MCPcopy Create free account
hub / github.com/DeepRec-AI/DeepRec / _apply_sparse

Method _apply_sparse

tensorflow/python/training/adadelta.py:107–119  ·  view source on GitHub ↗
(self, grad, var)

Source from the content-addressed store, hash-verified

105 use_locking=self._use_locking)
106
107 def _apply_sparse(self, grad, var):
108 accum = self.get_slot(var, "accum")
109 accum_update = self.get_slot(var, "accum_update")
110 return training_ops.sparse_apply_adadelta(
111 var,
112 accum,
113 accum_update,
114 math_ops.cast(self._lr_t, var.dtype.base_dtype),
115 math_ops.cast(self._rho_t, var.dtype.base_dtype),
116 math_ops.cast(self._epsilon_t, var.dtype.base_dtype),
117 grad.values,
118 grad.indices,
119 use_locking=self._use_locking)
120
121 def _resource_apply_sparse(self, grad, var, indices):
122 accum = self.get_slot(var, "accum")

Callers

nothing calls this directly

Calls 2

get_slotMethod · 0.45
castMethod · 0.45

Tested by

no test coverage detected