Update README.md

README.md

@@ -18,6 +18,8 @@ tags:
 - rslora
 - liger
 - sociology
+datasets:
+- agentlans/finewebedu-multiple-choice
 ---
 # Qwen2.5-1.5B-Instruct-Multiple-Choice-Maker
 
@@ -30,11 +32,11 @@ Special thanks to [mradermacher](https://huggingface.co/mradermacher) for quanti
 
 ### Training Details
 - **Data Source**:
-  - The training dataset was generated from open-source sociology textbooks using a custom prompt powered by the [agentlans/Llama3.1-LexiHermes-SuperStorm](https://huggingface.co/agentlans/Llama3.1-LexiHermes-SuperStorm) model. The dataset contains 3,739 rows. Due to licensing restrictions, the dataset is not provided here.
-  - Additional finetuning on [agentlans/finewebedu-multiple-choice](https://huggingface.co/datasets/agentlans/finewebedu-multiple-choice) dataset
+  - The training dataset was generated from open-source sociology textbooks using a custom prompt powered by the [agentlans/Llama3.1-LexiHermes-SuperStorm](https://huggingface.co/agentlans/Llama3.1-LexiHermes-SuperStorm) model. The dataset contains 3739 rows. Due to licensing restrictions, the dataset is not provided here.
+  - Additional finetuning on [agentlans/finewebedu-multiple-choice](https://huggingface.co/datasets/agentlans/finewebedu-multiple-choice) dataset.
 - **Training Method**:
   - Fine-tuning was conducted using [LLaMA-Factory](https://github.com/hiyouga/LLaMA-Factory) with rank 16 LoRA, alpha = 32, and rslora, leveraging the Liger kernel.
-  - The additional finetuning was done with the same settings but with 0.2 LoRA dropout
+  - The additional finetuning was done with the same settings but with 0.2 LoRA dropout.
 
 ### Potential Applications
 - **Education**: Automate the creation of multiple-choice questions for exams or quizzes.