Update README.md
Browse files
README.md
CHANGED
@@ -25,3 +25,152 @@ configs:
|
|
25 |
- split: test
|
26 |
path: data/test-*
|
27 |
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
25 |
- split: test
|
26 |
path: data/test-*
|
27 |
---
|
28 |
+
|
29 |
+
# Random ASCII Dataset
|
30 |
+
|
31 |
+
This dataset contains random sequences of ASCII characters, with "train," "validation," and "test" splits, designed to simulate text-like structures using all printable ASCII characters. Each sequence consists of pseudo-randomly generated "words" of various lengths, separated by spaces to mimic natural language text.
|
32 |
+
|
33 |
+
## Dataset Details
|
34 |
+
|
35 |
+
- **Splits**: Train, Validation, and Test
|
36 |
+
- **Number of sequences**:
|
37 |
+
- Train: 5000 sequences
|
38 |
+
- Validation: 5000 sequences
|
39 |
+
- Test: 5000 sequences
|
40 |
+
- **Sequence length**: 512 characters per sequence
|
41 |
+
- **Character pool**: All printable ASCII characters, including letters, digits, punctuation, and whitespace.
|
42 |
+
|
43 |
+
## Sample Usage
|
44 |
+
|
45 |
+
To load this dataset in Python, you can use the Hugging Face `datasets` library:
|
46 |
+
|
47 |
+
```python
|
48 |
+
from datasets import load_dataset
|
49 |
+
|
50 |
+
# Load the dataset
|
51 |
+
dataset = load_dataset("brando/random-ascii-dataset")
|
52 |
+
|
53 |
+
# Access the train, validation, and test splits
|
54 |
+
train_data = dataset["train"]
|
55 |
+
val_data = dataset["validation"]
|
56 |
+
test_data = dataset["test"]
|
57 |
+
|
58 |
+
# Print a sample
|
59 |
+
print(train_data[0]["text"])
|
60 |
+
```
|
61 |
+
|
62 |
+
# Example Data
|
63 |
+
Below are examples of random sequences generated in this dataset:
|
64 |
+
```python
|
65 |
+
#Example 1:
|
66 |
+
"!Q4$^V3w L@#12 Vd&$%4B+ (k#yFw! [7*9z"
|
67 |
+
|
68 |
+
#Example 2:
|
69 |
+
"T^&3xR f$xH&ty ^23M* qW@# Lm5&"
|
70 |
+
|
71 |
+
#Example 3:
|
72 |
+
"b7$W %&6Zn!!R xT&8N z#G m93T +%^0"
|
73 |
+
```
|
74 |
+
|
75 |
+
# License
|
76 |
+
|
77 |
+
This dataset is released under the Apache License 2.0. You are free to use, modify, and distribute this dataset under the terms of the Apache License.
|
78 |
+
|
79 |
+
# Citation
|
80 |
+
|
81 |
+
```bibtex
|
82 |
+
@misc{miranda2021ultimateutils,
|
83 |
+
title={Ultimate Utils - the Ultimate Utils Library for Machine Learning and Artificial Intelligence},
|
84 |
+
author={Brando Miranda},
|
85 |
+
year={2021},
|
86 |
+
url={https://github.com/brando90/ultimate-utils},
|
87 |
+
note={Available at: \url{https://www.ideals.illinois.edu/handle/2142/112797}},
|
88 |
+
abstract={Ultimate Utils is a comprehensive library providing utility functions and tools to facilitate efficient machine learning and AI research, including efficient tensor manipulations and gradient handling with methods such as `detach()` for creating gradient-free tensors.}
|
89 |
+
}
|
90 |
+
```
|
91 |
+
|
92 |
+
# Code that generate it
|
93 |
+
|
94 |
+
```python
|
95 |
+
#ref: https://chatgpt.com/c/671ff56a-563c-8001-afd5-94632fe63d67
|
96 |
+
import os
|
97 |
+
import random
|
98 |
+
import string
|
99 |
+
from huggingface_hub import login
|
100 |
+
from datasets import Dataset, DatasetDict
|
101 |
+
|
102 |
+
# Function to load the Hugging Face API token from a file
|
103 |
+
def load_token(file_path: str) -> str:
|
104 |
+
"""Load API token from a specified file path."""
