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

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

A fast approximation of the cumulative distribution function of the standard normal.

implement SGM through grad through ReLU (This code is copied from https://github.com/csdongxian/skip-connections-matter)

Calculate gradient ∇_x L(x, y; θ)

Optimized gradient calculation with caching, mixed precision and chunking to avoid OOM

Evaluate a function without caching intermediate activations, allowing for reduced memory at the expense of extra compute in the backward pas

Return: rot -- torch.tensor, size (B, 3, 3) pts @ trans_mat Parameters: angles -- torc

If comm is provided, average all numerical stats across that comm

A counter for the `thop` package to count the operations in an attention operation. Meant to be used like: macs, params = thop.pr

Discrete Cosine Transform (DCT)

Discrete Cosine Transform, Type II (a.k.a. the DCT) (This code is copied from https://github.com/yuyang-long/SSA/blob/master/dct.py)

Discrete Cosine Transform, Type II (a.k.a. the DCT) (This code is copied from https://github.com/yuyang-long/SSA/blob/master/dct.py)

Discrete Cosine Transform, Type II (a.k.a. the DCT) (This code is copied from https://github.com/yuyang-long/SSA/blob/master/dct.py)

Discrete Cosine Transform, Type II (a.k.a. the DCT) (This code is copied from https://github.com/yuyang-long/SSA/blob/master/dct.py)

2-dimentional Discrete Cosine Transform, Type II (a.k.a. the DCT) (This code is copied from https://github.com/yuyang-long/SSA/blob/m

Defaults for image and classifier training.

Calculate gradients with respect to perturbations

The inference stage, which should be overridden when the attack need to change the models (e.g., ensemble-model attack, ghost, etc.) or the i

Calculate loss for the attack Args: logits: Model output logits label: Ground truth labels num_c

The gradient calculation, which should be overridden when the attack need to tune the gradient (e.g., TIM, variance tuning, enhanced momentum

Wraps `score_fn` so that the model output corresponds to a real time-dependent score function. Args: sde: An `sde_lib.SDE` object that repr

The inverse to DCT-II, which is a scaled Discrete Cosine Transform

The inverse to DCT-II, which is a scaled Discrete Cosine Transform, Type III Our definition of idct is that idct(dct(x)) == x

The inverse to DCT-II, which is a scaled Discrete Cosine Transform, Type III Our definition of idct is that idct(dct(x)) == x

The inverse to DCT-II, which is a scaled Discrete Cosine Transform, Type III Our definition of idct is that idct(dct(x)) == x

The inverse to DCT-II, which is a scaled Discrete Cosine Transform, Type III Our definition of idct is that idct(dct(x)) == x

The inverse to 2D DCT-II, which is a scaled Discrete Cosine Transform, Type III Our definition of idct is that idct_2d(dct_2d(x)) ==

Create the drift and diffusion functions for the reverse SDE

Load the NAT generator model from HuggingFace or local checkpoint Arguments: model_name (str): the name of surro

Copy model parameters into a (differently-shaped) list of full-precision parameters.

Compute the KL divergence between two gaussians. Shapes are automatically broadcasted, so batches can be compared to scalars, among othe

Sample x_{t-1} from the model at the given timestep. :param model: the model to sample from. :param x: the current tensor at

Generate one sample from the prior distribution, $p_T(x)$.

Diffuse the data for a given number of diffusion steps. In other words, sample from q(x_t | x_0). :param x_start: the initi

r"""ResNet-50 model from `"Deep Residual Learning for Image Recognition" <https://arxiv.org/pdf/1512.03385.pdf>'_ Args: pretrained (bo

r"""ResNet-50 model from `"Deep Residual Learning for Image Recognition" <https://arxiv.org/pdf/1512.03385.pdf>`_ Args: pretrained (b

Create the drift and diffusion functions for the reverse SDE