下载
中文
注册
我要评分
文档获取效率
文档正确性
内容完整性
文档易理解
在线提单
论坛求助
昇腾小AI

torch

Tensors

表1 Tensors

序号

API名称

是否支持

1

arange

2

as_strided

3

as_tensor

4

cat

5

chunk

6

column_stack

7

complex

8

dequantize

9

dstack

10

empty

11

empty_like

12

empty_strided

13

eye

14

from_numpy

15

full

16

full_like

17

gather

18

get_default_dtype

是,但是不支持NPU数据类型

19

heaviside

20

hstack

21

index_select

22

is_complex

是,支持判断,但当前硬件限制不支持复数

23

is_floating_point

24

is_nonzero

25

is_storage

26

is_tensor

27

linspace

28

logspace

29

masked_select

30

moveaxis

31

movedim

32

narrow

33

nonzero

34

numel

35

ones

36

ones_like

37

polar

38

quantize_per_channel

39

quantize_per_tensor

40

range

41

reshape

42

row_stack

43

scatter

44

scatter_add

45

set_default_dtype

是,但是不支持NPU数据类型

46

set_default_tensor_type

是,但是不支持NPU数据类型

47

set_flush_denormal

48

set_printoptions

49

sparse_coo_tensor

50

split

51

squeeze

52

stack

53

swapaxes

54

swapdims

55

t

56

take

57

tensor

58

tensor_split

59

tile

60

transpose

61

unbind

62

unsqueeze

63

vstack

64

where

65

zeros

66

zeros_like

Generators

表2 Generators

序号

API名称

是否支持

1

Generator

是,但不支持get_state和set_state接口

Random sampling

表3 Random sampling

序号

API名称

是否支持

1

bernoulli

2

get_rng_state

3

initial_seed

4

manual_seed

5

multinomial

6

normal

7

poisson

8

quasirandom.SobolEngine

9

rand

10

rand_like

11

randint

12

randint_like

13

randn

14

randn_like

15

randperm

16

seed

17

set_rng_state

18

torch.default_generator

19

torch.Tensor.bernoulli_

20

torch.Tensor.cauchy_

21

torch.Tensor.exponential_

22

torch.Tensor.geometric_

23

torch.Tensor.log_normal_

24

torch.Tensor.normal_

25

torch.Tensor.random_

26

torch.Tensor.uniform_

Serialization

表4 Serialization

序号

API名称

是否支持

1

load

2

save

Parallelism

表5 Parallelism

序号

API名称

是否支持

1

get_num_interop_threads

2

get_num_threads

3

set_num_interop_threads

4

set_num_threads

Locally disabling gradient computation

表6 Locally disabling gradient computation

序号

API名称

是否支持

1

enable_grad

2

no_grad

3

set_grad_enabled

Math operations

表7 Math operations

序号

API名称

是否支持

1

abs

2

absolute

3

acos

4

acosh

5

add

6

addbmm

7

addcdiv

8

addcmul

9

addmm

10

addmv

11

addr

12

all

13

allclose

14

amax

15

amin

16

angle

17

any

18

arccos

19

arccosh

20

arcsin

21

arcsinh

22

arctan

23

arctanh

24

argmax

25

argmin

26

argsort

27

asin

28

asinh

29

atan

30

atan2

31

atanh

32

atleast_1d

33

atleast_2d

34

atleast_3d

35

baddbmm

36

bartlett_window

37

bincount

38

bitwise_and

39

bitwise_not

40

bitwise_or

41

bitwise_xor

42

blackman_window

43

block_diag

44

bmm

45

broadcast_shapes

46

broadcast_tensors

47

broadcast_to

48

bucketize

49

cartesian_prod

50

cdist

是,仅支持mode=donot_use_mm_for_euclid_dist

51

ceil

52

chain_matmul

53

cholesky

54

cholesky_inverse

55

cholesky_solve

56

clamp

57

clip

58

clone

59

combinations

60

conj

61

copysign

62

cos

63

cosh

64

count_nonzero

65

cross

66

cummax

67

cummin

68

cumprod

69

cumsum

70

deg2rad

71

det

72

diag

是,仅支持diagonal=0场景

73

diag_embed

74

diagflat

75

diagonal

76

diff

77

digamma

78

dist

79

div

80

divide

81

dot

82

eig

83

einsum

84

eq

85

equal

86

erf

87

erfc

88

erfinv

89

exp

90

exp2

91

expm1

92

fake_quantize_per_channel_affine

93

fake_quantize_per_tensor_affine

94

fix

95

flatten

96

flip

97

fliplr

98

flipud

99

float_power

100

floor

101

floor_divide

102

fmax

103

fmin

104

fmod

105

frac

106

gcd

107

ge

108

geqrf

109

ger

110

greater

111

greater_equal

112

gt

113

hamming_window

114

hann_window

115

histc

116

hypot

117

i0

118

igamma

119

igammac

120

imag

121

inner

122

inverse

123

isclose

124

isfinite

125

isinf

126

isnan

127

isneginf

128

isposinf

129

isreal

130

istft

131

kaiser_window

132

kron

133

kthvalue

134

lcm

135

ldexp

136

le

137

lerp

138

less

139

less_equal

140

lgamma

141

lobpcg

142

log

143

log10

144

log1p

145

log2

146

logaddexp

147

logaddexp2

148

logcumsumexp

149

logdet

150

logical_and

151

logical_not

152

logical_or

153

logical_xor

154

logit

155

logsumexp

156

lstsq

157

lt

158

lu

159

lu_solve

160

lu_unpack

161

matmul

162

matrix_exp

163

matrix_power

164

matrix_rank

165

max

166

maximum

167

mean

168

median

169

meshgrid

170

min

171

minimum

172

mm

173

mode

174

msort

175

mul

176

multiply

177

mv

178

mvlgamma

179

nan_to_num

180

nanmedian

181

nanquantile

182

nansum

183

ne

184

neg

185

negative

186

nextafter

187

norm

188

not_equal

189

orgqr

190

ormqr

191

outer

192

pca_lowrank

193

pinverse

194

polygamma

195

pow

196

prod

197

qr

198

quantile

199

rad2deg

200

ravel

201

real

202

reciprocal

203

remainder

204

renorm

205

repeat_interleave

206

roll

207

rot90

208

round

209

rsqrt

210

searchsorted

211

sgn

212

sigmoid

213

sign

214

signbit

215

sin

216

sinc

217

sinh

218

slogdet

219

solve

220

sort

221

sqrt

222

square

223

std

224

std_mean

225

stft

226

sub

227

subtract

228

sum

229

svd

230

svd_lowrank

231

symeig

232

tan

233

tanh

234

tensordot

235

topk

236

trace

237

trapz

238

triangular_solve

239

tril

240

tril_indices

241

triu

242

triu_indices

243

true_divide

244

trunc

245

unique

246

unique_consecutive

是,传参时必须使用关键字,否则精度不达标。return_inverse=return_inverse,return_counts=return_counts,dim=dim

247

vander

248

var

249

var_mean

250

vdot

251

view_as_complex

252

view_as_real

253

xlogy

Utilities

表8 Utilities

序号

API名称

是否支持

1

compiled_with_cxx11_abi

2

result_type

3

can_cast

4

promote_types

5

use_deterministic_algorithms

6

are_deterministic_algorithms_enabled

7

_assert

搜索结果
找到“0”个结果

当前产品无相关内容

未找到相关内容,请尝试其他搜索词