| /external/tensorflow/tensorflow/python/keras/optimizer_v2/ |
| D | adagrad.py | 107 def _prepare_local(self, var_device, var_dtype, apply_state): argument
108 super(Adagrad, self)._prepare_local(var_device, var_dtype, apply_state)
109 apply_state[(var_device, var_dtype)].update(dict(
111 neg_lr_t=-apply_state[(var_device, var_dtype)]['lr_t'],
147 def _resource_apply_dense(self, grad, var, apply_state=None): argument
149 coefficients = ((apply_state or {}).get((var_device, var_dtype))
161 def _resource_apply_sparse(self, grad, var, indices, apply_state=None): argument
163 coefficients = ((apply_state or {}).get((var_device, var_dtype))
|
| D | adadelta.py | 111 def _prepare_local(self, var_device, var_dtype, apply_state): argument
112 super(Adadelta, self)._prepare_local(var_device, var_dtype, apply_state)
113 apply_state[(var_device, var_dtype)].update(dict(
127 def _resource_apply_dense(self, grad, var, apply_state=None): argument
129 coefficients = ((apply_state or {}).get((var_device, var_dtype))
144 def _resource_apply_sparse(self, grad, var, indices, apply_state=None): argument
146 coefficients = ((apply_state or {}).get((var_device, var_dtype))
|
| D | adamax.py | 115 def _prepare_local(self, var_device, var_dtype, apply_state): argument
116 super(Adamax, self)._prepare_local(var_device, var_dtype, apply_state)
122 lr_t = apply_state[(var_device, var_dtype)]['lr_t']
124 apply_state[(var_device, var_dtype)].update(dict(
134 def _resource_apply_dense(self, grad, var, apply_state=None): argument
136 coefficients = ((apply_state or {}).get((var_device, var_dtype))
154 def _resource_apply_sparse(self, grad, var, indices, apply_state=None): argument
156 coefficients = ((apply_state or {}).get((var_device, var_dtype))
|
| D | gradient_descent.py | 103 def _prepare_local(self, var_device, var_dtype, apply_state): argument
104 super(SGD, self)._prepare_local(var_device, var_dtype, apply_state)
105 apply_state[(var_device, var_dtype)]["momentum"] = array_ops.identity(
108 def _resource_apply_dense(self, grad, var, apply_state=None): argument
110 coefficients = ((apply_state or {}).get((var_device, var_dtype))
140 def _resource_apply_sparse(self, grad, var, indices, apply_state=None): argument
143 coefficients = ((apply_state or {}).get((var_device, var_dtype))
|
| D | ftrl.py | 148 def _prepare_local(self, var_device, var_dtype, apply_state): argument
149 super(Ftrl, self)._prepare_local(var_device, var_dtype, apply_state)
150 apply_state[(var_device, var_dtype)].update(dict(
161 def _resource_apply_dense(self, grad, var, apply_state=None): argument
163 coefficients = ((apply_state or {}).get((var_device, var_dtype))
193 def _resource_apply_sparse(self, grad, var, indices, apply_state=None): argument
195 coefficients = ((apply_state or {}).get((var_device, var_dtype))
|
| D | adam.py | 161 def _prepare_local(self, var_device, var_dtype, apply_state): argument
162 super(Adam, self)._prepare_local(var_device, var_dtype, apply_state)
169 lr = (apply_state[(var_device, var_dtype)]['lr_t'] *
171 apply_state[(var_device, var_dtype)].update(dict(
192 def _resource_apply_dense(self, grad, var, apply_state=None): argument
194 coefficients = ((apply_state or {}).get((var_device, var_dtype))
229 def _resource_apply_sparse(self, grad, var, indices, apply_state=None): argument
231 coefficients = ((apply_state or {}).get((var_device, var_dtype))
|
| D | rmsprop.py | 132 def _prepare_local(self, var_device, var_dtype, apply_state): argument
133 super(RMSprop, self)._prepare_local(var_device, var_dtype, apply_state)
136 apply_state[(var_device, var_dtype)].update(dict(
137 neg_lr_t=-apply_state[(var_device, var_dtype)]["lr_t"],
144 def _resource_apply_dense(self, grad, var, apply_state=None): argument
146 coefficients = ((apply_state or {}).get((var_device, var_dtype))
190 def _resource_apply_sparse(self, grad, var, indices, apply_state=None): argument
192 coefficients = ((apply_state or {}).get((var_device, var_dtype))
|
| D | optimizer_v2.py | 440 apply_state = self._prepare(var_list)
442 functools.partial(self._distributed_apply, apply_state=apply_state),
446 def _distributed_apply(self, distribution, grads_and_vars, name, apply_state): argument
464 apply_kwargs["apply_state"] = apply_state
469 apply_kwargs["apply_state"] = apply_state
621 apply_state = {}
623 apply_state[(var_device, var_dtype)] = {}
625 self._prepare_local(var_device, var_dtype, apply_state)
627 return apply_state
629 def _prepare_local(self, var_device, var_dtype, apply_state): argument
[all …]
|
| D | nadam.py | 124 def _prepare_local(self, var_device, var_dtype, apply_state): argument
144 apply_state[(var_device, var_dtype)] = dict(
166 def _resource_apply_dense(self, grad, var, apply_state=None): argument
168 coefficients = ((apply_state or {}).get((var_device, var_dtype))
189 def _resource_apply_sparse(self, grad, var, indices, apply_state=None): argument
191 coefficients = ((apply_state or {}).get((var_device, var_dtype))
|