PyTorch
torch functions
torch.Tensor: A multi-dimensional array containing elements of a single data type.torch.tensor: Constructs a tensor with data and properties.torch.zeros: Returns a tensor filled with zeros.torch.ones: Returns a tensor filled with ones.torch.arange: Returns a 1-D tensor of evenly spaced values.torch.linspace: Returns a 1-D tensor of evenly spaced values over a specified interval.torch.rand: Returns a tensor filled with random numbers from a uniform distribution on the interval [0, 1).torch.randn: Returns a tensor filled with random numbers from a normal distribution with mean 0 and variance 1.torch.empty: Returns a tensor with uninitialized data.torch.full: Creates a tensor of specified size filled with a specified value.torch.eye: Returns a 2-D tensor with ones on the diagonal and zeros elsewhere.
Tensor operations
torch.add: Adds two tensors element-wise.torch.sub: Subtracts two tensors element-wise.torch.mul: Multiplies two tensors element-wise.torch.div: Divides two tensors element-wise.torch.matmul: Performs matrix multiplication.torch.pow: Raises each element of a tensor to a power.torch.exp: Calculates the exponential of each element of a tensor.torch.log: Calculates the natural logarithm of each element of a tensor.torch.sqrt: Calculates the square root of each element of a tensor.torch.abs: Calculates the absolute value of each element of a tensor.torch.neg: Negates each element of a tensor.torch.round: Rounds each element of a tensor to the nearest integer.torch.floor: Returns the floor of each element of a tensor.torch.ceil: Returns the ceiling of each element of a tensor.torch.clamp: Clamps all elements in input into the range [ min, max ].torch.sum: Returns the sum of all elements in the input tensor.torch.mean: Returns the mean of all elements in the input tensor.torch.std: Returns the standard deviation of all elements in the input tensor.torch.var: Returns the variance of all elements in the input tensor.torch.max: Returns the maximum value of all elements in the input tensor.torch.min: Returns the minimum value of all elements in the input tensor.torch.argmax: Returns the index of the maximum value of all elements in the input tensor.torch.argmin: Returns the index of the minimum value of all elements in the input tensor.torch.sort: Sorts the elements of the input tensor along a given dimension.torch.topk: Returns the k largest elements of the input tensor along a given dimension.torch.reshape: Returns a tensor with the same data and number of elements as input, but with the specified shape.torch.transpose: Returns a view of the input tensor with its dimensions swapped.torch.squeeze: Returns a tensor with all the dimensions of input of size 1 removed.torch.unsqueeze: Returns a new tensor with a dimension of size one inserted at the specified position.torch.cat: Concatenates the given tensors in the given dimension.torch.stack: Concatenates a sequence of tensors along a new dimension.torch.chunk: Splits a tensor into a specific number of chunks.torch.split: Splits a tensor into chunks of a specific size.
Neural network modules
torch.nn.Module: Base class for all neural network modules.torch.nn.Linear: Applies a linear transformation to the incoming data.torch.nn.Conv2d: Applies a 2D convolution over an input signal composed of several input planes.torch.nn.MaxPool2d: Applies a 2D max pooling over an input signal.torch.nn.ReLU: Applies the rectified linear unit function element-wise.torch.nn.Sigmoid: Applies the sigmoid function element-wise.torch.nn.Tanh: Applies the hyperbolic tangent function element-wise.torch.nn.BatchNorm2d: Applies Batch Normalization over a 4D input.torch.nn.Dropout: During training, randomly zeroes some of the elements of the input tensor with probability p.torch.nn.Embedding: A simple lookup table that stores embeddings of a fixed dictionary and size.
Loss functions
torch.nn.MSELoss: Creates a criterion that measures the mean squared error (squared L2 norm) between each element in the input and target.torch.nn.CrossEntropyLoss: This criterion computes the cross entropy loss between input and target.torch.nn.BCELoss: Creates a criterion that measures the Binary Cross Entropy between the target and the output.torch.nn.L1Loss: Creates a criterion that measures the mean absolute error (MAE) between each element in the input and target.
Optimizers
torch.optim.SGD: Implements stochastic gradient descent (optionally with momentum).torch.optim.Adam: Implements Adam algorithm.torch.optim.RMSprop: Implements RMSprop algorithm.
Autograd
torch.autograd.grad: Computes and returns the sum of gradients of outputs with respect to the inputs.
Utilities
torch.device: Represents the device on which a torch.Tensor is or will be allocated.torch.cuda.is_available: Returns True if CUDA is available.torch.save: Saves a tensor to disk.torch.load: Loads a tensor from disk.
Other
torch.no_grad: Context-manager that disables gradient calculation.torch.set_grad_enabled: Enable or disable grad, depending on its argument.