Spaces:
Running
Running
File size: 56,420 Bytes
c61ccee |
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 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 |
import sys
import torch
from torch._C import _add_docstr, _fft # type: ignore[attr-defined]
from torch._torch_docs import factory_common_args, common_args
__all__ = ['fft', 'ifft', 'fft2', 'ifft2', 'fftn', 'ifftn',
'rfft', 'irfft', 'rfft2', 'irfft2', 'rfftn', 'irfftn',
'hfft', 'ihfft', 'fftfreq', 'rfftfreq', 'fftshift', 'ifftshift',
'Tensor']
Tensor = torch.Tensor
# Note: This not only adds the doc strings for the spectral ops, but
# connects the torch.fft Python namespace to the torch._C._fft builtins.
fft = _add_docstr(_fft.fft_fft, r"""
fft(input, n=None, dim=-1, norm=None, *, out=None) -> Tensor
Computes the one dimensional discrete Fourier transform of :attr:`input`.
Note:
The Fourier domain representation of any real signal satisfies the
Hermitian property: `X[i] = conj(X[-i])`. This function always returns both
the positive and negative frequency terms even though, for real inputs, the
negative frequencies are redundant. :func:`~torch.fft.rfft` returns the
more compact one-sided representation where only the positive frequencies
are returned.
Note:
Supports torch.half and torch.chalf on CUDA with GPU Architecture SM53 or greater.
However it only supports powers of 2 signal length in every transformed dimension.
Args:
input (Tensor): the input tensor
n (int, optional): Signal length. If given, the input will either be zero-padded
or trimmed to this length before computing the FFT.
dim (int, optional): The dimension along which to take the one dimensional FFT.
norm (str, optional): Normalization mode. For the forward transform
(:func:`~torch.fft.fft`), these correspond to:
* ``"forward"`` - normalize by ``1/n``
* ``"backward"`` - no normalization
* ``"ortho"`` - normalize by ``1/sqrt(n)`` (making the FFT orthonormal)
Calling the backward transform (:func:`~torch.fft.ifft`) with the same
normalization mode will apply an overall normalization of ``1/n`` between
the two transforms. This is required to make :func:`~torch.fft.ifft`
the exact inverse.
Default is ``"backward"`` (no normalization).
Keyword args:
{out}
Example:
>>> t = torch.arange(4)
>>> t
tensor([0, 1, 2, 3])
>>> torch.fft.fft(t)
tensor([ 6.+0.j, -2.+2.j, -2.+0.j, -2.-2.j])
>>> t = torch.tensor([0.+1.j, 2.+3.j, 4.+5.j, 6.+7.j])
>>> torch.fft.fft(t)
tensor([12.+16.j, -8.+0.j, -4.-4.j, 0.-8.j])
""".format(**common_args))
ifft = _add_docstr(_fft.fft_ifft, r"""
ifft(input, n=None, dim=-1, norm=None, *, out=None) -> Tensor
Computes the one dimensional inverse discrete Fourier transform of :attr:`input`.
Note:
Supports torch.half and torch.chalf on CUDA with GPU Architecture SM53 or greater.
However it only supports powers of 2 signal length in every transformed dimension.
Args:
input (Tensor): the input tensor
n (int, optional): Signal length. If given, the input will either be zero-padded
or trimmed to this length before computing the IFFT.
dim (int, optional): The dimension along which to take the one dimensional IFFT.
norm (str, optional): Normalization mode. For the backward transform
(:func:`~torch.fft.ifft`), these correspond to:
* ``"forward"`` - no normalization
* ``"backward"`` - normalize by ``1/n``
* ``"ortho"`` - normalize by ``1/sqrt(n)`` (making the IFFT orthonormal)
Calling the forward transform (:func:`~torch.fft.fft`) with the same
normalization mode will apply an overall normalization of ``1/n`` between
the two transforms. This is required to make :func:`~torch.fft.ifft`
the exact inverse.
Default is ``"backward"`` (normalize by ``1/n``).
Keyword args:
{out}
Example:
>>> t = torch.tensor([ 6.+0.j, -2.+2.j, -2.+0.j, -2.-2.j])
>>> torch.fft.ifft(t)
tensor([0.+0.j, 1.+0.j, 2.+0.j, 3.+0.j])
""".format(**common_args))
fft2 = _add_docstr(_fft.fft_fft2, r"""
fft2(input, s=None, dim=(-2, -1), norm=None, *, out=None) -> Tensor
Computes the 2 dimensional discrete Fourier transform of :attr:`input`.
Equivalent to :func:`~torch.fft.fftn` but FFTs only the last two dimensions by default.
Note:
The Fourier domain representation of any real signal satisfies the
Hermitian property: ``X[i, j] = conj(X[-i, -j])``. This
function always returns all positive and negative frequency terms even
though, for real inputs, half of these values are redundant.
:func:`~torch.fft.rfft2` returns the more compact one-sided representation
where only the positive frequencies of the last dimension are returned.
Note:
Supports torch.half and torch.chalf on CUDA with GPU Architecture SM53 or greater.
However it only supports powers of 2 signal length in every transformed dimensions.
Args:
input (Tensor): the input tensor
s (Tuple[int], optional): Signal size in the transformed dimensions.
If given, each dimension ``dim[i]`` will either be zero-padded or
trimmed to the length ``s[i]`` before computing the FFT.
If a length ``-1`` is specified, no padding is done in that dimension.
Default: ``s = [input.size(d) for d in dim]``
dim (Tuple[int], optional): Dimensions to be transformed.
Default: last two dimensions.
norm (str, optional): Normalization mode. For the forward transform
(:func:`~torch.fft.fft2`), these correspond to:
* ``"forward"`` - normalize by ``1/n``
* ``"backward"`` - no normalization
* ``"ortho"`` - normalize by ``1/sqrt(n)`` (making the FFT orthonormal)
Where ``n = prod(s)`` is the logical FFT size.
Calling the backward transform (:func:`~torch.fft.ifft2`) with the same
normalization mode will apply an overall normalization of ``1/n``
between the two transforms. This is required to make
:func:`~torch.fft.ifft2` the exact inverse.
Default is ``"backward"`` (no normalization).
