Skip to content

D
DI-treetensor

flybirding10011 / DI-treetensor
与 Fork 源项目一致

Fork自 OpenDILab开源决策智能平台 / DI-treetensor

通知 1
Star 0
Fork 0
D
DI-treetensor

体验新版 GitCode，发现更多精彩内容 >>

提交 0ba6d5f2 编写于 作者: HansBug's avatar HansBug 😆
浏览文件
操作

dev, doc, test(hansbug): complete exp, exp2, sqrt, log, log2, log10

上级 18f597d7
隐藏空白更改
内联 并排
Showing with 1001 addition and 25 deletion
test/torch/test_funcs.py
浏览文件 @ 0ba6d5f2
import math
import torch
import treetensor.torch as ttorch
......@@ -721,6 +723,33 @@ class TestTorchFuncs:
'b': {'x': [[False, False, False], [False, False, True]]},
})).all()
@choose_mark()
def test_isclose(self):
t1 = ttorch.isclose(
ttorch.tensor((1., 2, 3)),
ttorch.tensor((1 + 1e-10, 3, 4))
)
assert isinstance(t1, torch.Tensor)
assert (t1 == ttorch.tensor([True, False, False])).all()
t2 = ttorch.isclose(
ttorch.tensor({
'a': [1., 2, 3],
'b': {'x': [[float('inf'), 4, 1e20],
[-math.inf, 2.2943, 9483.32]]},
}),
ttorch.tensor({
'a': [1 + 1e-10, 3, 4],
'b': {'x': [[math.inf, 6, 1e20 + 1],
[-float('inf'), 2.294300000001, 9484.32]]},
}),
)
assert (t2 == ttorch.tensor({
'a': [True, False, False],
'b': {'x': [[True, False, True],
[True, True, False]]},
})).all()
@choose_mark()
def test_abs(self):
t1 = ttorch.abs(ttorch.tensor([12, 0, -3]))
......@@ -1194,3 +1223,243 @@ class TestTorchFuncs:
[-7, -2, -7]]],
}
}))
@choose_mark()
def test_exp(self):
t1 = ttorch.exp(ttorch.tensor([-4.0, -1.0, 0, 2.0, 4.8, 8.0]))
assert isinstance(t1, torch.Tensor)
assert ttorch.isclose(t1, ttorch.tensor(
[1.8316e-02, 3.6788e-01, 1.0000e+00, 7.3891e+00, 1.2151e+02, 2.9810e+03]), rtol=1e-4).all()
t2 = ttorch.exp(ttorch.tensor({
'a': [-4.0, -1.0, 0, 2.0, 4.8, 8.0],
'b': {'x': [[-2.0, 1.2, 0.25],
[16.0, 3.75, -2.34]]},
}))
assert ttorch.isclose(t2, ttorch.tensor({
'a': [1.8316e-02, 3.6788e-01, 1.0000e+00, 7.3891e+00, 1.2151e+02, 2.9810e+03],
'b': {'x': [[1.3534e-01, 3.3201e+00, 1.2840e+00],
[8.8861e+06, 4.2521e+01, 9.6328e-02]]},
}), rtol=1e-4).all()
@choose_mark()
def test_exp_(self):
t1 = ttorch.tensor([-4.0, -1.0, 0, 2.0, 4.8, 8.0])
t1r = ttorch.exp_(t1)
assert t1r is t1
assert isinstance(t1, torch.Tensor)
assert ttorch.isclose(t1, ttorch.tensor(
[1.8316e-02, 3.6788e-01, 1.0000e+00, 7.3891e+00, 1.2151e+02, 2.9810e+03]), rtol=1e-4).all()
t2 = ttorch.tensor({
'a': [-4.0, -1.0, 0, 2.0, 4.8, 8.0],
'b': {'x': [[-2.0, 1.2, 0.25],
[16.0, 3.75, -2.34]]},
})
t2r = ttorch.exp_(t2)
assert t2r is t2
assert ttorch.isclose(t2, ttorch.tensor({
'a': [1.8316e-02, 3.6788e-01, 1.0000e+00, 7.3891e+00, 1.2151e+02, 2.9810e+03],
'b': {'x': [[1.3534e-01, 3.3201e+00, 1.2840e+00],
[8.8861e+06, 4.2521e+01, 9.6328e-02]]},
}), rtol=1e-4).all()
@choose_mark()
def test_exp2(self):
t1 = ttorch.exp2(ttorch.tensor([-4.0, -1.0, 0, 2.0, 4.8, 8.0]))
assert isinstance(t1, torch.Tensor)
assert ttorch.isclose(t1, ttorch.tensor(
[6.2500e-02, 5.0000e-01, 1.0000e+00, 4.0000e+00, 2.7858e+01, 2.5600e+02]), rtol=1e-4).all()
t2 = ttorch.exp2(ttorch.tensor({
'a': [-4.0, -1.0, 0, 2.0, 4.8, 8.0],
'b': {'x': [[-2.0, 1.2, 0.25],
[16.0, 3.75, -2.34]]},
}))
assert ttorch.isclose(t2, ttorch.tensor({
'a': [6.2500e-02, 5.0000e-01, 1.0000e+00, 4.0000e+00, 2.7858e+01, 2.5600e+02],
'b': {'x': [[2.5000e-01, 2.2974e+00, 1.1892e+00],
[6.5536e+04, 1.3454e+01, 1.9751e-01]]},
}), rtol=1e-4).all()
@choose_mark()
def test_exp2_(self):
t1 = ttorch.tensor([-4.0, -1.0, 0, 2.0, 4.8, 8.0])
t1r = ttorch.exp2_(t1)
assert t1r is t1
assert isinstance(t1, torch.Tensor)
assert ttorch.isclose(t1, ttorch.tensor(
[6.2500e-02, 5.0000e-01, 1.0000e+00, 4.0000e+00, 2.7858e+01, 2.5600e+02]), rtol=1e-4).all()
t2 = ttorch.tensor({
'a': [-4.0, -1.0, 0, 2.0, 4.8, 8.0],
'b': {'x': [[-2.0, 1.2, 0.25],
[16.0, 3.75, -2.34]]},
