Loaded some domain-specific downstream image classification content moderation models, which is essentially the practice of monitoring and filtering user-generated content on platforms, based on SigLIP-2 Base Patch16 with newly initialized trainable parameters. 🥠