Keyword args:
{out}
Example:
>>> x = torch.rand(10, 10, dtype=torch.complex64)
>>> fft2 = torch.fft.fft2(x)
The discrete Fourier transform is separable, so :func:`~torch.fft.fft2`
here is equivalent to two one-dimensional :func:`~torch.fft.fft` calls:
>>> two_ffts = torch.fft.fft(torch.fft.fft(x, dim=0), dim=1)
>>> torch.testing.assert_close(fft2, two_ffts, check_stride=False)
""".format(**common_args))
ifft2 = _add_docstr(_fft.fft_ifft2, r"""
ifft2(input, s=None, dim=(-2, -1), norm=None, *, out=None) -> Tensor
Computes the 2 dimensional inverse discrete Fourier transform of :attr:`input`.
Equivalent to :func:`~torch.fft.ifftn` but IFFTs only the last two dimensions by default.
Note:
Supports torch.half and torch.chalf on CUDA with GPU Architecture SM53 or greater.
However it only supports powers of 2 signal length in every transformed dimensions.
Args:
input (Tensor): the input tensor
s (Tuple[int], optional): Signal size in the transformed dimensions.
If given, each dimension ``dim[i]`` will either be zero-padded or
trimmed to the length ``s[i]`` before computing the IFFT.
If a length ``-1`` is specified, no padding is done in that dimension.
Default: ``s = [input.size(d) for d in dim]``
dim (Tuple[int], optional): Dimensions to be transformed.
Default: last two dimensions.
norm (str, optional): Normalization mode. For the backward transform
(:func:`~torch.fft.ifft2`), these correspond to:
* ``"forward"`` - no normalization
* ``"backward"`` - normalize by ``1/n``
* ``"ortho"`` - normalize by ``1/sqrt(n)`` (making the IFFT orthonormal)
Where ``n = prod(s)`` is the logical IFFT size.
Calling the forward transform (:func:`~torch.fft.fft2`) with the same
normalization mode will apply an overall normalization of ``1/n`` between
the two transforms. This is required to make :func:`~torch.fft.ifft2`
the exact inverse.
Default is ``"backward"`` (normalize by ``1/n``).
Keyword args:
{out}
Example:
>>> x = torch.rand(10, 10, dtype=torch.complex64)
>>> ifft2 = torch.fft.ifft2(x)
The discrete Fourier transform is separable, so :func:`~torch.fft.ifft2`
here is equivalent to two one-dimensional :func:`~torch.fft.ifft` calls:
>>> two_iffts = torch.fft.ifft(torch.fft.ifft(x, dim=0), dim=1)
>>> torch.testing.assert_close(ifft2, two_iffts, check_stride=False)
""".format(**common_args))
fftn = _add_docstr(_fft.fft_fftn, r"""
fftn(input, s=None, dim=None, norm=None, *, out=None) -> Tensor
Computes the N dimensional discrete Fourier transform of :attr:`input`.
Note:
The Fourier domain representation of any real signal satisfies the
Hermitian property: ``X[i_1, ..., i_n] = conj(X[-i_1, ..., -i_n])``. This
function always returns all positive and negative frequency terms even
though, for real inputs, half of these values are redundant.
:func:`~torch.fft.rfftn` returns the more compact one-sided representation
where only the positive frequencies of the last dimension are returned.
Note:
Supports torch.half and torch.chalf on CUDA with GPU Architecture SM53 or greater.
However it only supports powers of 2 signal length in every transformed dimensions.
Args:
input (Tensor): the input tensor
s (Tuple[int], optional): Signal size in the transformed dimensions.
If given, each dimension ``dim[i]`` will either be zero-padded or
trimmed to the length ``s[i]`` before computing the FFT.
If a length ``-1`` is specified, no padding is done in that dimension.
Default: ``s = [input.size(d) for d in dim]``
dim (Tuple[int], optional): Dimensions to be transformed.
Default: all dimensions, or the last ``len(s)`` dimensions if :attr:`s` is given.
norm (str, optional): Normalization mode. For the forward transform
(:func:`~torch.fft.fftn`), these correspond to:
* ``"forward"`` - normalize by ``1/n``
* ``"backward"`` - no normalization
* ``"ortho"`` - normalize by ``1/sqrt(n)`` (making the FFT orthonormal)
Where ``n = prod(s)`` is the logical FFT size.
Calling the backward transform (:func:`~torch.fft.ifftn`) with the same
normalization mode will apply an overall normalization of ``1/n``
between the two transforms. This is required to make
:func:`~torch.fft.ifftn` the exact inverse.
Default is ``"backward"`` (no normalization).
Keyword args:
{out}
Example:
>>> x = torch.rand(10, 10, dtype=torch.complex64)
>>> fftn = torch.fft.fftn(x)
The discrete Fourier transform is separable, so :func:`~torch.fft.fftn`
here is equivalent to two one-dimensional :func:`~torch.fft.fft` calls:
>>> two_ffts = torch.fft.fft(torch.fft.fft(x, dim=0), dim=1)
>>> torch.testing.assert_close(fftn, two_ffts, check_stride=False)
""".format(**common_args))
ifftn = _add_docstr(_fft.fft_ifftn, r"""
ifftn(input, s=None, dim=None, norm=None, *, out=None) -> Tensor
Computes the N dimensional inverse discrete Fourier transform of :attr:`input`.
Note:
Supports torch.half and torch.chalf on CUDA with GPU Architecture SM53 or greater.
However it only supports powers of 2 signal length in every transformed dimensions.
Args:
input (Tensor): the input tensor
s (Tuple[int], optional): Signal size in the transformed dimensions.
If given, each dimension ``dim[i]`` will either be zero-padded or
trimmed to the length ``s[i]`` before computing the IFFT.
If a length ``-1`` is specified, no padding is done in that dimension.
Default: ``s = [input.size(d) for d in dim]``
dim (Tuple[int], optional): Dimensions to be transformed.
Default: all dimensions, or the last ``len(s)`` dimensions if :attr:`s` is given.
norm (str, optional): Normalization mode. For the backward transform
(:func:`~torch.fft.ifftn`), these correspond to:
* ``"forward"`` - no normalization
* ``"backward"`` - normalize by ``1/n``
* ``"ortho"`` - normalize by ``1/sqrt(n)`` (making the IFFT orthonormal)
Where ``n = prod(s)`` is the logical IFFT size.
Calling the forward transform (:func:`~torch.fft.fftn`) with the same
normalization mode will apply an overall normalization of ``1/n`` between
the two transforms. This is required to make :func:`~torch.fft.ifftn`
the exact inverse.
Default is ``"backward"`` (normalize by ``1/n``).
