Merge pull request ChiaXinLiang#28 from weian312/dev

remove duplicate

MLLM_latex/chapter4/chapter4.tex

@@ -86,9 +86,6 @@ \subsection{Multitask Fine-Tuning}
 
 In some cases, models are fine-tuned on multiple tasks simultaneously, a technique known as \textbf{multitask learning}. This helps the model generalize better across various related tasks. For example, a model might be fine-tuned on both image captioning and visual question answering datasets simultaneously, allowing it to perform well in both scenarios.
 
-\subsection{Cross-Modal Tasks}
-
-Fine-tuning is essential for tasks that require the model to reason across modalities, such as cross-modal retrieval or referring expression comprehension (where the model must identify specific objects in an image based on a text description). The goal is to align the visual and textual representations effectively during this phase.
 
 \subsection{Cross-Modal Tasks}