})
t2r = ttorch.exp2_(t2)
assert t2r is t2
assert ttorch.isclose(t2, ttorch.tensor({
'a': [6.2500e-02, 5.0000e-01, 1.0000e+00, 4.0000e+00, 2.7858e+01, 2.5600e+02],
'b': {'x': [[2.5000e-01, 2.2974e+00, 1.1892e+00],
[6.5536e+04, 1.3454e+01, 1.9751e-01]]},
}), rtol=1e-4).all()
@choose_mark()
def test_sqrt(self):
t1 = ttorch.sqrt(ttorch.tensor([-4.0, -1.0, 0, 2.0, 4.8, 8.0]))
assert isinstance(t1, torch.Tensor)
assert ttorch.isclose(t1, ttorch.tensor(
[math.nan, math.nan, 0.0000, 1.4142, 2.1909, 2.8284]), rtol=1e-4, equal_nan=True).all()
t2 = ttorch.sqrt(ttorch.tensor({
'a': [-4.0, -1.0, 0, 2.0, 4.8, 8.0],
'b': {'x': [[-2.0, 1.2, 0.25],
[16.0, 3.75, -2.34]]},
}))
assert ttorch.isclose(t2, ttorch.tensor({
'a': [math.nan, math.nan, 0.0000, 1.4142, 2.1909, 2.8284],
'b': {'x': [[math.nan, 1.0954, 0.5000],
[4.0000, 1.9365, math.nan]]},
}), rtol=1e-4, equal_nan=True).all()
@choose_mark()
def test_sqrt_(self):
t1 = ttorch.tensor([-4.0, -1.0, 0, 2.0, 4.8, 8.0])
t1r = ttorch.sqrt_(t1)
assert t1r is t1
assert isinstance(t1, torch.Tensor)
assert ttorch.isclose(t1, ttorch.tensor(
[math.nan, math.nan, 0.0000, 1.4142, 2.1909, 2.8284]), rtol=1e-4, equal_nan=True).all()
t2 = ttorch.tensor({
'a': [-4.0, -1.0, 0, 2.0, 4.8, 8.0],
'b': {'x': [[-2.0, 1.2, 0.25],
[16.0, 3.75, -2.34]]},
})
t2r = ttorch.sqrt_(t2)
assert t2r is t2
assert ttorch.isclose(t2, ttorch.tensor({
'a': [math.nan, math.nan, 0.0000, 1.4142, 2.1909, 2.8284],
'b': {'x': [[math.nan, 1.0954, 0.5000],
[4.0000, 1.9365, math.nan]]},
}), rtol=1e-4, equal_nan=True).all()
@choose_mark()
def test_log(self):
t1 = ttorch.log(ttorch.tensor([-4.0, -1.0, 0, 2.0, 4.8, 8.0]))
assert isinstance(t1, torch.Tensor)
assert ttorch.isclose(t1, ttorch.tensor(
[math.nan, math.nan, -math.inf, 0.6931, 1.5686, 2.0794]), rtol=1e-4, equal_nan=True).all()
t2 = ttorch.log(ttorch.tensor({
'a': [-4.0, -1.0, 0, 2.0, 4.8, 8.0],
'b': {'x': [[-2.0, 1.2, 0.25],
[16.0, 3.75, -2.34]]},
}))
assert ttorch.isclose(t2, ttorch.tensor({
'a': [math.nan, math.nan, -math.inf, 0.6931, 1.5686, 2.0794],
'b': {'x': [[math.nan, 0.1823, -1.3863],
[2.7726, 1.3218, math.nan]]},
}), rtol=1e-4, atol=1e-4, equal_nan=True).all()
@choose_mark()
def test_log_(self):
t1 = ttorch.tensor([-4.0, -1.0, 0, 2.0, 4.8, 8.0])
t1r = ttorch.log_(t1)
assert t1r is t1
assert isinstance(t1, torch.Tensor)
assert ttorch.isclose(t1, ttorch.tensor(
[math.nan, math.nan, -math.inf, 0.6931, 1.5686, 2.0794]), rtol=1e-4, equal_nan=True).all()
t2 = ttorch.tensor({
'a': [-4.0, -1.0, 0, 2.0, 4.8, 8.0],
'b': {'x': [[-2.0, 1.2, 0.25],
[16.0, 3.75, -2.34]]},
})
t2r = ttorch.log_(t2)
assert t2r is t2
assert ttorch.isclose(t2, ttorch.tensor({
'a': [math.nan, math.nan, -math.inf, 0.6931, 1.5686, 2.0794],
'b': {'x': [[math.nan, 0.1823, -1.3863],
[2.7726, 1.3218, math.nan]]},
}), rtol=1e-4, atol=1e-4, equal_nan=True).all()
@choose_mark()
def test_log2(self):
t1 = ttorch.log2(ttorch.tensor([-4.0, -1.0, 0, 2.0, 4.8, 8.0]))
assert isinstance(t1, torch.Tensor)
assert ttorch.isclose(t1, ttorch.tensor(
[math.nan, math.nan, -math.inf, 1.0000, 2.2630, 3.0000]), rtol=1e-4, equal_nan=True).all()
t2 = ttorch.log2(ttorch.tensor({
'a': [-4.0, -1.0, 0, 2.0, 4.8, 8.0],
'b': {'x': [[-2.0, 1.2, 0.25],
[16.0, 3.75, -2.34]]},
}))
assert ttorch.isclose(t2, ttorch.tensor({
'a': [math.nan, math.nan, -math.inf, 1.0000, 2.2630, 3.0000],
'b': {'x': [[math.nan, 0.2630, -2.0000],
[4.0000, 1.9069, math.nan]]},
}), rtol=1e-4, atol=1e-4, equal_nan=True).all()
@choose_mark()
def test_log2_(self):
t1 = ttorch.tensor([-4.0, -1.0, 0, 2.0, 4.8, 8.0])
t1r = ttorch.log2_(t1)
assert t1r is t1
assert isinstance(t1, torch.Tensor)
assert ttorch.isclose(t1, ttorch.tensor(
