Improved Fine-Tuning of Large Multimodal Models for Hateful Meme Detection