PyTorch torch.poisson Function
Pytorch torch Reference Manual
torch.poissonIt is a function in PyTorch used to generate random numbers from a Poisson distribution.
Function Definition
torch.poisson(lam, generator=None, out=None)
Parameter Description
lam- The parameter lambda (expected value/mean) of the Poisson distribution, must be non-negativegenerator- Random number generator (optional)out- Output tensor (optional)
Usage Example
Example
import torch
# Generate a random tensor using a single lambda value
result1 = torch.poisson(lam=5.0, size=(3, 3))
print("3x3 Poisson distribution random tensor with lambda=5:")
print(result1)
# Use a tensor as the lambda parameter
lam_tensor = torch.tensor([1.0, 2.0, 5.0, 10.0])
result2 = torch.poisson(lam_tensor)
print("nSampling result using tensor lambda:")
print(result2)
# Simulate count data
lam = torch.tensor([0.5, 1.0, 2.0, 5.0, 10.0])
samples = torch.poisson(lam)
print("nPoisson sampling with different lambda values:")
for i, (l, s) in enumerate(zip(lam.tolist(), samples.tolist())):
print(f" lambda={l}: {s}")
# Generate a random tensor using a single lambda value
result1 = torch.poisson(lam=5.0, size=(3, 3))
print("3x3 Poisson distribution random tensor with lambda=5:")
print(result1)
# Use a tensor as the lambda parameter
lam_tensor = torch.tensor([1.0, 2.0, 5.0, 10.0])
result2 = torch.poisson(lam_tensor)
print("nSampling result using tensor lambda:")
print(result2)
# Simulate count data
lam = torch.tensor([0.5, 1.0, 2.0, 5.0, 10.0])
samples = torch.poisson(lam)
print("nPoisson sampling with different lambda values:")
for i, (l, s) in enumerate(zip(lam.tolist(), samples.tolist())):
print(f" lambda={l}: {s}")
Other Extensions