[math.nan, math.nan, -math.inf, 1.0000, 2.2630, 3.0000]), rtol=1e-4, equal_nan=True).all()
t2 = ttorch.tensor({
'a': [-4.0, -1.0, 0, 2.0, 4.8, 8.0],
'b': {'x': [[-2.0, 1.2, 0.25],
[16.0, 3.75, -2.34]]},
})
t2r = ttorch.log2_(t2)
assert t2r is t2
assert ttorch.isclose(t2, ttorch.tensor({
'a': [math.nan, math.nan, -math.inf, 1.0000, 2.2630, 3.0000],
'b': {'x': [[math.nan, 0.2630, -2.0000],
[4.0000, 1.9069, math.nan]]},
}), rtol=1e-4, atol=1e-4, equal_nan=True).all()
@choose_mark()
def test_log10(self):
t1 = ttorch.log10(ttorch.tensor([-4.0, -1.0, 0, 2.0, 4.8, 8.0]))
assert isinstance(t1, torch.Tensor)
assert ttorch.isclose(t1, ttorch.tensor(
[math.nan, math.nan, -math.inf, 0.3010, 0.6812, 0.9031]), rtol=1e-4, equal_nan=True).all()
t2 = ttorch.log10(ttorch.tensor({
'a': [-4.0, -1.0, 0, 2.0, 4.8, 8.0],
'b': {'x': [[-2.0, 1.2, 0.25],
[16.0, 3.75, -2.34]]},
}))
assert ttorch.isclose(t2, ttorch.tensor({
'a': [math.nan, math.nan, -math.inf, 0.3010, 0.6812, 0.9031],
'b': {'x': [[math.nan, 0.0792, -0.6021],
[1.2041, 0.5740, math.nan]]},
}), rtol=1e-4, atol=1e-4, equal_nan=True).all()
@choose_mark()
def test_log10_(self):
t1 = ttorch.tensor([-4.0, -1.0, 0, 2.0, 4.8, 8.0])
t1r = ttorch.log10_(t1)
assert t1r is t1
assert isinstance(t1, torch.Tensor)
assert ttorch.isclose(t1, ttorch.tensor(
[math.nan, math.nan, -math.inf, 0.3010, 0.6812, 0.9031]), rtol=1e-4, equal_nan=True).all()
t2 = ttorch.tensor({
'a': [-4.0, -1.0, 0, 2.0, 4.8, 8.0],
'b': {'x': [[-2.0, 1.2, 0.25],
[16.0, 3.75, -2.34]]},
})
t2r = ttorch.log10_(t2)
assert t2r is t2
assert ttorch.isclose(t2, ttorch.tensor({
'a': [math.nan, math.nan, -math.inf, 0.3010, 0.6812, 0.9031],
'b': {'x': [[math.nan, 0.0792, -0.6021],
[1.2041, 0.5740, math.nan]]},
}), rtol=1e-4, atol=1e-4, equal_nan=True).all()
test/torch/test_tensor.py
浏览文件 @ 0ba6d5f2
import math
import numpy as np
import torch
from treevalue import func_treelize, typetrans, TreeValue
......@@ -349,6 +351,27 @@ class TestTorchTensor:
'b': {'x': [[False, False, False], [False, False, True]]},
})).all()
@choose_mark()
def test_isclose(self):
t1 = ttorch.tensor((1., 2, 3)).isclose(ttorch.tensor((1 + 1e-10, 3, 4)))
assert isinstance(t1, torch.Tensor)
assert (t1 == ttorch.tensor([True, False, False])).all()
t2 = ttorch.tensor({
'a': [1., 2, 3],
'b': {'x': [[float('inf'), 4, 1e20],
[-math.inf, 2.2943, 9483.32]]},
}).isclose(ttorch.tensor({
'a': [1 + 1e-10, 3, 4],
'b': {'x': [[math.inf, 6, 1e20 + 1],
[-float('inf'), 2.294300000001, 9484.32]]},
}))
assert (t2 == ttorch.tensor({
'a': [True, False, False],
'b': {'x': [[True, False, True],
[True, True, False]]},
})).all()
@choose_mark()
def test_abs(self):
t1 = ttorch.tensor([12, 0, -3]).abs()
......@@ -990,3 +1013,243 @@ class TestTorchTensor:
[-7, -2, -7]]],
}
}))
@choose_mark()
def test_exp(self):
t1 = ttorch.tensor([-4.0, -1.0, 0, 2.0, 4.8, 8.0]).exp()
assert isinstance(t1, torch.Tensor)
assert ttorch.isclose(t1, ttorch.tensor(
[1.8316e-02, 3.6788e-01, 1.0000e+00, 7.3891e+00, 1.2151e+02, 2.9810e+03]), rtol=1e-4).all()
t2 = ttorch.tensor({
'a': [-4.0, -1.0, 0, 2.0, 4.8, 8.0],
'b': {'x': [[-2.0, 1.2, 0.25],
[16.0, 3.75, -2.34]]},
}).exp()
assert ttorch.isclose(t2, ttorch.tensor({
'a': [1.8316e-02, 3.6788e-01, 1.0000e+00, 7.3891e+00, 1.2151e+02, 2.9810e+03],
'b': {'x': [[1.3534e-01, 3.3201e+00, 1.2840e+00],
[8.8861e+06, 4.2521e+01, 9.6328e-02]]},
}), rtol=1e-4).all()
@choose_mark()
def test_exp_(self):
t1 = ttorch.tensor([-4.0, -1.0, 0, 2.0, 4.8, 8.0])
t1r = t1.exp_()
assert t1r is t1
assert isinstance(t1, torch.Tensor)
assert ttorch.isclose(t1, ttorch.tensor(
[1.8316e-02, 3.6788e-01, 1.0000e+00, 7.3891e+00, 1.2151e+02, 2.9810e+03]), rtol=1e-4).all()
t2 = ttorch.tensor({
'a': [-4.0, -1.0, 0, 2.0, 4.8, 8.0],
'b': {'x': [[-2.0, 1.2, 0.25],
