forced dataset sizes to be 3 for testing

src/deepeval/base_task.py

@@ -188,7 +188,7 @@ class BaseTask(ABC):
         print("Loading dataset from Hugging Face.")
         dataset= load_dataset(self.dataset_repo, token=HF_TOKEN, split="train")
         print("Dataset loaded.")
-        return dataset
+        return dataset.select(range(min(3, len(dataset))))
 
     @abstractmethod
     def evaluate(self):