Quick Start¶
Create a Tree-based Tensor¶
You can create a tree-based tensor or a native tensor like the following example code.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 | import builtins
import os
from functools import partial
import treetensor.torch as torch
print = partial(builtins.print, sep=os.linesep)
if __name__ == '__main__':
t1 = torch.tensor([[1, 2, 3],
[4, 5, 6]])
print('new native tensor:', t1)
t2 = torch.tensor({
'a': [1, 2, 3],
'b': {'x': [[4, 5], [6, 7]]},
})
print('new tree tensor:', t2)
t3 = torch.randn(2, 3)
print('new random native tensor:', t3)
t4 = torch.randn({
'a': (2, 3),
'b': {'x': (3, 4)},
})
print('new random tree tensor:', t4)
|
The output should be like below.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 | new native tensor:
tensor([[1, 2, 3],
[4, 5, 6]])
new tree tensor:
<Tensor 0x7fe7ecae71d0>
├── a --> tensor([1, 2, 3])
└── b --> <Tensor 0x7fe7ecae7410>
└── x --> tensor([[4, 5],
[6, 7]])
new random native tensor:
tensor([[ 0.6374, 0.5619, 1.2875],
[-0.1674, 1.4108, -0.1149]])
new random tree tensor:
<Tensor 0x7fe7e2ec7410>
├── a --> tensor([[-1.0148, 0.7347, -1.6171],
│ [-1.1030, -2.0437, 0.1575]])
└── b --> <Tensor 0x7fe7e2ec74d0>
└── x --> tensor([[ 1.0864, 0.0259, -0.2414, 0.0771],
[ 0.4995, 0.4199, 0.1481, 0.0871],
[-0.5012, 0.1868, 1.5312, -0.5145]])
|