Update README.md

One-to-One Korean-English Parallel Sentence Dataset
This dataset contains a collection of parallel sentences in Korean and English. It has been compiled from multiple sources to support training and evaluation of machine translation models.

Dataset Overview
Size: 74,000 sentence pairs (Korean-English)
Columns:
kor_sent: Korean sentence
eng_sent: English sentence
Purpose: The dataset is designed to help train machine translation models for Korean to English and vice versa.
Dataset Sources
This dataset is constructed from multiple publicly available parallel corpora:

PAWS-X (Korean & English): The PAWS-X dataset is designed for cross-lingual sentence pair classification and includes high-quality pairs of sentences in various languages, including Korean and English.

OPUS-100 (Korean-English): This dataset is part of the OPUS collection of multilingual parallel corpora and is used for machine translation research.

Sentence-Transformer Parallel En-Ko with Similarity: A parallel corpus used for training models in sentence embedding and similarity tasks, focusing on English and Korean.

IWSLT2017 (Korean-English): The IWSLT 2017 dataset provides high-quality sentence pairs collected from spoken language datasets used for translation and interpretation research.

Each dataset brings a unique style and variation of parallel sentences, making this collection highly diverse and applicable for various NLP and machine translation tasks.

File Format
The dataset is provided as a CSV file, where each row contains:

A sentence in Korean in the kor_sent column.
The corresponding English sentence in the eng_sent column.

Training Models
This dataset is ideal for training machine translation models, such as those using mBART-50, BERT, or any other multilingual model. You can use this data for training, fine-tuning, or evaluating a translation model that converts between English and Korean sentences.

Data Quality
Length of Sentences: The dataset contains a mix of short and long sentences, making it useful for various training purposes.
Text Cleaning: Text has been cleaned to remove unnecessary characters and formatting.
Language: The dataset focuses on conversational and general text, so it may not contain specialized terminology.
Ethical Considerations
While the data has been sourced from publicly available datasets, it’s important to recognize that:

The dataset should not be used to generate harmful, abusive, or unethical content.
Use of this dataset must comply with the terms and conditions of the original dataset

README.md

@@ -1,28 +1,36 @@
----
-license: apache-2.0
-configs:
-- config_name: default
-  data_files:
-  - split: train
-    path: data/train-*
-dataset_info:
-  features:
-  - name: kor_sent
-    dtype: string
-  - name: eng_sent
-    dtype: string
-  - name: source
-    dtype: string
-  - name: similarity
-    dtype: float64
-  - name: from
-    dtype: string
-  - name: __index_level_0__
-    dtype: float64
-  splits:
-  - name: train
-    num_bytes: 784539402
-    num_examples: 3332436
-  download_size: 374217193
-  dataset_size: 784539402
----
+---
+license: apache-2.0
+configs:
+- config_name: default
+  data_files:
+  - split: train
+    path: data/train-*
+dataset_info:
+  features:
+  - name: kor_sent
+    dtype: string
+  - name: eng_sent
+    dtype: string
+  - name: source
+    dtype: string
+  - name: similarity
+    dtype: float64
+  - name: from
+    dtype: string
+  - name: __index_level_0__
+    dtype: float64
+  splits:
+  - name: train
+    num_bytes: 784539402
+    num_examples: 3332436
+  download_size: 374217193
+  dataset_size: 784539402
+task_categories:
+- translation
+language:
+- en
+- ko
+pretty_name: loki
+size_categories:
+- 1M<n<10M
+---