LostTech.TensorFlow : API Documentation

Type TowerOptimizer

Namespace tensorflow_estimator.contrib.estimator

Parent Optimizer

Interfaces ITowerOptimizer

Public instance methods

object apply_gradients(IEnumerable<object> grads_and_vars, IGraphNodeBase global_step, PythonFunctionContainer name)

Apply gradients to variables.

This is the second part of `minimize()`. It returns an `Operation` that applies gradients.
Parameters
IEnumerable<object> grads_and_vars
List of (gradient, variable) pairs as returned by `compute_gradients()`.
IGraphNodeBase global_step
Optional `Variable` to increment by one after the variables have been updated.
PythonFunctionContainer name
Optional name for the returned operation. Default to the name passed to the `Optimizer` constructor.
Returns
object
An `Operation` that applies the specified gradients. If `global_step` was not None, that operation also increments `global_step`.

object apply_gradients(object grads_and_vars, int global_step, IDictionary<string, object> kwargs)

object apply_gradients(object grads_and_vars, BaseResourceVariable global_step, IDictionary<string, object> kwargs)

object apply_gradients(IEnumerable<object> grads_and_vars, int global_step, PythonFunctionContainer name)

Apply gradients to variables.

This is the second part of `minimize()`. It returns an `Operation` that applies gradients.
Parameters
IEnumerable<object> grads_and_vars
List of (gradient, variable) pairs as returned by `compute_gradients()`.
int global_step
Optional `Variable` to increment by one after the variables have been updated.
PythonFunctionContainer name
Optional name for the returned operation. Default to the name passed to the `Optimizer` constructor.
Returns
object
An `Operation` that applies the specified gradients. If `global_step` was not None, that operation also increments `global_step`.

object apply_gradients(ValueTuple<IEnumerable<object>, object> grads_and_vars, IGraphNodeBase global_step, IDictionary<string, object> kwargs)

object apply_gradients(ValueTuple<IEnumerable<object>, object> grads_and_vars, BaseResourceVariable global_step, IDictionary<string, object> kwargs)

object apply_gradients(ValueTuple<IEnumerable<object>, object> grads_and_vars, int global_step, IDictionary<string, object> kwargs)

object apply_gradients(IEnumerable<object> grads_and_vars, int global_step, IDictionary<string, object> kwargs)

object apply_gradients(object grads_and_vars, IGraphNodeBase global_step, IDictionary<string, object> kwargs)

object apply_gradients(IEnumerable<object> grads_and_vars, BaseResourceVariable global_step, IDictionary<string, object> kwargs)

object apply_gradients(IEnumerable<object> grads_and_vars, BaseResourceVariable global_step, PythonFunctionContainer name)

Apply gradients to variables.

This is the second part of `minimize()`. It returns an `Operation` that applies gradients.
Parameters
IEnumerable<object> grads_and_vars
List of (gradient, variable) pairs as returned by `compute_gradients()`.
BaseResourceVariable global_step
Optional `Variable` to increment by one after the variables have been updated.
PythonFunctionContainer name
Optional name for the returned operation. Default to the name passed to the `Optimizer` constructor.
Returns
object
An `Operation` that applies the specified gradients. If `global_step` was not None, that operation also increments `global_step`.

object apply_gradients(IEnumerable<object> grads_and_vars, IGraphNodeBase global_step, IDictionary<string, object> kwargs)

object apply_gradients_dyn(object grads_and_vars, object global_step, IDictionary<string, object> kwargs)

object compute_gradients(PythonFunctionContainer loss, Object[] args)

Compute gradients of `loss` for the variables in `var_list`.

This is the first part of `minimize()`. It returns a list of (gradient, variable) pairs where "gradient" is the gradient for "variable". Note that "gradient" can be a `Tensor`, an `IndexedSlices`, or `None` if there is no gradient for the given variable.
Parameters
PythonFunctionContainer loss
A Tensor containing the value to minimize or a callable taking no arguments which returns the value to minimize. When eager execution is enabled it must be a callable.
Object[] args
Returns
object
A list of (gradient, variable) pairs. Variable is always present, but gradient can be `None`.

object compute_gradients(object loss, IEnumerable<object> var_list, ImplicitContainer<T> gate_gradients, object aggregation_method, bool colocate_gradients_with_ops, IGraphNodeBase grad_loss)

Compute gradients of `loss` for the variables in `var_list`.