Keyword args:
{out}
Example:
>>> x = torch.rand(10, 10, dtype=torch.complex64)
>>> ifftn = torch.fft.ifftn(x)
The discrete Fourier transform is separable, so :func:`~torch.fft.ifftn`
here is equivalent to two one-dimensional :func:`~torch.fft.ifft` calls:
>>> two_iffts = torch.fft.ifft(torch.fft.ifft(x, dim=0), dim=1)
>>> torch.testing.assert_close(ifftn, two_iffts, check_stride=False)
""".format(**common_args))
rfft = _add_docstr(_fft.fft_rfft, r"""
rfft(input, n=None, dim=-1, norm=None, *, out=None) -> Tensor
Computes the one dimensional Fourier transform of real-valued :attr:`input`.
The FFT of a real signal is Hermitian-symmetric, ``X[i] = conj(X[-i])`` so
the output contains only the positive frequencies below the Nyquist frequency.
To compute the full output, use :func:`~torch.fft.fft`
Note:
Supports torch.half on CUDA with GPU Architecture SM53 or greater.
However it only supports powers of 2 signal length in every transformed dimension.
Args:
input (Tensor): the real input tensor
n (int, optional): Signal length. If given, the input will either be zero-padded
or trimmed to this length before computing the real FFT.
dim (int, optional): The dimension along which to take the one dimensional real FFT.
norm (str, optional): Normalization mode. For the forward transform
(:func:`~torch.fft.rfft`), these correspond to:
* ``"forward"`` - normalize by ``1/n``
* ``"backward"`` - no normalization
* ``"ortho"`` - normalize by ``1/sqrt(n)`` (making the FFT orthonormal)
Calling the backward transform (:func:`~torch.fft.irfft`) with the same
normalization mode will apply an overall normalization of ``1/n`` between
the two transforms. This is required to make :func:`~torch.fft.irfft`
the exact inverse.
Default is ``"backward"`` (no normalization).
Keyword args:
{out}
Example:
>>> t = torch.arange(4)
>>> t
tensor([0, 1, 2, 3])
>>> torch.fft.rfft(t)
tensor([ 6.+0.j, -2.+2.j, -2.+0.j])
Compare against the full output from :func:`~torch.fft.fft`:
>>> torch.fft.fft(t)
tensor([ 6.+0.j, -2.+2.j, -2.+0.j, -2.-2.j])
Notice that the symmetric element ``T[-1] == T[1].conj()`` is omitted.
At the Nyquist frequency ``T[-2] == T[2]`` is it's own symmetric pair,
and therefore must always be real-valued.
""".format(**common_args))
irfft = _add_docstr(_fft.fft_irfft, r"""
irfft(input, n=None, dim=-1, norm=None, *, out=None) -> Tensor
Computes the inverse of :func:`~torch.fft.rfft`.
:attr:`input` is interpreted as a one-sided Hermitian signal in the Fourier
domain, as produced by :func:`~torch.fft.rfft`. By the Hermitian property, the
output will be real-valued.
Note:
Some input frequencies must be real-valued to satisfy the Hermitian
property. In these cases the imaginary component will be ignored.
For example, any imaginary component in the zero-frequency term cannot
be represented in a real output and so will always be ignored.
Note:
The correct interpretation of the Hermitian input depends on the length of
the original data, as given by :attr:`n`. This is because each input shape
could correspond to either an odd or even length signal. By default, the
signal is assumed to be even length and odd signals will not round-trip
properly. So, it is recommended to always pass the signal length :attr:`n`.
Note:
Supports torch.half and torch.chalf on CUDA with GPU Architecture SM53 or greater.
However it only supports powers of 2 signal length in every transformed dimension.
With default arguments, size of the transformed dimension should be (2^n + 1) as argument
`n` defaults to even output size = 2 * (transformed_dim_size - 1)
Args:
input (Tensor): the input tensor representing a half-Hermitian signal
n (int, optional): Output signal length. This determines the length of the
output signal. If given, the input will either be zero-padded or trimmed to this
length before computing the real IFFT.
Defaults to even output: ``n=2*(input.size(dim) - 1)``.
dim (int, optional): The dimension along which to take the one dimensional real IFFT.
norm (str, optional): Normalization mode. For the backward transform
(:func:`~torch.fft.irfft`), these correspond to:
* ``"forward"`` - no normalization
* ``"backward"`` - normalize by ``1/n``
* ``"ortho"`` - normalize by ``1/sqrt(n)`` (making the real IFFT orthonormal)
Calling the forward transform (:func:`~torch.fft.rfft`) with the same
normalization mode will apply an overall normalization of ``1/n`` between
the two transforms. This is required to make :func:`~torch.fft.irfft`
the exact inverse.
Default is ``"backward"`` (normalize by ``1/n``).
Keyword args:
{out}
Example:
>>> t = torch.linspace(0, 1, 5)
>>> t
tensor([0.0000, 0.2500, 0.5000, 0.7500, 1.0000])
>>> T = torch.fft.rfft(t)
>>> T
tensor([ 2.5000+0.0000j, -0.6250+0.8602j, -0.6250+0.2031j])
Without specifying the output length to :func:`~torch.fft.irfft`, the output
will not round-trip properly because the input is odd-length:
>>> torch.fft.irfft(T)
tensor([0.1562, 0.3511, 0.7812, 1.2114])
So, it is recommended to always pass the signal length :attr:`n`:
>>> roundtrip = torch.fft.irfft(T, t.numel())
>>> torch.testing.assert_close(roundtrip, t, check_stride=False)
""".format(**common_args))
rfft2 = _add_docstr(_fft.fft_rfft2, r"""
rfft2(input, s=None, dim=(-2, -1), norm=None, *, out=None) -> Tensor
Computes the 2-dimensional discrete Fourier transform of real :attr:`input`.
Equivalent to :func:`~torch.fft.rfftn` but FFTs only the last two dimensions by default.
The FFT of a real signal is Hermitian-symmetric, ``X[i, j] = conj(X[-i, -j])``,
so the full :func:`~torch.fft.fft2` output contains redundant information.
:func:`~torch.fft.rfft2` instead omits the negative frequencies in the last
dimension.
Note:
Supports torch.half on CUDA with GPU Architecture SM53 or greater.
However it only supports powers of 2 signal length in every transformed dimensions.
Args:
input (Tensor): the input tensor
s (Tuple[int], optional): Signal size in the transformed dimensions.
If given, each dimension ``dim[i]`` will either be zero-padded or
trimmed to the length ``s[i]`` before computing the real FFT.
If a length ``-1`` is specified, no padding is done in that dimension.
