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

Method Apply

lingvo/core/optimizer.py:99–160  ·  view source on GitHub ↗

Applies the gradient to the variable. Args: lr: A scalar or a callable that returns the learning rate. var_grad: A `.NestedMap` of (var, grad) pairs. weight: A scalar tensor that stores the weight of the loss and gradient. In most cases, weight is the number of trainin

(self, lr, var_grad, weight=None)

Source from the content-addressed store, hash-verified

source not stored for this graph (policy: none)

Callers 10

_ApplyAndResetMethod · 0.45
_ApplyAndResetMethod · 0.45
_FpropBpropMethod · 0.45
_Apply1Method · 0.45
_Apply2Method · 0.45
__init__Method · 0.45
_BPropGenTrainOpsMethod · 0.45
testAccumulatorMethod · 0.45
_testLearnerMethod · 0.45

Calls 4

GetOptimizerMethod · 0.95
VarReuseForSlotVarsMethod · 0.95
AddSummaryMethod · 0.95
GetLrValueFunction · 0.85

Tested by 6

_FpropBpropMethod · 0.36
_Apply1Method · 0.36
_Apply2Method · 0.36
testAccumulatorMethod · 0.36
_testLearnerMethod · 0.36