![]() |
MyCaffe
1.12.2.41
Deep learning software for Windows C# programmers.
|
The MyCaffe.layers.gpt namespace contains all GPT related layers. More...
Classes | |
| class | CausalSelfAttentionLayer |
| The CausalSelfAttention provides a vanilla multi-head self-attention layer with projection at the end. More... | |
| class | CausalSelfAttentionLayer2 |
| The CausalSelfAttention provides a vanilla multi-head self-attention layer with projection at the end. More... | |
| class | CustomListData |
| The CustomData supports external data input via an external Assembly DLL that supports the ICustomTokenInput interface. More... | |
| class | GeluLayer |
| The GeluLayer implements the New GELU activation function currently in Google BERT repo (same as OpenAI GPT) More... | |
| interface | ICustomTokenInput |
| The ICustomTokenInput interface specifies the interface that all custom token inputs implement. More... | |
| class | InputData |
| The InputData is an abstract class used to get training data and tokenize input data. More... | |
| interface | IVocabulary |
| The IVocabulary interface specifies the interface that all Vocabularies implement. More... | |
| class | LayerFactory |
| The LayerFactor is responsible for creating all layers implemented in the MyCaffe.layers.gpt namespace. More... | |
| class | LayerNormLayer |
| The LayerNormalizationLayer performs layer normalization similar to the PyTorch LayerNorm layer. More... | |
| class | MultiheadAttentionLayer |
| The MultiheadAttention provides a vanilla multi-head layer. More... | |
| class | NLLLossLayer |
| Computes the nll loss for a one-of-many classification task, passing real-valued predictions (from a softmax or logsoftmax) to get a probability distribution over classes. More... | |
| class | PositionalEncodingLayer |
| The PositionalEncodingLayer is a neuron layer that adds positional encoding to the input. More... | |
| class | TextInputData |
| The TextInputData manages character data read in from a text file. Data is tokenized into indexes that reference each character within the vocabulary. More... | |
| class | TextListData |
| The TextListData manages parallel lists of data where the first list contains the encoder input data and the second the decoder input/target data. More... | |
| class | TokenizedDataLayer |
| The TokenizedDataLayer loads and tokenizes data for a transformer model where data is loaded in the form: data, pos, target(optional) More... | |
| class | TokenizedDataPairsLayer |
| The TokenizedDataPairsLayer loads and tokenizes data for a transformer model where data is loaded in the form: data, pos, target(optional) More... | |
| class | TransformerBlockLayer |
| The TransformerBlock provides a generic transformer block More... | |
| class | VocabularyCharacter |
| The VocabularyCharacters class manages the data vocabulary of characters. More... | |
| class | VocabularySentencePiece |
| The VocabularySentencePieces class manages the data vocabulary of words. More... | |
| class | VocabularyWord |
| The VocabularyWords class manages the data vocabulary of words. More... | |
The MyCaffe.layers.gpt namespace contains all GPT related layers.
WORK IN PROGRESS.