Overview¶

The autograd package enables automatic differentiation of NDArray operations. In machine learning applications, autograd is often used to calculate the gradients of loss functions with respect to parameters.

>>> x = mx.nd.array([1,2,3,4])
>>> x.attach_grad()
>>> with mx.autograd.record():
...     y = x * x + 1
>>> y.backward()
>>> print(x.grad)
[ 2.  4.  6.  8.]

Train mode and Predict Mode¶

Some operators (Dropout, BatchNorm, etc) behave differently in when training and when making predictions. This can be controlled with train_mode and predict_mode scope.

By default, MXNet is in predict_mode. A with autograd.record() block by default turns on train_mode (equivalent to with autograd.record(train_mode=True)). To compute a gradient in prediction mode (as when generating adversarial examples), call record with train_mode=False and then call backward(train_mode=False)

Although training usually coincides with recording, this isn’t always the case. To control training vs predict_mode without changing recording vs not recording, Use a with autograd.train_mode(): or with autograd.predict_mode(): block.

Detailed tutorials are available in Part 1 of the MXNet gluon book.

Autograd¶

API Reference¶

Autograd for NDArray.

mxnet.autograd.set_recording(is_recording)[source]¶

Set status to recording/not recording. When recording, graph will be constructed for gradient computation.

Parameters:	is_recording (bool) –
Returns:
Return type:	previous state before this set.

mxnet.autograd.set_training(train_mode)[source]¶

Set status to training/predicting. This affects ctx.is_train in operator running context. For example, Dropout will drop inputs randomly when train_mode=True while simply passing through if train_mode=False.

Parameters:	train_mode (bool) –
Returns:
Return type:	previous state before this set.

mxnet.autograd.is_recording()[source]¶

Get status on recording/not recording.

Returns:
Return type:	Current state of recording.

mxnet.autograd.is_training()[source]¶

Get status on training/predicting.

Returns:
Return type:	Current state of training/predicting.

mxnet.autograd.record(train_mode=True)[source]¶

Returns an autograd recording scope context to be used in ‘with’ statement and captures code that needs gradients to be calculated.

Note

When forwarding with train_mode=False, the corresponding backward should also use train_mode=False, otherwise gradient is undefined.

Example:

with autograd.record():
    y = model(x)
    backward([y])
metric.update(...)
optim.step(...)

Parameters:	train_mode (bool, default True) – Whether the forward pass is in training or predicting mode. This controls the behavior of some layers such as Dropout, BatchNorm.

mxnet.autograd.pause(train_mode=False)[source]¶

Returns a scope context to be used in ‘with’ statement for codes that do not need gradients to be calculated.

Example:

with autograd.record():
    y = model(x)
    backward([y])
    with autograd.pause():
        # testing, IO, gradient updates...

Parameters:	train_mode (bool, default False) – Whether to do forward for training or predicting.

mxnet.autograd.train_mode()[source]¶

Returns a scope context to be used in ‘with’ statement in which forward pass behavior is set to training mode, without changing the recording states.

Example:

y = model(x)
with autograd.train_mode():
    y = dropout(y)

mxnet.autograd.predict_mode()[source]¶

Returns a scope context to be used in ‘with’ statement in which forward pass behavior is set to inference mode, without changing the recording states.

Example:

with autograd.record():
    y = model(x)
    with autograd.predict_mode():
        y = sampling(y)
    backward([y])

mxnet.autograd.mark_variables(variables, gradients, grad_reqs='write')[source]¶

Mark NDArrays as variables to compute gradient for autograd.

Parameters:	variables (NDArray or list of NDArray) – gradients (NDArray or list of NDArray) – grad_reqs (str or list of str) –

mxnet.autograd.backward(heads, head_grads=None, retain_graph=False, train_mode=True)[source]¶

Compute the gradients of heads w.r.t previously marked variables.

Parameters:	heads (NDArray or list of NDArray) – Output NDArray(s) head_grads (NDArray or list of NDArray or None) – Gradients with respect to heads. train_mode (bool, optional) – Whether to do backward for training or predicting.

mxnet.autograd.get_symbol(x)[source]¶

Retrieve recorded computation history as Symbol.

Parameters:	x (NDArray) – Array representing the head of computation graph.
Returns:	The retrieved Symbol.
Return type:	Symbol

class mxnet.autograd.Function[source]¶

User-defined differentiable function.

Function allows defining both forward and backward computation for custom operators. During gradient computation, the used-defined backward function will be used instead of the default chain-rule. You can also cast to numpy array and back for some operations in forward and backward.

For example, a stable sigmoid function can be defined as:

class sigmoid(Function):
    def forward(self, x):
        y = 1 / (1 + mx.nd.exp(-x))
        self.save_for_backward(y)
        return y

    def backward(self, dy):
        # backward takes as many inputs as forward's return value,
        # and returns as many NDArrays as forward's arguments.
        y, = self.saved_tensors
        return y * (1-y)

forward(*inputs)[source]¶

Forward computation.

backward(*output_grads)[source]¶

Backward computation.

Takes as many inputs as forward’s outputs, and returns as many NDArrays as forward’s inputs.