Update README.md

README.md

@@ -25,7 +25,22 @@ The **Twirling Mizo News Dataset** is a collection of 6,731 news articles writte
       - *hmarchhak*
 
 - **Largest Category:** *tualchhung* (1,686 articles)
+- **Training Set (80%)**: This set contains 80% of the data for each category and will be used for training machine learning models.
+- **Testing Set (20%)**: This set contains the remaining 20% of the data for each category and can be used for evaluating the performance of the models.
 
+### Example Split
+
+For each category, the dataset is split as follows:
+
+1. **Category**: "tualchhung"
+   - **Training Set**: 80% of articles in this category.
+   - **Testing Set**: 20% of articles in this category.
+   
+2. **Category**: "khawvel"
+   - **Training Set**: 80% of articles in this category.
+   - **Testing Set**: 20% of articles in this category.
+
+This pattern is applied to all categories in the dataset, ensuring that the splits are balanced and representative of each category.
 
 ## How to use
 
@@ -35,17 +50,16 @@ The `datasets` library allows you to load and pre-process your dataset in pure P
 from datasets import load_dataset
 
 twirling_mizo_news_train = load_dataset("andrewbawitlung/twirling_mizo_news", split="train")
+twirling_mizo_news_test = load_dataset("andrewbawitlung/twirling_mizo_news", split="test")
 
 ```
 display 3 random indices
 ```python
 import random
-random_indices = random.sample(range(len(twirling_mizo_news_train)), 3)
-
-for idx in random_indices:
-    random_article = twirling_mizo_news_train[idx]
-    print(f"Category: {random_article['Category']}")
-    print(f"Article: {random_article['Article']}\n")
+for split, dataset in [("train", twirling_mizo_news_train), ("test", twirling_mizo_news_test)]:
+    print(f"Random samples from the {split} dataset:")
+    for idx in random.sample(range(len(dataset)), 5):
+        print(f"Index: {idx}\n{dataset[idx]}\n{'-' * 50}")
 
 ```