Default: ``s = [input.size(d) for d in dim]``
dim (Tuple[int], optional): Dimensions to be transformed.
Default: last two dimensions.
norm (str, optional): Normalization mode. For the forward transform
(:func:`~torch.fft.rfft2`), these correspond to:
* ``"forward"`` - normalize by ``1/n``
* ``"backward"`` - no normalization
* ``"ortho"`` - normalize by ``1/sqrt(n)`` (making the real FFT orthonormal)
Where ``n = prod(s)`` is the logical FFT size.
Calling the backward transform (:func:`~torch.fft.irfft2`) with the same
normalization mode will apply an overall normalization of ``1/n`` between
the two transforms. This is required to make :func:`~torch.fft.irfft2`
the exact inverse.
Default is ``"backward"`` (no normalization).
Keyword args:
{out}
Example:
>>> t = torch.rand(10, 10)
>>> rfft2 = torch.fft.rfft2(t)
>>> rfft2.size()
torch.Size([10, 6])
Compared against the full output from :func:`~torch.fft.fft2`, we have all
elements up to the Nyquist frequency.
>>> fft2 = torch.fft.fft2(t)
>>> torch.testing.assert_close(fft2[..., :6], rfft2, check_stride=False)
The discrete Fourier transform is separable, so :func:`~torch.fft.rfft2`
here is equivalent to a combination of :func:`~torch.fft.fft` and
:func:`~torch.fft.rfft`:
>>> two_ffts = torch.fft.fft(torch.fft.rfft(t, dim=1), dim=0)
>>> torch.testing.assert_close(rfft2, two_ffts, check_stride=False)
""".format(**common_args))
irfft2 = _add_docstr(_fft.fft_irfft2, r"""
irfft2(input, s=None, dim=(-2, -1), norm=None, *, out=None) -> Tensor
Computes the inverse of :func:`~torch.fft.rfft2`.
Equivalent to :func:`~torch.fft.irfftn` but IFFTs only the last two dimensions by default.
:attr:`input` is interpreted as a one-sided Hermitian signal in the Fourier
domain, as produced by :func:`~torch.fft.rfft2`. By the Hermitian property, the
output will be real-valued.
Note:
Some input frequencies must be real-valued to satisfy the Hermitian
property. In these cases the imaginary component will be ignored.
For example, any imaginary component in the zero-frequency term cannot
be represented in a real output and so will always be ignored.
Note:
The correct interpretation of the Hermitian input depends on the length of
the original data, as given by :attr:`s`. This is because each input shape
could correspond to either an odd or even length signal. By default, the
signal is assumed to be even length and odd signals will not round-trip
properly. So, it is recommended to always pass the signal shape :attr:`s`.
Note:
Supports torch.half and torch.chalf on CUDA with GPU Architecture SM53 or greater.
However it only supports powers of 2 signal length in every transformed dimensions.
With default arguments, the size of last dimension should be (2^n + 1) as argument
`s` defaults to even output size = 2 * (last_dim_size - 1)
Args:
input (Tensor): the input tensor
s (Tuple[int], optional): Signal size in the transformed dimensions.
If given, each dimension ``dim[i]`` will either be zero-padded or
trimmed to the length ``s[i]`` before computing the real FFT.
If a length ``-1`` is specified, no padding is done in that dimension.
Defaults to even output in the last dimension:
``s[-1] = 2*(input.size(dim[-1]) - 1)``.
dim (Tuple[int], optional): Dimensions to be transformed.
The last dimension must be the half-Hermitian compressed dimension.
Default: last two dimensions.
norm (str, optional): Normalization mode. For the backward transform
(:func:`~torch.fft.irfft2`), these correspond to:
* ``"forward"`` - no normalization
* ``"backward"`` - normalize by ``1/n``
* ``"ortho"`` - normalize by ``1/sqrt(n)`` (making the real IFFT orthonormal)
Where ``n = prod(s)`` is the logical IFFT size.
Calling the forward transform (:func:`~torch.fft.rfft2`) with the same
normalization mode will apply an overall normalization of ``1/n`` between
the two transforms. This is required to make :func:`~torch.fft.irfft2`
the exact inverse.
Default is ``"backward"`` (normalize by ``1/n``).
Keyword args:
{out}
Example:
>>> t = torch.rand(10, 9)
>>> T = torch.fft.rfft2(t)
Without specifying the output length to :func:`~torch.fft.irfft2`, the output
will not round-trip properly because the input is odd-length in the last
dimension:
>>> torch.fft.irfft2(T).size()
torch.Size([10, 8])
So, it is recommended to always pass the signal shape :attr:`s`.
>>> roundtrip = torch.fft.irfft2(T, t.size())
>>> roundtrip.size()
torch.Size([10, 9])
>>> torch.testing.assert_close(roundtrip, t, check_stride=False)
""".format(**common_args))
rfftn = _add_docstr(_fft.fft_rfftn, r"""
rfftn(input, s=None, dim=None, norm=None, *, out=None) -> Tensor
Computes the N-dimensional discrete Fourier transform of real :attr:`input`.
The FFT of a real signal is Hermitian-symmetric,
``X[i_1, ..., i_n] = conj(X[-i_1, ..., -i_n])`` so the full
:func:`~torch.fft.fftn` output contains redundant information.
:func:`~torch.fft.rfftn` instead omits the negative frequencies in the
last dimension.
Note:
Supports torch.half on CUDA with GPU Architecture SM53 or greater.
However it only supports powers of 2 signal length in every transformed dimensions.
Args:
input (Tensor): the input tensor
s (Tuple[int], optional): Signal size in the transformed dimensions.
If given, each dimension ``dim[i]`` will either be zero-padded or
trimmed to the length ``s[i]`` before computing the real FFT.
If a length ``-1`` is specified, no padding is done in that dimension.
Default: ``s = [input.size(d) for d in dim]``
dim (Tuple[int], optional): Dimensions to be transformed.
Default: all dimensions, or the last ``len(s)`` dimensions if :attr:`s` is given.
norm (str, optional): Normalization mode. For the forward transform
(:func:`~torch.fft.rfftn`), these correspond to:
* ``"forward"`` - normalize by ``1/n``
* ``"backward"`` - no normalization
* ``"ortho"`` - normalize by ``1/sqrt(n)`` (making the real FFT orthonormal)
Where ``n = prod(s)`` is the logical FFT size.
Calling the backward transform (:func:`~torch.fft.irfftn`) with the same
normalization mode will apply an overall normalization of ``1/n`` between
the two transforms. This is required to make :func:`~torch.fft.irfftn`
the exact inverse.
