GroupNorm¶

class nvtripy.GroupNorm(num_groups: int, num_channels: int, dtype: dtype = float32, eps: float = 1e-05)[source]¶

Bases: Module

Applies group normalization over the input tensor:

\(\text{GroupNorm}(x) = \Large \frac{x - \bar{x}}{ \sqrt{\sigma^2 + \epsilon}} \normalsize * \gamma + \beta\)

where \(\bar{x}\) is the mean and \(\sigma^2\) is the variance.

Parameters:

num_groups (int) – The number of groups to split the channels into.
num_channels (int) – The number of channels expected in the input.
dtype (dtype) – The data type to use for the weight and bias parameters.
eps (float) – \(\epsilon\) value to prevent division by zero.

Example

group_norm = tp.GroupNorm(2, 2)

group_norm.weight = tp.iota(group_norm.weight.shape)
group_norm.bias = tp.iota(group_norm.bias.shape)

input = tp.iota((1, 2, 2, 2), dim=1)
output = group_norm(input)

Local Variables¶

>>> group_norm
GroupNorm(
    weight: Parameter = (shape=(2,), dtype=float32),
    bias: Parameter = (shape=(2,), dtype=float32),
)
>>> group_norm.state_dict()
{
    weight: tensor([0, 1], dtype=float32, loc=gpu:0, shape=(2,)),
    bias: tensor([0, 1], dtype=float32, loc=gpu:0, shape=(2,)),
}

>>> input
tensor(
    [[[[0, 0],
       [0, 0]],

      [[1, 1],
       [1, 1]]]], 
    dtype=float32, loc=gpu:0, shape=(1, 2, 2, 2))

>>> output
tensor(
    [[[[0, 0],
       [0, 0]],

      [[1, 1],
       [1, 1]]]], 
    dtype=float32, loc=gpu:0, shape=(1, 2, 2, 2))

__call__(*args: Any, **kwargs: Any) → Any¶

Calls the module with the specified arguments.

Parameters:

*args (Any) – Positional arguments to the module.
**kwargs (Any) – Keyword arguments to the module.

Returns:

The outputs computed by the module.

Return type:

Any

Example

class Module(tp.Module):
    def forward(self, x):
        return tp.relu(x)


module = Module()

input = tp.arange(-3, 3)
out = module(input)  # Note that we do not call `forward` directly.

Local Variables¶

>>> module
Module(
)
>>> module.state_dict()
{}

>>> input
tensor([-3, -2, -1, 0, 1, 2], dtype=float32, loc=gpu:0, shape=(6,))

>>> out
tensor([0, 0, 0, 0, 1, 2], dtype=float32, loc=gpu:0, shape=(6,))

load_state_dict(state_dict: Dict[str, Tensor], strict: bool = True) → Tuple[Set[str], Set[str]]¶

Loads parameters from the provided state_dict into the current module. This will recurse over any nested child modules.

Parameters:

state_dict (Dict[str, Tensor]) – A dictionary mapping names to parameters.
strict (bool) – If True, keys in state_dict must exactly match those in this module. If not, an error will be raised.

Returns:

missing_keys: keys that are expected by this module but not provided in state_dict.
unexpected_keys: keys that are not expected by this module but provided in state_dict.

Return type:

A tuple of two sets of strings representing

Example

class MyModule(tp.Module):
    def __init__(self):
        super().__init__()
        self.param = tp.ones((2,), dtype=tp.float32)


module = MyModule()

print(f"Before: {module.param}")

module.load_state_dict({"param": tp.zeros((2,), dtype=tp.float32)})

print(f"After: {module.param}")

Output¶

Before: tensor([1, 1], dtype=float32, loc=gpu:0, shape=(2,))
After: tensor([0, 0], dtype=float32, loc=gpu:0, shape=(2,))

See also

state_dict()

named_children() → Iterator[Tuple[str, Module]]¶

Returns an iterator over immediate children of this module, yielding tuples containing the name of the child module and the child module itself.

Returns:: An iterator over tuples containing the name of the child module and the child module itself.
Return type:: Iterator[Tuple[str, Module]]

Example

class StackedLinear(tp.Module):
    def __init__(self):
        super().__init__()
        self.linear1 = tp.Linear(2, 2)
        self.linear2 = tp.Linear(2, 2)


stacked_linear = StackedLinear()

for name, module in stacked_linear.named_children():
    print(f"{name}: {type(module).__name__}")

Output¶

linear1: Linear
linear2: Linear

named_parameters() → Iterator[Tuple[str, Tensor]]¶

Returns:: An iterator over tuples containing the name of a parameter and the parameter itself.
Return type:: Iterator[Tuple[str, Tensor]]

Example

class MyModule(tp.Module):
    def __init__(self):
        super().__init__()
        self.alpha = tp.Tensor(1)
        self.beta = tp.Tensor(2)


linear = MyModule()

for name, parameter in linear.named_parameters():
    print(f"{name}: {parameter}")

Output¶

alpha: tensor(1, dtype=int32, loc=cpu:0, shape=())
beta: tensor(2, dtype=int32, loc=cpu:0, shape=())

state_dict() → Dict[str, Tensor]¶

Returns a dictionary mapping names to parameters in the module. This will recurse over any nested child modules.

Returns:: A dictionary mapping names to parameters.
Return type:: Dict[str, Tensor]

Example

class MyModule(tp.Module):
    def __init__(self):
        super().__init__()
        self.param = tp.ones((2,), dtype=tp.float32)
        self.linear1 = tp.Linear(2, 2)
        self.linear2 = tp.Linear(2, 2)


module = MyModule()

state_dict = module.state_dict()

Local Variables¶

>>> state_dict
{
    param: tensor([1, 1], dtype=float32, loc=gpu:0, shape=(2,)),
    linear1.weight: <nvtripy.frontend.module.parameter.DefaultParameter object at 0x79774be1c0d0>,
    linear1.bias: <nvtripy.frontend.module.parameter.DefaultParameter object at 0x79774be21400>,
    linear2.weight: <nvtripy.frontend.module.parameter.DefaultParameter object at 0x79774bdc0340>,
    linear2.bias: <nvtripy.frontend.module.parameter.DefaultParameter object at 0x79774bdc03d0>,
}

num_groups: int¶: The number of groups to split the channels into.

num_channels: int¶: The number of channels expected in the input.

dtype: dtype¶: The data type used to perform the operation.

weight: Tensor¶: The \(\gamma\) parameter of shape \([\text{num_channels}]\).

bias: Tensor¶: The \(\beta\) parameter of shape \([\text{num_channels}]\).

eps: float¶: A value added to the denominator to prevent division by zero. Defaults to 1e-5.

forward(x: Tensor) → Tensor[source]¶

Parameters:: x (Tensor) – The input tensor.
Returns:: A tensor of the same shape as the input.
Return type:: Tensor