Types & classes — github.com/QiangAIResearcher/sahp_repo

CodeHub MCPcopy Create free account

hub / github.com/QiangAIResearcher/sahp_repo / types & classes

Types & classes15 in github.com/QiangAIResearcher/sahp_repo

⨍ Functions 54 ◇ Types & classes 15

↓ 5 callersClassMaskBatch

object for holding a batch of data with mask during training

train_functions/train_sahp.py:37

↓ 3 callersClassNoamOpt

Optim wrapper that implements rate.

utils/atten_optimizer.py:3

↓ 2 callersClassGELU

Paper Section 3.4, last paragraph notice that BERT used the GELU instead of RELU

models/utils/gelu.py:6

↓ 2 callersClassSublayerConnection

A residual connection followed by a layer norm. Note for code simplicity the norm is first as opposed to last.

models/utils/sublayer.py:5

↓ 1 callersClassAttention

Compute 'Scaled Dot Product Attention

models/attention/single.py:8

↓ 1 callersClassBiasedPositionalEmbedding

models/embedding/position.py:35

↓ 1 callersClassLayerNorm

Construct a layernorm module (See citation for details).

models/utils/layer_norm.py:5

↓ 1 callersClassMultiHeadedAttention

Take in models size and number of heads.

models/attention/multi_head.py:5

↓ 1 callersClassPositionalEmbedding

models/embedding/position.py:6

↓ 1 callersClassPositionwiseFeedForward

Implements FFN equation.

models/utils/feed_forward.py:5

↓ 1 callersClassSAHP

Generic N layer attentive Hawkes with masking

models/sahp.py:23

↓ 1 callersClassTypeEmbedding

models/embedding/event_type.py:4

ClassEventEmbedding

Event Embedding which is consisted with under features 1. TypeEmbedding : normal embedding matrix 2. PositionalEmbedding : adding

models/embedding/event_embedding.py:6

models/sahp.py:310

ClassSeqGenerator

models/base.py:7