[16.0, 3.75, -2.34]]},
})
t2r = t2.exp_()
assert t2r is t2
assert ttorch.isclose(t2, ttorch.tensor({
'a': [1.8316e-02, 3.6788e-01, 1.0000e+00, 7.3891e+00, 1.2151e+02, 2.9810e+03],
'b': {'x': [[1.3534e-01, 3.3201e+00, 1.2840e+00],
[8.8861e+06, 4.2521e+01, 9.6328e-02]]},
}), rtol=1e-4).all()
@choose_mark()
def test_exp2(self):
t1 = ttorch.tensor([-4.0, -1.0, 0, 2.0, 4.8, 8.0]).exp2()
assert isinstance(t1, torch.Tensor)
assert ttorch.isclose(t1, ttorch.tensor(
[6.2500e-02, 5.0000e-01, 1.0000e+00, 4.0000e+00, 2.7858e+01, 2.5600e+02]), rtol=1e-4).all()
t2 = ttorch.tensor({
'a': [-4.0, -1.0, 0, 2.0, 4.8, 8.0],
'b': {'x': [[-2.0, 1.2, 0.25],
[16.0, 3.75, -2.34]]},
}).exp2()
assert ttorch.isclose(t2, ttorch.tensor({
'a': [6.2500e-02, 5.0000e-01, 1.0000e+00, 4.0000e+00, 2.7858e+01, 2.5600e+02],
'b': {'x': [[2.5000e-01, 2.2974e+00, 1.1892e+00],
[6.5536e+04, 1.3454e+01, 1.9751e-01]]},
}), rtol=1e-4).all()
@choose_mark()
def test_exp2_(self):
t1 = ttorch.tensor([-4.0, -1.0, 0, 2.0, 4.8, 8.0])
t1r = t1.exp2_()
assert t1r is t1
assert isinstance(t1, torch.Tensor)
assert ttorch.isclose(t1, ttorch.tensor(
[6.2500e-02, 5.0000e-01, 1.0000e+00, 4.0000e+00, 2.7858e+01, 2.5600e+02]), rtol=1e-4).all()
t2 = ttorch.tensor({
'a': [-4.0, -1.0, 0, 2.0, 4.8, 8.0],
'b': {'x': [[-2.0, 1.2, 0.25],
[16.0, 3.75, -2.34]]},
})
t2r = t2.exp2_()
assert t2r is t2
assert ttorch.isclose(t2, ttorch.tensor({
'a': [6.2500e-02, 5.0000e-01, 1.0000e+00, 4.0000e+00, 2.7858e+01, 2.5600e+02],
'b': {'x': [[2.5000e-01, 2.2974e+00, 1.1892e+00],
[6.5536e+04, 1.3454e+01, 1.9751e-01]]},
}), rtol=1e-4).all()
@choose_mark()
def test_sqrt(self):
t1 = ttorch.tensor([-4.0, -1.0, 0, 2.0, 4.8, 8.0]).sqrt()
assert isinstance(t1, torch.Tensor)
assert ttorch.isclose(t1, ttorch.tensor(
[math.nan, math.nan, 0.0000, 1.4142, 2.1909, 2.8284]), rtol=1e-4, equal_nan=True).all()
t2 = ttorch.tensor({
'a': [-4.0, -1.0, 0, 2.0, 4.8, 8.0],
'b': {'x': [[-2.0, 1.2, 0.25],
[16.0, 3.75, -2.34]]},
}).sqrt()
assert ttorch.isclose(t2, ttorch.tensor({
'a': [math.nan, math.nan, 0.0000, 1.4142, 2.1909, 2.8284],
'b': {'x': [[math.nan, 1.0954, 0.5000],
[4.0000, 1.9365, math.nan]]},
}), rtol=1e-4, equal_nan=True).all()
@choose_mark()
def test_sqrt_(self):
t1 = ttorch.tensor([-4.0, -1.0, 0, 2.0, 4.8, 8.0])
t1r = t1.sqrt_()
assert t1r is t1
assert isinstance(t1, torch.Tensor)
assert ttorch.isclose(t1, ttorch.tensor(
[math.nan, math.nan, 0.0000, 1.4142, 2.1909, 2.8284]), rtol=1e-4, equal_nan=True).all()
t2 = ttorch.tensor({
'a': [-4.0, -1.0, 0, 2.0, 4.8, 8.0],
'b': {'x': [[-2.0, 1.2, 0.25],
[16.0, 3.75, -2.34]]},
})
t2r = t2.sqrt_()
assert t2r is t2
assert ttorch.isclose(t2, ttorch.tensor({
'a': [math.nan, math.nan, 0.0000, 1.4142, 2.1909, 2.8284],
'b': {'x': [[math.nan, 1.0954, 0.5000],
[4.0000, 1.9365, math.nan]]},
}), rtol=1e-4, equal_nan=True).all()
@choose_mark()
def test_log(self):
t1 = ttorch.tensor([-4.0, -1.0, 0, 2.0, 4.8, 8.0]).log()
assert isinstance(t1, torch.Tensor)
assert ttorch.isclose(t1, ttorch.tensor(
[math.nan, math.nan, -math.inf, 0.6931, 1.5686, 2.0794]), rtol=1e-4, equal_nan=True).all()
t2 = ttorch.tensor({
'a': [-4.0, -1.0, 0, 2.0, 4.8, 8.0],
'b': {'x': [[-2.0, 1.2, 0.25],
[16.0, 3.75, -2.34]]},
}).log()
assert ttorch.isclose(t2, ttorch.tensor({
'a': [math.nan, math.nan, -math.inf, 0.6931, 1.5686, 2.0794],
'b': {'x': [[math.nan, 0.1823, -1.3863],
[2.7726, 1.3218, math.nan]]},
}), rtol=1e-4, atol=1e-4, equal_nan=True).all()
@choose_mark()
def test_log_(self):
t1 = ttorch.tensor([-4.0, -1.0, 0, 2.0, 4.8, 8.0])
t1r = t1.log_()
assert t1r is t1
assert isinstance(t1, torch.Tensor)
assert ttorch.isclose(t1, ttorch.tensor(
[math.nan, math.nan, -math.inf, 0.6931, 1.5686, 2.0794]), rtol=1e-4, equal_nan=True).all()
