Implementation reorganized from d2l
[Click to expand]
- Transformer (transformer.py)
- TransformerEncoder (transformer.py)
- Embedding
- PositionalEncoding (embeddings.py)
- EncoderBlock (Nx) (layers.py)
- MultiHeadAttention (Self) (sublayers.py)
- Add & Norm (sublayers.py)
- PositionWiseFFN (sublayers.py)
- Add & Norm
- TransformerDecoder (transformer.py)
- Embedding
- PositionalEncoding
- DecoderBlock (Nx) (layers.py)
- MultiHeadAttention (Self)
- Add & Norm
- MultiHeadAttention (Encoder-Decoder)
- Add & Norm
- PositionWiseFFN
- Add & Norm
- Linear
- TransformerEncoder (transformer.py)
./experiments/scripts/demo.sh