Project structure of: kyegomez/Gemini

• Gemini Multi-modal Transformer models for text/image generation and testing
  • example.py Python model optimized, text-image inputs, outputs text sequence.
  • gemini_torch Multi-modal Transformer models for EIA data
    • __init__.py Imports and adds to all list.
    • model.py Transformer, Gemini models, multi-layer, concatenation
    • tokenizer.py Multimodal SentencePiece tokenizer with modality, BOS, and EOS.
    • transformer.py Transformer model with customizable layers and dropout
    • utils.py Embeddings from Images and Audio.
  • pyproject.toml PyProject config: Poetry, Gemini-Torch, Python 3.9
  • README.md Multimodal transformer model for text/image generation
  • requirements.txt Python packages: zetascale, torch, einops, pytest, sentencepiece.
  • tests Test models, encoders, and tokenizers.
    • test_audio_embedder.py Test audio embedding function's behavior and output integrity.
    • test_gemini.py Gemini model test functions: parameters, forward pass, error handling.
    • test_img_encoder.py Tests ImageToTextEmbeddings model initialization and output shape.
    • test_img_to_transformer.py ImageToTextEmbeddings initialization and forward pass demonstrated.
    • test_tokenizer.py Test multi-modal sentence piece tokenizer functions.