Default is ``"backward"`` (no normalization).
Keyword args:
{out}
Example:
>>> t = torch.rand(10, 10)
>>> rfftn = torch.fft.rfftn(t)
>>> rfftn.size()
torch.Size([10, 6])
Compared against the full output from :func:`~torch.fft.fftn`, we have all
elements up to the Nyquist frequency.
>>> fftn = torch.fft.fftn(t)
>>> torch.testing.assert_close(fftn[..., :6], rfftn, check_stride=False)
The discrete Fourier transform is separable, so :func:`~torch.fft.rfftn`
here is equivalent to a combination of :func:`~torch.fft.fft` and
:func:`~torch.fft.rfft`:
>>> two_ffts = torch.fft.fft(torch.fft.rfft(t, dim=1), dim=0)
>>> torch.testing.assert_close(rfftn, two_ffts, check_stride=False)
""".format(**common_args))
irfftn = _add_docstr(_fft.fft_irfftn, r"""
irfftn(input, s=None, dim=None, norm=None, *, out=None) -> Tensor
Computes the inverse of :func:`~torch.fft.rfftn`.
:attr:`input` is interpreted as a one-sided Hermitian signal in the Fourier
domain, as produced by :func:`~torch.fft.rfftn`. By the Hermitian property, the
output will be real-valued.
Note:
Some input frequencies must be real-valued to satisfy the Hermitian
property. In these cases the imaginary component will be ignored.
For example, any imaginary component in the zero-frequency term cannot
be represented in a real output and so will always be ignored.
Note:
The correct interpretation of the Hermitian input depends on the length of
the original data, as given by :attr:`s`. This is because each input shape
could correspond to either an odd or even length signal. By default, the
signal is assumed to be even length and odd signals will not round-trip
properly. So, it is recommended to always pass the signal shape :attr:`s`.
Note:
Supports torch.half and torch.chalf on CUDA with GPU Architecture SM53 or greater.
However it only supports powers of 2 signal length in every transformed dimensions.
With default arguments, the size of last dimension should be (2^n + 1) as argument
`s` defaults to even output size = 2 * (last_dim_size - 1)
Args:
input (Tensor): the input tensor
s (Tuple[int], optional): Signal size in the transformed dimensions.
If given, each dimension ``dim[i]`` will either be zero-padded or
trimmed to the length ``s[i]`` before computing the real FFT.
If a length ``-1`` is specified, no padding is done in that dimension.
Defaults to even output in the last dimension:
``s[-1] = 2*(input.size(dim[-1]) - 1)``.
dim (Tuple[int], optional): Dimensions to be transformed.
The last dimension must be the half-Hermitian compressed dimension.
Default: all dimensions, or the last ``len(s)`` dimensions if :attr:`s` is given.
norm (str, optional): Normalization mode. For the backward transform
(:func:`~torch.fft.irfftn`), these correspond to:
* ``"forward"`` - no normalization
* ``"backward"`` - normalize by ``1/n``
* ``"ortho"`` - normalize by ``1/sqrt(n)`` (making the real IFFT orthonormal)
Where ``n = prod(s)`` is the logical IFFT size.
Calling the forward transform (:func:`~torch.fft.rfftn`) with the same
normalization mode will apply an overall normalization of ``1/n`` between
the two transforms. This is required to make :func:`~torch.fft.irfftn`
the exact inverse.
Default is ``"backward"`` (normalize by ``1/n``).
Keyword args:
{out}
Example:
>>> t = torch.rand(10, 9)
>>> T = torch.fft.rfftn(t)
Without specifying the output length to :func:`~torch.fft.irfft`, the output
will not round-trip properly because the input is odd-length in the last
dimension:
>>> torch.fft.irfftn(T).size()
torch.Size([10, 8])
So, it is recommended to always pass the signal shape :attr:`s`.
>>> roundtrip = torch.fft.irfftn(T, t.size())
>>> roundtrip.size()
torch.Size([10, 9])
>>> torch.testing.assert_close(roundtrip, t, check_stride=False)
""".format(**common_args))
hfft = _add_docstr(_fft.fft_hfft, r"""
hfft(input, n=None, dim=-1, norm=None, *, out=None) -> Tensor
Computes the one dimensional discrete Fourier transform of a Hermitian
symmetric :attr:`input` signal.
Note:
:func:`~torch.fft.hfft`/:func:`~torch.fft.ihfft` are analogous to
:func:`~torch.fft.rfft`/:func:`~torch.fft.irfft`. The real FFT expects
a real signal in the time-domain and gives a Hermitian symmetry in the
frequency-domain. The Hermitian FFT is the opposite; Hermitian symmetric in
the time-domain and real-valued in the frequency-domain. For this reason,
special care needs to be taken with the length argument :attr:`n`, in the
same way as with :func:`~torch.fft.irfft`.
Note:
Because the signal is Hermitian in the time-domain, the result will be
real in the frequency domain. Note that some input frequencies must be
real-valued to satisfy the Hermitian property. In these cases the imaginary
component will be ignored. For example, any imaginary component in
``input[0]`` would result in one or more complex frequency terms which
cannot be represented in a real output and so will always be ignored.
Note:
The correct interpretation of the Hermitian input depends on the length of
the original data, as given by :attr:`n`. This is because each input shape
could correspond to either an odd or even length signal. By default, the
signal is assumed to be even length and odd signals will not round-trip
properly. So, it is recommended to always pass the signal length :attr:`n`.
Note:
Supports torch.half and torch.chalf on CUDA with GPU Architecture SM53 or greater.
However it only supports powers of 2 signal length in every transformed dimension.
With default arguments, size of the transformed dimension should be (2^n + 1) as argument
`n` defaults to even output size = 2 * (transformed_dim_size - 1)
Args:
input (Tensor): the input tensor representing a half-Hermitian signal
n (int, optional): Output signal length. This determines the length of the
real output. If given, the input will either be zero-padded or trimmed to this
length before computing the Hermitian FFT.
Defaults to even output: ``n=2*(input.size(dim) - 1)``.
dim (int, optional): The dimension along which to take the one dimensional Hermitian FFT.
norm (str, optional): Normalization mode. For the forward transform
(:func:`~torch.fft.hfft`), these correspond to:
* ``"forward"`` - normalize by ``1/n``
* ``"backward"`` - no normalization
* ``"ortho"`` - normalize by ``1/sqrt(n)`` (making the Hermitian FFT orthonormal)
Calling the backward transform (:func:`~torch.fft.ihfft`) with the same
normalization mode will apply an overall normalization of ``1/n`` between
the two transforms. This is required to make :func:`~torch.fft.ihfft`
the exact inverse.
