Update subset description

README.md

@@ -175,16 +175,79 @@ Please refer to the license of the [NiuTrans/Classical-Modern](https://github.co
 If you use this dataset in your research, please cite the original sources:
 
 1. [NiuTrans/Classical-Modern](https://github.com/NiuTrans/Classical-Modern)
-2. Gemini Pro (for English translations)
+2. [Gemini Pro](https://arxiv.org/abs/2403.05530)
 
 **Dataset Structure**
 
-The dataset is a tab-separated text file with four columns:
+The dataset have four subsets, namely `default`, `instruct`, `instruct-augment`, `instruct-large`
+1. **Default**
 
-1. **document\_info**: The title or source information of the text
-2. **modern\_chinese**: The translation of the original Classical Chinese text into modern Chinese
-3. **classical\_chinese**: The original text in Classical Chinese (Wenyanwen)
-4. **english**: The English translation of the Classical Chinese text
+   **Features:**
+   - `info`: string
+   - `modern`: string
+   - `classical`: string
+   - `english`: string
+  
+   **Splits:**
+   - Train: 972,467 examples
+
+   **Size:**
+   - Dataset size: 366.9 MB
+   - Download size: 256.4 MB
+
+2. **Instruct**
+
+   **Features:**
+   - `info`: string
+   - `modern`: string
+   - `classical`: string
+   - `english`: string
+   - `text`: string
+   - `messages`: list of {`content`: string, `role`: string}
+
+   **Splits:**
+   - Train: 9,000 examples
+   - Test: 1,000 examples
+
+   **Size:**
+   - Dataset size: 10.98 MB
+   - Download size: 6.89 MB
+
+3. **Instruct-Augmented**
+
+   **Features:**
+   - `info`: string
+   - `modern`: string
+   - `classical`: string
+   - `english`: string
+   - `text`: string
+   - `messages`: list of {`content`: string, `role`: string}
+
+   **Splits:**
+   - Train: 9,000 examples
+   - Test: 1,000 examples
+
+   **Size:**
+   - Dataset size: 12.4 MB
+   - Download size: 7.56 MB
+   
+4. **Instruct-Large**
+
+   **Features:**
+   - `info`: string
+   - `modern`: string
+   - `classical`: string
+   - `english`: string
+   - `text`: string
+   - `messages`: list of {`content`: string, `role`: string}
+
+   **Splits:**
+   - Train: 875,214 examples
+   - Test: 97,246 examples
+
+   **Size:**
+   - Dataset size: 1.07 GB
+   - Download size: 673.41 MB
 
 Here is an example of a dataset entry:
 
@@ -197,15 +260,14 @@ Here is an example of a dataset entry:
 The dataset contains approximately 972k entries.
 
 **Data Fields**
-
-- **document\_info**: A string representing the title or source information of the text.
-- **modern\_chinese**: A string containing the translation of the original Classical Chinese text into modern Chinese.
-- **classical\_chinese**: A string containing the original text in Classical Chinese (Wenyanwen).
-- **english**: A string containing the English translation of the Classical Chinese text.
-
-**Data Splits**
-
-There are no official data splits for this dataset. We recommend splitting the dataset into train, validation, and test sets at a ratio of 80:10:10.
+- `info`: A string representing the title or source information of the text.
+- `classical`: Classical Chinese (Wenyanwen) text corresponding to the modern text.
+- `modern`: A string containing the translation of the original Classical Chinese text into modern Chinese.
+- `english`: English translation of the Chinese text.
+- `text`: instruction/answer pair in string format
+- `messages`: instruction/answer pair in conversation format:
+  - `content`: String representing the content of a message.
+  - `role`: String representing the role associated with the message (e.g., sender, receiver).
 
 **Potential Bias**