t2 = ttorch.tensor({
'a': [-4.0, -1.0, 0, 2.0, 4.8, 8.0],
'b': {'x': [[-2.0, 1.2, 0.25],
[16.0, 3.75, -2.34]]},
})
t2r = t2.log_()
assert t2r is t2
assert ttorch.isclose(t2, ttorch.tensor({
'a': [math.nan, math.nan, -math.inf, 0.6931, 1.5686, 2.0794],
'b': {'x': [[math.nan, 0.1823, -1.3863],
[2.7726, 1.3218, math.nan]]},
}), rtol=1e-4, atol=1e-4, equal_nan=True).all()
@choose_mark()
def test_log2(self):
t1 = ttorch.tensor([-4.0, -1.0, 0, 2.0, 4.8, 8.0]).log2()
assert isinstance(t1, torch.Tensor)
assert ttorch.isclose(t1, ttorch.tensor(
[math.nan, math.nan, -math.inf, 1.0000, 2.2630, 3.0000]), rtol=1e-4, equal_nan=True).all()
t2 = ttorch.tensor({
'a': [-4.0, -1.0, 0, 2.0, 4.8, 8.0],
'b': {'x': [[-2.0, 1.2, 0.25],
[16.0, 3.75, -2.34]]},
}).log2()
assert ttorch.isclose(t2, ttorch.tensor({
'a': [math.nan, math.nan, -math.inf, 1.0000, 2.2630, 3.0000],
'b': {'x': [[math.nan, 0.2630, -2.0000],
[4.0000, 1.9069, math.nan]]},
}), rtol=1e-4, atol=1e-4, equal_nan=True).all()
@choose_mark()
def test_log2_(self):
t1 = ttorch.tensor([-4.0, -1.0, 0, 2.0, 4.8, 8.0])
t1r = t1.log2_()
assert t1r is t1
assert isinstance(t1, torch.Tensor)
assert ttorch.isclose(t1, ttorch.tensor(
[math.nan, math.nan, -math.inf, 1.0000, 2.2630, 3.0000]), rtol=1e-4, equal_nan=True).all()
t2 = ttorch.tensor({
'a': [-4.0, -1.0, 0, 2.0, 4.8, 8.0],
'b': {'x': [[-2.0, 1.2, 0.25],
[16.0, 3.75, -2.34]]},
})
t2r = t2.log2_()
assert t2r is t2
assert ttorch.isclose(t2, ttorch.tensor({
'a': [math.nan, math.nan, -math.inf, 1.0000, 2.2630, 3.0000],
'b': {'x': [[math.nan, 0.2630, -2.0000],
[4.0000, 1.9069, math.nan]]},
}), rtol=1e-4, atol=1e-4, equal_nan=True).all()
@choose_mark()
def test_log10(self):
t1 = ttorch.tensor([-4.0, -1.0, 0, 2.0, 4.8, 8.0]).log10()
assert isinstance(t1, torch.Tensor)
assert ttorch.isclose(t1, ttorch.tensor(
[math.nan, math.nan, -math.inf, 0.3010, 0.6812, 0.9031]), rtol=1e-4, equal_nan=True).all()
t2 = ttorch.tensor({
'a': [-4.0, -1.0, 0, 2.0, 4.8, 8.0],
'b': {'x': [[-2.0, 1.2, 0.25],
[16.0, 3.75, -2.34]]},
}).log10()
assert ttorch.isclose(t2, ttorch.tensor({
'a': [math.nan, math.nan, -math.inf, 0.3010, 0.6812, 0.9031],
'b': {'x': [[math.nan, 0.0792, -0.6021],
[1.2041, 0.5740, math.nan]]},
}), rtol=1e-4, atol=1e-4, equal_nan=True).all()
@choose_mark()
def test_log10_(self):
t1 = ttorch.tensor([-4.0, -1.0, 0, 2.0, 4.8, 8.0])
t1r = t1.log10_()
assert t1r is t1
assert isinstance(t1, torch.Tensor)
assert ttorch.isclose(t1, ttorch.tensor(
[math.nan, math.nan, -math.inf, 0.3010, 0.6812, 0.9031]), rtol=1e-4, equal_nan=True).all()
t2 = ttorch.tensor({
'a': [-4.0, -1.0, 0, 2.0, 4.8, 8.0],
'b': {'x': [[-2.0, 1.2, 0.25],
[16.0, 3.75, -2.34]]},
})
t2r = t2.log10_()
assert t2r is t2
assert ttorch.isclose(t2, ttorch.tensor({
'a': [math.nan, math.nan, -math.inf, 0.3010, 0.6812, 0.9031],
'b': {'x': [[math.nan, 0.0792, -0.6021],
[1.2041, 0.5740, math.nan]]},
}), rtol=1e-4, atol=1e-4, equal_nan=True).all()
treetensor/torch/funcs.py
浏览文件 @ 0ba6d5f2
......@@ -24,7 +24,7 @@ __all__ = [
'eq', 'ne', 'lt', 'le', 'gt', 'ge',
'equal', 'tensor', 'clone',
'dot', 'matmul', 'mm',
'isfinite', 'isinf', 'isnan',
'isfinite', 'isinf', 'isnan', 'isclose',
'abs', 'abs_', 'clamp', 'clamp_', 'sign', 'sigmoid', 'sigmoid_',
'round', 'round_', 'floor', 'floor_', 'ceil', 'ceil_',
'add', 'sub', 'mul', 'div', 'pow', 'neg', 'neg_',
......@@ -1049,6 +1049,46 @@ def isnan(input):
return torch.isnan(input)
# noinspection PyShadowingBuiltins
@doc_from_base()
@func_treelize()
def isclose(input, other, *args, **kwargs):
"""
Returns a new tensor with boolean elements representing
if each element of ``input`` is “close” to the corresponding element of ``other``.
