Insert module :code:`new` with name :code:`name` before element :code:`key` in sequential :code:`base` and return the new sequence.

Get a submodule at any depth of a net by its name Args: net (nn.Module): the base module containing other modules name (str)

Patch Discriminator from pix2pix

The architecture used in `Pix2Pix <https://arxiv.org/abs/1611.07004>`_, able to train on 128x128 or 512x512 images.

Set the content. Args: content_img (3xHxW tensor): an image tensor content_layer (str, optional): the layer

Set the style. Args: style_img (3xHxW tensor): an image tensor style_ratio (float): a multiplier for the sty

Tempered log softmax. Computes log softmax along dimension 1 Args: x (tensor): activations t (float): temperature n_

Set the classifier architecture to avgpool-flatten-linear1-relu-linear2.

Classification training and testing loop. Both forward functions must return a per-batch loss and logits predictions. It expands from :co

Computes the accuracy over the k top predictions for the specified values of k

Allow to edit any part of a model by recursively editing its modules. For instance, in order to delete all dropout layers and change relus i

Run the recipe Args: n_iters (int): number of iterations to run neuron (int): the feature map to maximize

Run the recipe Args: n_iters (int): number of iterations to run content (PIL.Image): content image

r""" Returns the focal loss between `target` and `input` :math:`\text{FL}(p_t)=-(1-p_t)^\gamma\log(p_t)`

Return the std needed to initialize a weight matrix with given parameters.

Remove the ReLU

r""" Returns the orthogonal loss for weight matrix `m`, from Big GAN. https://arxiv.org/abs/1809.11096 :math:`R_{\beta}(W)= ||W^T W \od

Remove the hook :code:`hook_name` applied on member :code:`name` of :code:`module`. Args: module (nn.Module): the module on whic

Recreate the block as a simple conv-BatchNorm-ReLU

Rotate the image by an angle randomly sampled between [-theta, theta] Args: theta (float): an angle in degrees

The bi-tempered loss from https://arxiv.org/abs/1906.03361 Args: x (tensor): a tensor of batched logits like for cross_entropy

Compute tempered nll loss Args: x (tensor): activations of log softmax y (tensor): labels t1 (float): temperature 1

Place the normalization and ReLU before the convolution.

Returns the total variation loss for batch of images `i`