Default is ``"backward"`` (no normalization).
Keyword args:
{out}
Example:
Taking a real-valued frequency signal and bringing it into the time domain
gives Hermitian symmetric output:
>>> t = torch.linspace(0, 1, 5)
>>> t
tensor([0.0000, 0.2500, 0.5000, 0.7500, 1.0000])
>>> T = torch.fft.ifft(t)
>>> T
tensor([ 0.5000-0.0000j, -0.1250-0.1720j, -0.1250-0.0406j, -0.1250+0.0406j,
-0.1250+0.1720j])
Note that ``T[1] == T[-1].conj()`` and ``T[2] == T[-2].conj()`` is
redundant. We can thus compute the forward transform without considering
negative frequencies:
>>> torch.fft.hfft(T[:3], n=5)
tensor([0.0000, 0.2500, 0.5000, 0.7500, 1.0000])
Like with :func:`~torch.fft.irfft`, the output length must be given in order
to recover an even length output:
>>> torch.fft.hfft(T[:3])
tensor([0.1250, 0.2809, 0.6250, 0.9691])
""".format(**common_args))
ihfft = _add_docstr(_fft.fft_ihfft, r"""
ihfft(input, n=None, dim=-1, norm=None, *, out=None) -> Tensor
Computes the inverse of :func:`~torch.fft.hfft`.
:attr:`input` must be a real-valued signal, interpreted in the Fourier domain.
The IFFT of a real signal is Hermitian-symmetric, ``X[i] = conj(X[-i])``.
:func:`~torch.fft.ihfft` represents this in the one-sided form where only the
positive frequencies below the Nyquist frequency are included. To compute the
full output, use :func:`~torch.fft.ifft`.
Note:
Supports torch.half on CUDA with GPU Architecture SM53 or greater.
However it only supports powers of 2 signal length in every transformed dimension.
Args:
input (Tensor): the real input tensor
n (int, optional): Signal length. If given, the input will either be zero-padded
or trimmed to this length before computing the Hermitian IFFT.
dim (int, optional): The dimension along which to take the one dimensional Hermitian IFFT.
norm (str, optional): Normalization mode. For the backward transform
(:func:`~torch.fft.ihfft`), these correspond to:
* ``"forward"`` - no normalization
* ``"backward"`` - normalize by ``1/n``
* ``"ortho"`` - normalize by ``1/sqrt(n)`` (making the IFFT orthonormal)
Calling the forward transform (:func:`~torch.fft.hfft`) with the same
normalization mode will apply an overall normalization of ``1/n`` between
the two transforms. This is required to make :func:`~torch.fft.ihfft`
the exact inverse.
Default is ``"backward"`` (normalize by ``1/n``).
Keyword args:
{out}
Example:
>>> t = torch.arange(5)
>>> t
tensor([0, 1, 2, 3, 4])
>>> torch.fft.ihfft(t)
tensor([ 2.0000-0.0000j, -0.5000-0.6882j, -0.5000-0.1625j])
Compare against the full output from :func:`~torch.fft.ifft`:
>>> torch.fft.ifft(t)
tensor([ 2.0000-0.0000j, -0.5000-0.6882j, -0.5000-0.1625j, -0.5000+0.1625j,
-0.5000+0.6882j])
""".format(**common_args))
hfft2 = _add_docstr(_fft.fft_hfft2, r"""
hfft2(input, s=None, dim=(-2, -1), norm=None, *, out=None) -> Tensor
Computes the 2-dimensional discrete Fourier transform of a Hermitian symmetric
:attr:`input` signal. Equivalent to :func:`~torch.fft.hfftn` but only
transforms the last two dimensions by default.
:attr:`input` is interpreted as a one-sided Hermitian signal in the time
domain. By the Hermitian property, the Fourier transform will be real-valued.
Note:
Supports torch.half and torch.chalf on CUDA with GPU Architecture SM53 or greater.
However it only supports powers of 2 signal length in every transformed dimensions.
With default arguments, the size of last dimension should be (2^n + 1) as argument
`s` defaults to even output size = 2 * (last_dim_size - 1)
Args:
input (Tensor): the input tensor
s (Tuple[int], optional): Signal size in the transformed dimensions.
If given, each dimension ``dim[i]`` will either be zero-padded or
trimmed to the length ``s[i]`` before computing the Hermitian FFT.
If a length ``-1`` is specified, no padding is done in that dimension.
Defaults to even output in the last dimension:
``s[-1] = 2*(input.size(dim[-1]) - 1)``.
dim (Tuple[int], optional): Dimensions to be transformed.
The last dimension must be the half-Hermitian compressed dimension.
Default: last two dimensions.
norm (str, optional): Normalization mode. For the forward transform
(:func:`~torch.fft.hfft2`), these correspond to:
* ``"forward"`` - normalize by ``1/n``
* ``"backward"`` - no normalization
* ``"ortho"`` - normalize by ``1/sqrt(n)`` (making the Hermitian FFT orthonormal)
Where ``n = prod(s)`` is the logical FFT size.
Calling the backward transform (:func:`~torch.fft.ihfft2`) with the same
normalization mode will apply an overall normalization of ``1/n`` between
the two transforms. This is required to make :func:`~torch.fft.ihfft2`
the exact inverse.
Default is ``"backward"`` (no normalization).
Keyword args:
{out}
Example:
Starting from a real frequency-space signal, we can generate a
Hermitian-symmetric time-domain signal:
>>> T = torch.rand(10, 9)
>>> t = torch.fft.ihfft2(T)
Without specifying the output length to :func:`~torch.fft.hfftn`, the
output will not round-trip properly because the input is odd-length in the
last dimension:
>>> torch.fft.hfft2(t).size()
torch.Size([10, 10])
So, it is recommended to always pass the signal shape :attr:`s`.
>>> roundtrip = torch.fft.hfft2(t, T.size())
>>> roundtrip.size()
torch.Size([10, 9])
>>> torch.allclose(roundtrip, T)
True
""".format(**common_args))
ihfft2 = _add_docstr(_fft.fft_ihfft2, r"""
ihfft2(input, s=None, dim=(-2, -1), norm=None, *, out=None) -> Tensor
Computes the 2-dimensional inverse discrete Fourier transform of real
:attr:`input`. Equivalent to :func:`~torch.fft.ihfftn` but transforms only the
two last dimensions by default.
