MCPcopy Create free account
hub / github.com/tensorflow/lingvo / TransformerStack

Method TransformerStack

lingvo/core/self_attention_layer.py:353–372  ·  view source on GitHub ↗

Returns a stack of num_layers self-attention layers.

(self, name, num_layers=1, feed_forward_qdomain=None)

Source from the content-addressed store, hash-verified

source not stored for this graph (policy: none)

Callers 10

testTransformerStackMethod · 0.45
testTransformerStackV2Method · 0.45
testTransformerStackV2WithSimplifiedTransformerMethod · 0.45
testTransformerStackV2WithSimplifiedTransformerWithMQAMethod · 0.45
testTransformerLayerFPropMethod · 0.45
testTransformerStackAlternateLayersMethod · 0.45
testTransformerStackFPropWithPackedInputsMethod · 0.45
testTransparentTransformerStackTrainFPropMethod · 0.45
testTransparentTransformerStackEvalFPropMethod · 0.45
_EncoderParamsMethod · 0.45

Calls 3

_TransformerLayerBlockMethod · 0.95
_MaybeSplitMethod · 0.95
_SeqMethod · 0.45

Tested by 10

testTransformerStackMethod · 0.36
testTransformerStackV2Method · 0.36
testTransformerStackV2WithSimplifiedTransformerMethod · 0.36
testTransformerStackV2WithSimplifiedTransformerWithMQAMethod · 0.36
testTransformerLayerFPropMethod · 0.36
testTransformerStackAlternateLayersMethod · 0.36
testTransformerStackFPropWithPackedInputsMethod · 0.36
testTransparentTransformerStackTrainFPropMethod · 0.36
testTransparentTransformerStackEvalFPropMethod · 0.36
_EncoderParamsMethod · 0.36