Examples::
>>> import torch
>>> import treetensor.torch as ttorch
>>> import math
>>> ttorch.isclose(
... ttorch.tensor((1., 2, 3)),
... ttorch.tensor((1 + 1e-10, 3, 4))
... )
tensor([ True, False, False])
>>> ttorch.isclose(
... ttorch.tensor({
... 'a': [1., 2, 3],
... 'b': {'x': [[float('inf'), 4, 1e20],
... [-math.inf, 2.2943, 9483.32]]},
... }),
... ttorch.tensor({
... 'a': [1 + 1e-10, 3, 4],
... 'b': {'x': [[math.inf, 6, 1e20+1],
... [-float('inf'), 2.294300000001, 9484.32]]},
... }),
... )
<Tensor 0x7f5b3219f370>
├── a --> tensor([ True, False, False])
└── b --> <Tensor 0x7f5b3219f550>
└── x --> tensor([[ True, False, True],
[ True, True, False]])
"""
return torch.isclose(input, other, *args, **kwargs)
# noinspection PyShadowingBuiltins
@doc_from_base()
@func_treelize()
......@@ -1749,76 +1789,370 @@ def neg_(input):
return torch.neg_(input)
# noinspection PyShadowingBuiltins
@doc_from_base()
@func_treelize()
def exp():
pass
def exp(input, *args, **kwargs):
"""
Returns a new tensor with the exponential of the elements of the input tensor ``input``.
Examples::
>>> import torch
>>> import treetensor.torch as ttorch
>>> ttorch.exp(ttorch.tensor([-4.0, -1.0, 0, 2.0, 4.8, 8.0]))
tensor([1.8316e-02, 3.6788e-01, 1.0000e+00, 7.3891e+00, 1.2151e+02, 2.9810e+03])
>>> ttorch.exp(ttorch.tensor({
... 'a': [-4.0, -1.0, 0, 2.0, 4.8, 8.0],
... 'b': {'x': [[-2.0, 1.2, 0.25],
... [16.0, 3.75, -2.34]]},
... }))
<Tensor 0x7ff90a4b0a30>
├── a --> tensor([1.8316e-02, 3.6788e-01, 1.0000e+00, 7.3891e+00, 1.2151e+02, 2.9810e+03])
└── b --> <Tensor 0x7ff90a4b0af0>
└── x --> tensor([[1.3534e-01, 3.3201e+00, 1.2840e+00],
[8.8861e+06, 4.2521e+01, 9.6328e-02]])
"""
return torch.exp(input, *args, **kwargs)
# noinspection PyShadowingBuiltins
@doc_from_base()
@return_self
@func_treelize()
def exp_():
pass
def exp_(input):
"""
In-place version of :func:`exp`.
Examples::
>>> import torch
>>> import treetensor.torch as ttorch
>>> t = ttorch.tensor([-4.0, -1.0, 0, 2.0, 4.8, 8.0])
>>> ttorch.exp_(t)
>>> t
tensor([1.8316e-02, 3.6788e-01, 1.0000e+00, 7.3891e+00, 1.2151e+02, 2.9810e+03])
>>> t = ttorch.tensor({
... 'a': [-4.0, -1.0, 0, 2.0, 4.8, 8.0],
... 'b': {'x': [[-2.0, 1.2, 0.25],
... [16.0, 3.75, -2.34]]},
... })
>>> ttorch.exp_(t)
>>> t
<Tensor 0x7ff90a4bdb80>
├── a --> tensor([1.8316e-02, 3.6788e-01, 1.0000e+00, 7.3891e+00, 1.2151e+02, 2.9810e+03])
└── b --> <Tensor 0x7ff90a4bdc40>
└── x --> tensor([[1.3534e-01, 3.3201e+00, 1.2840e+00],
[8.8861e+06, 4.2521e+01, 9.6328e-02]])
"""
return torch.exp_(input)
# noinspection PyShadowingBuiltins
@doc_from_base()
@func_treelize()
def exp2():
pass
def exp2(input, *args, **kwargs):
"""
Computes the base two exponential function of ``input``.
Examples::
>>> import torch
>>> import treetensor.torch as ttorch
>>> ttorch.exp2(ttorch.tensor([-4.0, -1.0, 0, 2.0, 4.8, 8.0]))
tensor([6.2500e-02, 5.0000e-01, 1.0000e+00, 4.0000e+00, 2.7858e+01, 2.5600e+02])
>>> ttorch.exp2(ttorch.tensor({
... 'a': [-4.0, -1.0, 0, 2.0, 4.8, 8.0],
... 'b': {'x': [[-2.0, 1.2, 0.25],
... [16.0, 3.75, -2.34]]},
... }))
<Tensor 0x7ff90a4c3af0>
├── a --> tensor([6.2500e-02, 5.0000e-01, 1.0000e+00, 4.0000e+00, 2.7858e+01, 2.5600e+02])
└── b --> <Tensor 0x7ff90a4c3be0>
└── x --> tensor([[2.5000e-01, 2.2974e+00, 1.1892e+00],
[6.5536e+04, 1.3454e+01, 1.9751e-01]])
"""
return torch.exp2(input, *args, **kwargs)
# noinspection PyShadowingBuiltins
@doc_from_base()
@return_self
@func_treelize()
def exp2_():
pass
def exp2_(input):
"""
In-place version of :func:`exp2`.
Examples::
>>> import torch
>>> import treetensor.torch as ttorch
>>> t = ttorch.tensor([-4.0, -1.0, 0, 2.0, 4.8, 8.0])
>>> ttorch.exp2_(t)
>>> t
tensor([6.2500e-02, 5.0000e-01, 1.0000e+00, 4.0000e+00, 2.7858e+01, 2.5600e+02])
>>> t = ttorch.tensor({
... 'a': [-4.0, -1.0, 0, 2.0, 4.8, 8.0],
... 'b': {'x': [[-2.0, 1.2, 0.25],
... [16.0, 3.75, -2.34]]},
... })
>>> ttorch.exp2_(t)
>>> t
<Tensor 0x7ff90a4bd250>
├── a --> tensor([6.2500e-02, 5.0000e-01, 1.0000e+00, 4.0000e+00, 2.7858e+01, 2.5600e+02])
└── b --> <Tensor 0x7ff90a4bd130>
└── x --> tensor([[2.5000e-01, 2.2974e+00, 1.1892e+00],
[6.5536e+04, 1.3454e+01, 1.9751e-01]])
"""
return torch.exp2_(input)
# noinspection PyShadowingBuiltins
@doc_from_base()
@func_treelize()
def sqrt():
pass
def sqrt(input, *args, **kwargs):
"""
Returns a new tensor with the square-root of the elements of ``input``.
