PyTorch torch.mv Function
Pytorch torch Reference Manual
torch.mvis a function in PyTorch used to perform matrix-vector multiplication. It computes the product of a matrix and a vector.
Function Definition
torch.mv(input, vec, out=None)
Parameters:
input(Tensor): Input matrix, shape (n, m).vec(Tensor): Input vector, shape (m,) or (m, 1).out(Tensor, optional): Output tensor.
Return Value:
torch.Tensor: Returns the matrix-vector product, shape (n,).
Usage Example
Example
import torch
# Create matrix and vector
mat = torch.randn(3, 4)
vec = torch.randn(4)
# Matrix-vector multiplication
result = torch.mv(mat, vec)
print("Matrix shape:", mat.shape)
print("Vector shape:", vec.shape)
print("Result shape:", result.shape)
print("Result:", result)
# Create matrix and vector
mat = torch.randn(3, 4)
vec = torch.randn(4)
# Matrix-vector multiplication
result = torch.mv(mat, vec)
print("Matrix shape:", mat.shape)
print("Vector shape:", vec.shape)
print("Result shape:", result.shape)
print("Result:", result)
The output result is:
矩阵形状: torch.Size([3, 4]) 向量形状: torch.Size([4]) 结果形状: torch.Size([3]) tensor([-0.1861, 0.2482, -0.4375])
Example - Neural Network Layer Computation
import torch
# Simulate a fully connected layer of a neural network
W = torch.randn(512, 256) # Weight matrix
x = torch.randn(256) # Input vector
# Compute y = W @ x
y = torch.mv(W, x)
print("Weight matrix shape:", W.shape)
print("Input vector shape:", x.shape)
print("Output vector shape:", y.shape)
# Simulate a fully connected layer of a neural network
W = torch.randn(512, 256) # Weight matrix
x = torch.randn(256) # Input vector
# Compute y = W @ x
y = torch.mv(W, x)
print("Weight matrix shape:", W.shape)
print("Input vector shape:", x.shape)
print("Output vector shape:", y.shape)
Other Extensions