Wireless Local Area Network (WLAN) Access and\n Packet Data Networks (PDN)\n (Release 12)\n Foreword\n This Technical Specification (TS) has been produced...",
|
46 |
+
"Metadata":
|
47 |
+
"Series": "29",
|
48 |
+
"Release": "12",
|
49 |
+
"File_name": "29161-c00"
|
50 |
+
}
|
51 |
+
```
|
52 |
+
|
53 |
+
## Sample Code
|
54 |
+
|
55 |
+
Below, we share a code snippet on how to get quickly started with using the dataset. First, make sure to `pip install datasets`, then copy the snippet below and adapt it to your usecase.
|
56 |
+
|
57 |
+
#### Using the whole dataset
|
58 |
+
|
59 |
+
```python
|
60 |
+
from datasets import load_dataset
|
61 |
+
|
62 |
+
Tele_Data = load_dataset("AliMaatouk/Tele-Data")
|
63 |
+
data_sample = Tele_Data['train'][0]
|
64 |
+
print(f"ID: {data_sample['id']}\nCategory: {data_sample['category']} \nContent: {data_sample['content']}")
|
65 |
+
for key, value in json.loads(data_sample['metadata']).items():
|
66 |
+
print(f"{key}: {value}")
|
67 |
+
```
|
68 |
+
|
69 |
+
#### Using a subset of the dataset
|
70 |
+
|
71 |
+
```python
|
72 |
+
from datasets import load_dataset
|
73 |
+
|
74 |
+
Tele_Data = load_dataset("AliMaatouk/Tele-Data", name="standard")
|
75 |
+
data_sample = Tele_Data['train'][0]
|
76 |
+
print(f"ID: {data_sample['id']}\nCategory: {data_sample['category']} \nContent: {data_sample['content']}")
|
77 |
+
for key, value in json.loads(data_sample['metadata']).items():
|
78 |
+
print(f"{key}: {value}")
|
79 |
+
```
|
80 |
+
|
81 |
+
## Citation
|
82 |
+
|
83 |
+
You can find the paper with all details about the dataset at https://arxiv.org/abs/xxx. Please cite it as follows:
|
84 |
+
|
85 |
+
```
|
86 |
+
@misc{xx
|
87 |
+
}
|
88 |
+
```
|
Tele-Data.py
ADDED
@@ -0,0 +1,51 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import json
|
2 |
+
import datasets
|
3 |
+
|
4 |
+
class TeleData(datasets.GeneratorBasedBuilder):
|
5 |
+
"""Tele-Data dataset with multiple subsets: arxiv, standard, web, and wiki"""
|
6 |
+
|
7 |
+
BUILDER_CONFIGS = [
|
8 |
+
datasets.BuilderConfig(name="arxiv", version=datasets.Version("1.0.0"), description="ArXiv data"),
|
9 |
+
datasets.BuilderConfig(name="standard", version=datasets.Version("1.0.0"), description="Standard data"),
|
10 |
+
datasets.BuilderConfig(name="web", version=datasets.Version("1.0.0"), description="Web data"),
|
11 |
+
datasets.BuilderConfig(name="wiki", version=datasets.Version("1.0.0"), description="Wiki data"),
|
12 |
+
datasets.BuilderConfig(name="full", version=datasets.Version("1.0.0"), description="Full dataset"),
|
13 |
+
]
|
14 |
+
|
15 |
+
DEFAULT_CONFIG_NAME = "full"
|
16 |
+
|
17 |
+
def _info(self):
|
18 |
+
features = datasets.Features({
|
19 |
+
"id": datasets.Value("string"),
|
20 |
+
"category": datasets.Value("string"),
|
21 |
+
"content": datasets.Value("string"),
|
22 |
+
"metadata": datasets.Value("string"),
|
23 |
+
})
|
24 |
+
return datasets.DatasetInfo(features=features)
|
25 |
+
|
26 |
+
def _split_generators(self, dl_manager):
|
27 |
+
if self.config.name == "full":
|
28 |
+
urls = [f"{name}/{name}.jsonl" for name in ["arxiv", "standard", "web", "wiki"]]
|
29 |
+
else:
|
30 |
+
urls = [f"{self.config.name}/{self.config.name}.jsonl"]
|
31 |
+
|
32 |
+
data_files = dl_manager.download_and_extract(urls)
|
33 |
+
|
34 |
+
return [
|
35 |
+
datasets.SplitGenerator(
|
36 |
+
name=datasets.Split.TRAIN,
|
37 |
+
gen_kwargs={"filepaths": data_files if isinstance(data_files, list) else [data_files]},
|
38 |
+
)
|
39 |
+
]
|
40 |
+
|
41 |
+
def _generate_examples(self, filepaths):
|
42 |
+
for filepath in filepaths:
|
43 |
+
with open(filepath, "r", encoding="utf-8") as f:
|
44 |
+
for id_, line in enumerate(f):
|
45 |
+
data = json.loads(line)
|
46 |
+
yield id_, {
|
47 |
+
"id": data["id"],
|
48 |
+
"category": data["category"],
|
49 |
+
"content": data["content"],
|
50 |
+
"metadata": json.dumps(data["metadata"]),
|
51 |
+
}
|
arxiv/arxiv.jsonl
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:1f2d03a5cc1d8b0a13964d1a47e2456cd19813beddd8711e47d2e24102a824e7
|
3 |
+
size 4309790362
|
standard/standard.jsonl
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:b9560c7a62612c4724e986fc7443019d886a158a961ddc4c3e521a6650486fa9
|
3 |
+
size 350404101
|
web/web.jsonl
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:1a11dede469b0f5568298a8172ca451382c130e1e8db215a5158d3681598759e
|
3 |
+
size 7266835533
|
wiki/wiki.jsonl
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:0a5bb8acd97ad37a884a845a705206980039d337e49ab3ea5c278925b2e2977e
|
3 |
+
size 129518949
|