PyTorch torch.tensor_split Function
PyTorch torch Reference Manual
torch.tensor_splitIt is a function in PyTorch used to split tensors by index or number of segments.
Function Definition
torch.tensor_split(input, indices_or_sections, dim=0)
Usage Example
Example
import torch
x = torch.arange(12)
print("Original tensor:")
print(x)
# Split by index
result = torch.tensor_split(x, [2, 5, 8])
print("Split by indices [2, 5, 8]:")
for i, t in enumerate(result):
print(f" Block {i}: {t}")
# Split by number of segments (average into 3 parts)
y = torch.arange(9)
result = torch.tensor_split(y, 3)
print("nAverage split into 3 parts:")
for i, t in enumerate(result):
print(f" Block {i}: {t}")
# Split the 2D tensor by columns
z = torch.arange(12).reshape(3, 4)
print("n2D tensor:")
print(z)
result = torch.tensor_split(z, 2, dim=1)
print("Split into 2 parts by column:")
for i, t in enumerate(result):
print(f" Block {i}:n{t}")
x = torch.arange(12)
print("Original tensor:")
print(x)
# Split by index
result = torch.tensor_split(x, [2, 5, 8])
print("Split by indices [2, 5, 8]:")
for i, t in enumerate(result):
print(f" Block {i}: {t}")
# Split by number of segments (average into 3 parts)
y = torch.arange(9)
result = torch.tensor_split(y, 3)
print("nAverage split into 3 parts:")
for i, t in enumerate(result):
print(f" Block {i}: {t}")
# Split the 2D tensor by columns
z = torch.arange(12).reshape(3, 4)
print("n2D tensor:")
print(z)
result = torch.tensor_split(z, 2, dim=1)
print("Split into 2 parts by column:")
for i, t in enumerate(result):
print(f" Block {i}:n{t}")
The output result is:
原始张量:
tensor([ 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11])
按索引 [2, 5, 8] 分割:
块 0: tensor([0, 1])
Chunk 1: tensor([2, 3, 4])
Chunk 2: tensor([5, 6, 7])
Chunk 3: tensor([ 8, 9, 10, 11])
平均分为 3 份:
块 0: tensor([0, 1, 2])
Chunk 1: tensor([3, 4, 5])
Chunk 2: tensor([6, 7, 8])
二维张量:
tensor([[ 0, 1, 2, 3],
[ 4, 5, 6, 7],
[ 8, 9, 10, 11]])
按列分为 2 份:
块 0:
tensor([[0, 1],
[4, 5],
[8, 9]])
Chunk 1:
tensor([[ 2, 3],
[ 6, 7],
[10, 11]])
Other Extensions