Note:
Supports torch.half on CUDA with GPU Architecture SM53 or greater.
However it only supports powers of 2 signal length in every transformed dimensions.
Args:
input (Tensor): the input tensor
s (Tuple[int], optional): Signal size in the transformed dimensions.
If given, each dimension ``dim[i]`` will either be zero-padded or
trimmed to the length ``s[i]`` before computing the Hermitian IFFT.
If a length ``-1`` is specified, no padding is done in that dimension.
Default: ``s = [input.size(d) for d in dim]``
dim (Tuple[int], optional): Dimensions to be transformed.
Default: last two dimensions.
norm (str, optional): Normalization mode. For the backward transform
(:func:`~torch.fft.ihfft2`), these correspond to:
* ``"forward"`` - no normalization
* ``"backward"`` - normalize by ``1/n``
* ``"ortho"`` - normalize by ``1/sqrt(n)`` (making the Hermitian IFFT orthonormal)
Where ``n = prod(s)`` is the logical IFFT size.
Calling the forward transform (:func:`~torch.fft.hfft2`) with the same
normalization mode will apply an overall normalization of ``1/n`` between
the two transforms. This is required to make :func:`~torch.fft.ihfft2`
the exact inverse.
Default is ``"backward"`` (normalize by ``1/n``).
Keyword args:
{out}
Example:
>>> T = torch.rand(10, 10)
>>> t = torch.fft.ihfft2(t)
>>> t.size()
torch.Size([10, 6])
Compared against the full output from :func:`~torch.fft.ifft2`, the
Hermitian time-space signal takes up only half the space.
>>> fftn = torch.fft.ifft2(t)
>>> torch.allclose(fftn[..., :6], rfftn)
True
The discrete Fourier transform is separable, so :func:`~torch.fft.ihfft2`
here is equivalent to a combination of :func:`~torch.fft.ifft` and
:func:`~torch.fft.ihfft`:
>>> two_ffts = torch.fft.ifft(torch.fft.ihfft(t, dim=1), dim=0)
>>> torch.allclose(t, two_ffts)
True
""".format(**common_args))
hfftn = _add_docstr(_fft.fft_hfftn, r"""
hfftn(input, s=None, dim=None, norm=None, *, out=None) -> Tensor
Computes the n-dimensional discrete Fourier transform of a Hermitian symmetric
:attr:`input` signal.
:attr:`input` is interpreted as a one-sided Hermitian signal in the time
domain. By the Hermitian property, the Fourier transform will be real-valued.
Note:
:func:`~torch.fft.hfftn`/:func:`~torch.fft.ihfftn` are analogous to
:func:`~torch.fft.rfftn`/:func:`~torch.fft.irfftn`. The real FFT expects
a real signal in the time-domain and gives Hermitian symmetry in the
frequency-domain. The Hermitian FFT is the opposite; Hermitian symmetric in
the time-domain and real-valued in the frequency-domain. For this reason,
special care needs to be taken with the shape argument :attr:`s`, in the
same way as with :func:`~torch.fft.irfftn`.
Note:
Some input frequencies must be real-valued to satisfy the Hermitian
property. In these cases the imaginary component will be ignored.
For example, any imaginary component in the zero-frequency term cannot
be represented in a real output and so will always be ignored.
Note:
The correct interpretation of the Hermitian input depends on the length of
the original data, as given by :attr:`s`. This is because each input shape
could correspond to either an odd or even length signal. By default, the
signal is assumed to be even length and odd signals will not round-trip
properly. It is recommended to always pass the signal shape :attr:`s`.
Note:
Supports torch.half and torch.chalf on CUDA with GPU Architecture SM53 or greater.
However it only supports powers of 2 signal length in every transformed dimensions.
With default arguments, the size of last dimension should be (2^n + 1) as argument
`s` defaults to even output size = 2 * (last_dim_size - 1)
Args:
input (Tensor): the input tensor
s (Tuple[int], optional): Signal size in the transformed dimensions.
If given, each dimension ``dim[i]`` will either be zero-padded or
trimmed to the length ``s[i]`` before computing the real FFT.
If a length ``-1`` is specified, no padding is done in that dimension.
Defaults to even output in the last dimension:
``s[-1] = 2*(input.size(dim[-1]) - 1)``.
dim (Tuple[int], optional): Dimensions to be transformed.
The last dimension must be the half-Hermitian compressed dimension.
Default: all dimensions, or the last ``len(s)`` dimensions if :attr:`s` is given.
norm (str, optional): Normalization mode. For the forward transform
(:func:`~torch.fft.hfftn`), these correspond to:
* ``"forward"`` - normalize by ``1/n``
* ``"backward"`` - no normalization
* ``"ortho"`` - normalize by ``1/sqrt(n)`` (making the Hermitian FFT orthonormal)
Where ``n = prod(s)`` is the logical FFT size.
Calling the backward transform (:func:`~torch.fft.ihfftn`) with the same
normalization mode will apply an overall normalization of ``1/n`` between
the two transforms. This is required to make :func:`~torch.fft.ihfftn`
the exact inverse.
Default is ``"backward"`` (no normalization).
Keyword args:
{out}
Example:
Starting from a real frequency-space signal, we can generate a
Hermitian-symmetric time-domain signal:
>>> T = torch.rand(10, 9)
>>> t = torch.fft.ihfftn(T)
Without specifying the output length to :func:`~torch.fft.hfftn`, the
output will not round-trip properly because the input is odd-length in the
last dimension:
>>> torch.fft.hfftn(t).size()
torch.Size([10, 10])
So, it is recommended to always pass the signal shape :attr:`s`.
>>> roundtrip = torch.fft.hfftn(t, T.size())
>>> roundtrip.size()
torch.Size([10, 9])
>>> torch.allclose(roundtrip, T)
True
""".format(**common_args))
ihfftn = _add_docstr(_fft.fft_ihfftn, r"""
ihfftn(input, s=None, dim=None, norm=None, *, out=None) -> Tensor
Computes the N-dimensional inverse discrete Fourier transform of real :attr:`input`.
:attr:`input` must be a real-valued signal, interpreted in the Fourier domain.
The n-dimensional IFFT of a real signal is Hermitian-symmetric,
``X[i, j, ...] = conj(X[-i, -j, ...])``. :func:`~torch.fft.ihfftn` represents
this in the one-sided form where only the positive frequencies below the
Nyquist frequency are included in the last signal dimension. To compute the
full output, use :func:`~torch.fft.ifftn`.