This is the first part of `minimize()`. It returns a list of (gradient, variable) pairs where "gradient" is the gradient for "variable". Note that "gradient" can be a `Tensor`, an `IndexedSlices`, or `None` if there is no gradient for the given variable.
Parameters
object loss
A Tensor containing the value to minimize or a callable taking no arguments which returns the value to minimize. When eager execution is enabled it must be a callable.
IEnumerable<object> var_list
Optional list or tuple of tf.Variable to update to minimize `loss`. Defaults to the list of variables collected in the graph under the key `GraphKeys.TRAINABLE_VARIABLES`.
ImplicitContainer<T> gate_gradients
How to gate the computation of gradients. Can be `GATE_NONE`, `GATE_OP`, or `GATE_GRAPH`.
object aggregation_method
Specifies the method used to combine gradient terms. Valid values are defined in the class `AggregationMethod`.
bool colocate_gradients_with_ops
If True, try colocating gradients with the corresponding op.
IGraphNodeBase grad_loss
Optional. A `Tensor` holding the gradient computed for `loss`.
Returns
object
A list of (gradient, variable) pairs. Variable is always present, but gradient can be `None`.

object compute_gradients(PythonFunctionContainer loss, IDictionary<string, object> kwargs, Object[] args)

Compute gradients of `loss` for the variables in `var_list`.

This is the first part of `minimize()`. It returns a list of (gradient, variable) pairs where "gradient" is the gradient for "variable". Note that "gradient" can be a `Tensor`, an `IndexedSlices`, or `None` if there is no gradient for the given variable.
Parameters
PythonFunctionContainer loss
A Tensor containing the value to minimize or a callable taking no arguments which returns the value to minimize. When eager execution is enabled it must be a callable.
IDictionary<string, object> kwargs
Object[] args
Returns
object
A list of (gradient, variable) pairs. Variable is always present, but gradient can be `None`.

object compute_gradients_dyn(object loss, object var_list, ImplicitContainer<T> gate_gradients, object aggregation_method, ImplicitContainer<T> colocate_gradients_with_ops, object grad_loss)

Compute gradients of `loss` for the variables in `var_list`.

This is the first part of `minimize()`. It returns a list of (gradient, variable) pairs where "gradient" is the gradient for "variable". Note that "gradient" can be a `Tensor`, an `IndexedSlices`, or `None` if there is no gradient for the given variable.
Parameters
object loss
A Tensor containing the value to minimize or a callable taking no arguments which returns the value to minimize. When eager execution is enabled it must be a callable.
object var_list
Optional list or tuple of tf.Variable to update to minimize `loss`. Defaults to the list of variables collected in the graph under the key `GraphKeys.TRAINABLE_VARIABLES`.
ImplicitContainer<T> gate_gradients
How to gate the computation of gradients. Can be `GATE_NONE`, `GATE_OP`, or `GATE_GRAPH`.
object aggregation_method
Specifies the method used to combine gradient terms. Valid values are defined in the class `AggregationMethod`.
ImplicitContainer<T> colocate_gradients_with_ops
If True, try colocating gradients with the corresponding op.
object grad_loss
Optional. A `Tensor` holding the gradient computed for `loss`.
Returns
object
A list of (gradient, variable) pairs. Variable is always present, but gradient can be `None`.

object compute_gradients_dyn(object loss, Object[] args)

Compute gradients of `loss` for the variables in `var_list`.

This is the first part of `minimize()`. It returns a list of (gradient, variable) pairs where "gradient" is the gradient for "variable". Note that "gradient" can be a `Tensor`, an `IndexedSlices`, or `None` if there is no gradient for the given variable.
Parameters
object loss
A Tensor containing the value to minimize or a callable taking no arguments which returns the value to minimize. When eager execution is enabled it must be a callable.
Object[] args
Returns
object
A list of (gradient, variable) pairs. Variable is always present, but gradient can be `None`.

object compute_gradients_dyn(object loss, IDictionary<string, object> kwargs, Object[] args)

Compute gradients of `loss` for the variables in `var_list`.

This is the first part of `minimize()`. It returns a list of (gradient, variable) pairs where "gradient" is the gradient for "variable". Note that "gradient" can be a `Tensor`, an `IndexedSlices`, or `None` if there is no gradient for the given variable.
Parameters
object loss
A Tensor containing the value to minimize or a callable taking no arguments which returns the value to minimize. When eager execution is enabled it must be a callable.
IDictionary<string, object> kwargs
Object[] args
Returns
object
A list of (gradient, variable) pairs. Variable is always present, but gradient can be `None`.

string get_name(IDictionary<string, object> kwargs, Object[] args)

string get_name(Object[] args)

object get_name_dyn(Object[] args)

object get_name_dyn(IDictionary<string, object> kwargs, Object[] args)

object variables(Object[] args)

object variables(IDictionary<string, object> kwargs, Object[] args)

object variables_dyn(IDictionary<string, object> kwargs, Object[] args)

object variables_dyn(Object[] args)

Public static methods

bool has_been_used()

object has_been_used_dyn()

TowerOptimizer NewDyn(object optimizer_or_optimizer_fn)

Creates a `ClusterSpec`.

Public properties

object COLLECTION_FOR_GRAPH_STATES_dyn get; set;

object PythonObject get;

Public fields

string COLLECTION_FOR_GRAPH_STATES

return string