Examples::
>>> import torch
>>> import treetensor.torch as ttorch
>>> ttorch.sqrt(ttorch.tensor([-4.0, -1.0, 0, 2.0, 4.8, 8.0]))
tensor([ nan, nan, 0.0000, 1.4142, 2.1909, 2.8284])
>>> ttorch.sqrt(ttorch.tensor({
... 'a': [-4.0, -1.0, 0, 2.0, 4.8, 8.0],
... 'b': {'x': [[-2.0, 1.2, 0.25],
... [16.0, 3.75, -2.34]]},
... }))
<Tensor 0x7ff90a4cb760>
├── a --> tensor([ nan, nan, 0.0000, 1.4142, 2.1909, 2.8284])
└── b --> <Tensor 0x7ff90a4cb5b0>
└── x --> tensor([[ nan, 1.0954, 0.5000],
[4.0000, 1.9365, nan]])
"""
return torch.sqrt(input, *args, **kwargs)
# noinspection PyShadowingBuiltins
@doc_from_base()
@return_self
@func_treelize()
def sqrt_():
pass
def sqrt_(input):
"""
In-place version of :func:`sqrt`.
Examples::
>>> import torch
>>> import treetensor.torch as ttorch
>>> t = ttorch.tensor([-4.0, -1.0, 0, 2.0, 4.8, 8.0])
>>> ttorch.sqrt_(t)
>>> t
tensor([ nan, nan, 0.0000, 1.4142, 2.1909, 2.8284])
>>> t = ttorch.tensor({
... 'a': [-4.0, -1.0, 0, 2.0, 4.8, 8.0],
... 'b': {'x': [[-2.0, 1.2, 0.25],
... [16.0, 3.75, -2.34]]},
... })
>>> ttorch.sqrt_(t)
>>> t
<Tensor 0x7ff90a4b0af0>
├── a --> tensor([ nan, nan, 0.0000, 1.4142, 2.1909, 2.8284])
└── b --> <Tensor 0x7ff90a4b04f0>
└── x --> tensor([[ nan, 1.0954, 0.5000],
[4.0000, 1.9365, nan]])
"""
return torch.sqrt_(input)
# noinspection PyShadowingBuiltins
@doc_from_base()
@func_treelize()
def log():
pass
def log(input, *args, **kwargs):
"""
Returns a new tensor with the natural logarithm of the elements of ``input``.
Examples::
>>> import torch
>>> import treetensor.torch as ttorch
>>> ttorch.log(ttorch.tensor([-4.0, -1.0, 0, 2.0, 4.8, 8.0]))
tensor([ nan, nan, -inf, 0.6931, 1.5686, 2.0794])
>>> ttorch.log(ttorch.tensor({
... 'a': [-4.0, -1.0, 0, 2.0, 4.8, 8.0],
... 'b': {'x': [[-2.0, 1.2, 0.25],
... [16.0, 3.75, -2.34]]},
... }))
<Tensor 0x7ff90a4c9ca0>
├── a --> tensor([ nan, nan, -inf, 0.6931, 1.5686, 2.0794])
└── b --> <Tensor 0x7ff90a4c9e50>
└── x --> tensor([[ nan, 0.1823, -1.3863],
[ 2.7726, 1.3218, nan]])
"""
return torch.log(input, *args, **kwargs)
# noinspection PyShadowingBuiltins
@doc_from_base()
@return_self
@func_treelize()
def log_():
pass
def log_(input):
"""
In-place version of :func:`log`.
Examples::
>>> import torch
>>> import treetensor.torch as ttorch
>>> t = ttorch.tensor([-4.0, -1.0, 0, 2.0, 4.8, 8.0])
>>> ttorch.log_(t)
>>> t
tensor([ nan, nan, -inf, 0.6931, 1.5686, 2.0794])
>>> t = ttorch.tensor({
... 'a': [-4.0, -1.0, 0, 2.0, 4.8, 8.0],
... 'b': {'x': [[-2.0, 1.2, 0.25],
... [16.0, 3.75, -2.34]]},
... })
>>> ttorch.log_(t)
>>> t
<Tensor 0x7ff90a4bdf70>
├── a --> tensor([ nan, nan, -inf, 0.6931, 1.5686, 2.0794])
└── b --> <Tensor 0x7ff90a4bdcd0>
└── x --> tensor([[ nan, 0.1823, -1.3863],
[ 2.7726, 1.3218, nan]])
"""
return torch.log_(input)
# noinspection PyShadowingBuiltins
@doc_from_base()
@func_treelize()
def log2():
pass
def log2(input, *args, **kwargs):
"""
Returns a new tensor with the logarithm to the base 2 of the elements of ``input``.
Examples::
>>> import torch
>>> import treetensor.torch as ttorch
>>> ttorch.log2(ttorch.tensor([-4.0, -1.0, 0, 2.0, 4.8, 8.0]))
tensor([ nan, nan, -inf, 1.0000, 2.2630, 3.0000])
>>> ttorch.log2(ttorch.tensor({
... 'a': [-4.0, -1.0, 0, 2.0, 4.8, 8.0],
... 'b': {'x': [[-2.0, 1.2, 0.25],
... [16.0, 3.75, -2.34]]},
... }))
<Tensor 0x7ff90a4cff70>
├── a --> tensor([ nan, nan, -inf, 1.0000, 2.2630, 3.0000])
└── b --> <Tensor 0x7ff90a4bc070>
└── x --> tensor([[ nan, 0.2630, -2.0000],
[ 4.0000, 1.9069, nan]])
"""
return torch.log2(input, *args, **kwargs)
# noinspection PyShadowingBuiltins
@doc_from_base()
@return_self
@func_treelize()
def log2_():
pass
def log2_(input):
"""
In-place version of :func:`log2`.
