An image classification project leveraging the Transformer architecture from Large Language Models (LLMs).

• Multiple LLM Architectures: BERT, ELECTRA, GPT-2, RoBERTa, T5
• Diverse Feature Extractors: VGG, ResNet, DenseNet, EfficientNet, MobileNet, ConvNeXt
• Multi-Scale Feature Extraction: Extract image features at various scales for enhanced performance

├── models/
│   ├── bert.py                # BERT-based image classification model
│   ├── electra.py             # ELECTRA-based image classification model
│   ├── gpt.py                 # GPT-2-based image classification model
│   ├── roberta.py             # RoBERTa-based image classification model
│   ├── t5.py                  # T5-based image classification model
│   └── feature_extractor.py  # CNN-based feature extractors

1. Extract image features using CNN-based Feature Extractors
2. Project extracted features into Transformer embedding dimensions
3. Process features through LLM Transformer architecture
4. Perform final classification through MLP layers