Text this: Mobile3ViT: An Improved Hybrid CNN‐Visual Transformer Model for Automatic Gastrointestinal Image Recognition