| Package | Description |
|---|---|
| org.bytedeco.pytorch | |
| org.bytedeco.pytorch.global |
| Modifier and Type | Class and Description |
|---|---|
class |
TraceableFunction
See Node::is_traceable() for definition.
|
| Modifier and Type | Method and Description |
|---|---|
Node |
NodeSet.front() |
Node |
Edge.function()
The function this
Edge points to. |
Node |
NodeSet.Iterator.get() |
Node |
Node.getptr()
Nodes are neither copyable nor moveable.
|
Node |
AutogradMeta.grad_fn_() |
Node |
TensorBase.grad_fn()
Gets the gradient function of the
Variable. |
Node |
VariableHooksInterface.grad_fn(TensorBase arg0) |
| Modifier and Type | Method and Description |
|---|---|
void |
AnomalyMetadata.assign_parent(Node parent_node) |
void |
NodeSet.erase(Node value) |
Edge |
Edge.function(Node setter) |
AutogradMeta |
AutogradMeta.grad_fn_(Node setter) |
void |
NodeSet.insert(Node value) |
| Constructor and Description |
|---|
Edge(Node function_,
int input_nr_) |
| Modifier and Type | Method and Description |
|---|---|
static Node |
torch.get_current_node() |
static Node |
torch.grad_accumulator(Tensor arg0)
Gets the gradient accumulator of the
Variable if it has one, or else
create one on the fly and return it. |
static Node |
torch.grad_fn_unsafe(Tensor arg0)
Gets the raw gradient function pointer, whatever it currently is.
|
static Node |
torch.try_get_grad_accumulator(Tensor arg0)
Attempts to get a pointer to the gradient accumulator of the
Variable,
if it still exists. |
| Modifier and Type | Method and Description |
|---|---|
static void |
torch.create_gradient_edge(Tensor variable,
Node function)
Create an
Edge between the given variable and the function, which is
assumed to be the gradient function of this variable (i.e. |
static void |
torch.deleteNode(Node function) |
Copyright © 2024. All rights reserved.