Examples::
>>> import torch
>>> import treetensor.torch as ttorch
>>> t = ttorch.tensor([-4.0, -1.0, 0, 2.0, 4.8, 8.0])
>>> ttorch.log2_(t)
>>> t
tensor([ nan, nan, -inf, 1.0000, 2.2630, 3.0000])
>>> t = ttorch.tensor({
... 'a': [-4.0, -1.0, 0, 2.0, 4.8, 8.0],
... 'b': {'x': [[-2.0, 1.2, 0.25],
... [16.0, 3.75, -2.34]]},
... })
>>> ttorch.log2_(t)
>>> t
<Tensor 0x7ff90a4cbbe0>
├── a --> tensor([ nan, nan, -inf, 1.0000, 2.2630, 3.0000])
└── b --> <Tensor 0x7ff90a4cb940>
└── x --> tensor([[ nan, 0.2630, -2.0000],
[ 4.0000, 1.9069, nan]])
"""
return torch.log2_(input)
# noinspection PyShadowingBuiltins
@doc_from_base()
@func_treelize()
def log10():
pass
def log10(input, *args, **kwargs):
"""
Returns a new tensor with the logarithm to the base 10 of the elements of ``input``.
Examples::
>>> import torch
>>> import treetensor.torch as ttorch
>>> ttorch.log10(ttorch.tensor([-4.0, -1.0, 0, 2.0, 4.8, 8.0]))
tensor([ nan, nan, -inf, 0.3010, 0.6812, 0.9031])
>>> ttorch.log10(ttorch.tensor({
... 'a': [-4.0, -1.0, 0, 2.0, 4.8, 8.0],
... 'b': {'x': [[-2.0, 1.2, 0.25],
... [16.0, 3.75, -2.34]]},
... }))
<Tensor 0x7ff90a4bc4f0>
├── a --> tensor([ nan, nan, -inf, 0.3010, 0.6812, 0.9031])
└── b --> <Tensor 0x7ff90a4bc5b0>
└── x --> tensor([[ nan, 0.0792, -0.6021],
[ 1.2041, 0.5740, nan]])
"""
return torch.log10(input, *args, **kwargs)
# noinspection PyShadowingBuiltins
@doc_from_base()
@return_self
@func_treelize()
def log10_():
pass
def log10_(input):
"""
In-place version of :func:`log10`.
Examples::
>>> import torch
>>> import treetensor.torch as ttorch
>>> t = ttorch.tensor([-4.0, -1.0, 0, 2.0, 4.8, 8.0])
>>> ttorch.log10_(t)
>>> t
tensor([ nan, nan, -inf, 0.3010, 0.6812, 0.9031])
>>> t = ttorch.tensor({
... 'a': [-4.0, -1.0, 0, 2.0, 4.8, 8.0],
... 'b': {'x': [[-2.0, 1.2, 0.25],
... [16.0, 3.75, -2.34]]},
... })
>>> ttorch.log10_(t)
>>> t
<Tensor 0x7ff90a4acdc0>
├── a --> tensor([ nan, nan, -inf, 0.3010, 0.6812, 0.9031])
└── b --> <Tensor 0x7ff90a4acf40>
└── x --> tensor([[ nan, 0.0792, -0.6021],
[ 1.2041, 0.5740, nan]])
"""
return torch.log10_(input)
sys.modules[__name__] = module_autoremove(sys.modules[__name__])
treetensor/torch/tensor.py
浏览文件 @ 0ba6d5f2
......@@ -321,6 +321,14 @@ class Tensor(Torch, metaclass=clsmeta(_to_tensor, allow_dict=True)):
"""
return self.isnan()
@doc_from_base()
@method_treelize()
def isclose(self, other, *args, **kwargs):
"""
See :func:`treetensor.torch.isclose`.
"""
return self.isclose(other, *args, **kwargs)
@doc_from_base()
@method_treelize()
def abs(self, *args, **kwargs):
......@@ -541,3 +549,105 @@ class Tensor(Torch, metaclass=clsmeta(_to_tensor, allow_dict=True)):
In-place version of :meth:`Tensor.neg`.
"""
return self.neg_(*args, **kwargs)
@doc_from_base()
@method_treelize()
def exp(self, *args, **kwargs):
"""
See :func:`treetensor.torch.exp`.
"""
return self.exp(*args, **kwargs)
@doc_from_base()
@return_self
@method_treelize()
def exp_(self, *args, **kwargs):
"""
In-place version of :meth:`Tensor.exp`.
"""
return self.exp_(*args, **kwargs)
@doc_from_base()
@method_treelize()
def exp2(self, *args, **kwargs):
"""
See :func:`treetensor.torch.exp2`.
"""
return self.exp2(*args, **kwargs)
@doc_from_base()
@return_self
@method_treelize()
def exp2_(self, *args, **kwargs):
"""
In-place version of :meth:`Tensor.exp2`.
"""
return self.exp2_(*args, **kwargs)
@doc_from_base()
@method_treelize()
def sqrt(self, *args, **kwargs):
"""
See :func:`treetensor.torch.sqrt`.
"""
return self.sqrt(*args, **kwargs)
@doc_from_base()
@return_self
@method_treelize()
def sqrt_(self, *args, **kwargs):
"""
In-place version of :meth:`Tensor.sqrt`.
"""
return self.sqrt_(*args, **kwargs)
@doc_from_base()
@method_treelize()
def log(self, *args, **kwargs):
"""
See :func:`treetensor.torch.log`.
"""
return self.log(*args, **kwargs)
@doc_from_base()
@return_self
@method_treelize()
def log_(self, *args, **kwargs):
"""
In-place version of :meth:`Tensor.log`.
"""
return self.log_(*args, **kwargs)
@doc_from_base()
@method_treelize()
def log2(self, *args, **kwargs):
"""
See :func:`treetensor.torch.log2`.
"""
return self.log2(*args, **kwargs)
@doc_from_base()
@return_self
@method_treelize()
def log2_(self, *args, **kwargs):
"""
In-place version of :meth:`Tensor.log2`.
"""
return self.log2_(*args, **kwargs)
@doc_from_base()
@method_treelize()
def log10(self, *args, **kwargs):
"""
See :func:`treetensor.torch.log10`.
"""
return self.log10(*args, **kwargs)
@doc_from_base()
@return_self
@method_treelize()
def log10_(self, *args, **kwargs):
"""
In-place version of :meth:`Tensor.log10`.
"""
return self.log10_(*args, **kwargs)
Markdown is supported
0% 或 .
You are about to add 0 people to the discussion. Proceed with caution.
先完成此消息的编辑！
想要评论请 注册 或