|
105 |
+
with open(os.path.expanduser(file_path)) as f:
|
106 |
+
return f.read().strip()
|
107 |
+
|
108 |
+
# Function to log in to Hugging Face using a token
|
109 |
+
def login_to_huggingface(token: str) -> None:
|
110 |
+
"""Authenticate with Hugging Face Hub."""
|
111 |
+
login(token=token)
|
112 |
+
print("Login successful")
|
113 |
+
|
114 |
+
# Function to generate a random word of a given length
|
115 |
+
def generate_random_word(length: int, character_pool: str) -> str:
|
116 |
+
"""Generate a random word of specified length from a character pool."""
|
117 |
+
return "".join(random.choice(character_pool) for _ in range(length))
|
118 |
+
|
119 |
+
# Function to generate a single random sentence with "words" of random lengths
|
120 |
+
def generate_random_sentence(sequence_length: int, character_pool: str) -> str:
|
121 |
+
"""Generate a random sentence of approximately sequence_length characters."""
|
122 |
+
words = [
|
123 |
+
generate_random_word(random.randint(3, 10), character_pool)
|
124 |
+
for _ in range(sequence_length // 10) # Estimate number of words to fit length
|
125 |
+
]
|
126 |
+
sentence = " ".join(words)
|
127 |
+
# print(f"Generated sentence length: {len(sentence)}\a") # Print length and sound alert
|
128 |
+
return sentence
|
129 |
+
|
130 |
+
# Function to create a dataset of random "sentences"
|
131 |
+
def create_random_text_dataset(num_sequences: int, sequence_length: int, character_pool: str) -> Dataset:
|
132 |
+
"""Create a dataset with random text sequences."""
|
133 |
+
data = {
|
134 |
+
"text": [generate_random_sentence(sequence_length, character_pool) for _ in range(num_sequences)]
|
135 |
+
}
|
136 |
+
return Dataset.from_dict(data)
|
137 |
+
|
138 |
+
# Main function to generate, inspect, and upload dataset with train, validation, and test splits
|
139 |
+
def main() -> None:
|
140 |
+
# Step 1: Load token and log in
|
141 |
+
key_file_path: str = "/lfs/skampere1/0/brando9/keys/brandos_hf_token.txt"
|
142 |
+
token: str = load_token(key_file_path)
|
143 |
+
login_to_huggingface(token)
|
144 |
+
|
145 |
+
# Step 2: Dataset parameters
|
146 |
+
num_sequences_train: int = 5000
|
147 |
+
num_sequences_val: int = 5000
|
148 |
+
num_sequences_test: int = 5000
|
149 |
+
sequence_length: int = 512
|
150 |
+
character_pool: str = string.printable # All ASCII characters (letters, digits, punctuation, whitespace)
|
151 |
+
|
152 |
+
# Step 3: Create datasets for each split
|
153 |
+
train_dataset = create_random_text_dataset(num_sequences_train, sequence_length, character_pool)
|
154 |
+
val_dataset = create_random_text_dataset(num_sequences_val, sequence_length, character_pool)
|
155 |
+
test_dataset = create_random_text_dataset(num_sequences_test, sequence_length, character_pool)
|
156 |
+
|
157 |
+
# Step 4: Combine into a DatasetDict with train, validation, and test splits
|
158 |
+
dataset_dict = DatasetDict({
|
159 |
+
"train": train_dataset,
|
160 |
+
"validation": val_dataset,
|
161 |
+
"test": test_dataset
|
162 |
+
})
|
163 |
+
|
164 |
+
# Step 5: Print a sample of the train dataset for verification
|
165 |
+
print("Sample of train dataset:", train_dataset[:5])
|
166 |
+
|
167 |
+
# Step 6: Push the dataset to Hugging Face Hub
|
168 |
+
dataset_name: str = "brando/random-ascii-dataset"
|
169 |
+
dataset_dict.push_to_hub(dataset_name)
|
170 |
+
print(f"Dataset uploaded to https://huggingface.co/datasets/{dataset_name}")
|
171 |
+
|
172 |
+
# Run the main function
|
173 |
+
if __name__ == "__main__":
|
174 |
+
main()
|
175 |
+
|
176 |
+
```
|