Note:
Supports torch.half on CUDA with GPU Architecture SM53 or greater.
However it only supports powers of 2 signal length in every transformed dimensions.
Args:
input (Tensor): the input tensor
s (Tuple[int], optional): Signal size in the transformed dimensions.
If given, each dimension ``dim[i]`` will either be zero-padded or
trimmed to the length ``s[i]`` before computing the Hermitian IFFT.
If a length ``-1`` is specified, no padding is done in that dimension.
Default: ``s = [input.size(d) for d in dim]``
dim (Tuple[int], optional): Dimensions to be transformed.
Default: all dimensions, or the last ``len(s)`` dimensions if :attr:`s` is given.
norm (str, optional): Normalization mode. For the backward transform
(:func:`~torch.fft.ihfftn`), these correspond to:
* ``"forward"`` - no normalization
* ``"backward"`` - normalize by ``1/n``
* ``"ortho"`` - normalize by ``1/sqrt(n)`` (making the Hermitian IFFT orthonormal)
Where ``n = prod(s)`` is the logical IFFT size.
Calling the forward transform (:func:`~torch.fft.hfftn`) with the same
normalization mode will apply an overall normalization of ``1/n`` between
the two transforms. This is required to make :func:`~torch.fft.ihfftn`
the exact inverse.
Default is ``"backward"`` (normalize by ``1/n``).
Keyword args:
{out}
Example:
>>> T = torch.rand(10, 10)
>>> ihfftn = torch.fft.ihfftn(T)
>>> ihfftn.size()
torch.Size([10, 6])
Compared against the full output from :func:`~torch.fft.ifftn`, we have all
elements up to the Nyquist frequency.
>>> ifftn = torch.fft.ifftn(t)
>>> torch.allclose(ifftn[..., :6], ihfftn)
True
The discrete Fourier transform is separable, so :func:`~torch.fft.ihfftn`
here is equivalent to a combination of :func:`~torch.fft.ihfft` and
:func:`~torch.fft.ifft`:
>>> two_iffts = torch.fft.ifft(torch.fft.ihfft(t, dim=1), dim=0)
>>> torch.allclose(ihfftn, two_iffts)
True
""".format(**common_args))
fftfreq = _add_docstr(_fft.fft_fftfreq, r"""
fftfreq(n, d=1.0, *, out=None, dtype=None, layout=torch.strided, device=None, requires_grad=False) -> Tensor
Computes the discrete Fourier Transform sample frequencies for a signal of size :attr:`n`.
Note:
By convention, :func:`~torch.fft.fft` returns positive frequency terms
first, followed by the negative frequencies in reverse order, so that
``f[-i]`` for all :math:`0 < i \leq n/2`` in Python gives the negative
frequency terms. For an FFT of length :attr:`n` and with inputs spaced in
length unit :attr:`d`, the frequencies are::
f = [0, 1, ..., (n - 1) // 2, -(n // 2), ..., -1] / (d * n)
Note:
For even lengths, the Nyquist frequency at ``f[n/2]`` can be thought of as
either negative or positive. :func:`~torch.fft.fftfreq` follows NumPy's
convention of taking it to be negative.
Args:
n (int): the FFT length
d (float, optional): The sampling length scale.
The spacing between individual samples of the FFT input.
The default assumes unit spacing, dividing that result by the actual
spacing gives the result in physical frequency units.
Keyword Args:
{out}
{dtype}
{layout}
{device}
{requires_grad}
Example:
>>> torch.fft.fftfreq(5)
tensor([ 0.0000, 0.2000, 0.4000, -0.4000, -0.2000])
For even input, we can see the Nyquist frequency at ``f[2]`` is given as
negative:
>>> torch.fft.fftfreq(4)
tensor([ 0.0000, 0.2500, -0.5000, -0.2500])
""".format(**factory_common_args))
rfftfreq = _add_docstr(_fft.fft_rfftfreq, r"""
rfftfreq(n, d=1.0, *, out=None, dtype=None, layout=torch.strided, device=None, requires_grad=False) -> Tensor
Computes the sample frequencies for :func:`~torch.fft.rfft` with a signal of size :attr:`n`.
Note:
:func:`~torch.fft.rfft` returns Hermitian one-sided output, so only the
positive frequency terms are returned. For a real FFT of length :attr:`n`
and with inputs spaced in length unit :attr:`d`, the frequencies are::
f = torch.arange((n + 1) // 2) / (d * n)
Note:
For even lengths, the Nyquist frequency at ``f[n/2]`` can be thought of as
either negative or positive. Unlike :func:`~torch.fft.fftfreq`,
:func:`~torch.fft.rfftfreq` always returns it as positive.
Args:
n (int): the real FFT length
d (float, optional): The sampling length scale.
The spacing between individual samples of the FFT input.
The default assumes unit spacing, dividing that result by the actual
spacing gives the result in physical frequency units.
Keyword Args:
{out}
{dtype}
{layout}
{device}
{requires_grad}
Example:
>>> torch.fft.rfftfreq(5)
tensor([0.0000, 0.2000, 0.4000])
>>> torch.fft.rfftfreq(4)
tensor([0.0000, 0.2500, 0.5000])
Compared to the output from :func:`~torch.fft.fftfreq`, we see that the
Nyquist frequency at ``f[2]`` has changed sign:
>>> torch.fft.fftfreq(4)
tensor([ 0.0000, 0.2500, -0.5000, -0.2500])
""".format(**factory_common_args))
fftshift = _add_docstr(_fft.fft_fftshift, r"""
fftshift(input, dim=None) -> Tensor
Reorders n-dimensional FFT data, as provided by :func:`~torch.fft.fftn`, to have
negative frequency terms first.
This performs a periodic shift of n-dimensional data such that the origin
``(0, ..., 0)`` is moved to the center of the tensor. Specifically, to
``input.shape[dim] // 2`` in each selected dimension.
Note:
By convention, the FFT returns positive frequency terms first, followed by
the negative frequencies in reverse order, so that ``f[-i]`` for all
:math:`0 < i \leq n/2` in Python gives the negative frequency terms.
:func:`~torch.fft.fftshift` rearranges all frequencies into ascending order
from negative to positive with the zero-frequency term in the center.
Note:
For even lengths, the Nyquist frequency at ``f[n/2]`` can be thought of as
either negative or positive. :func:`~torch.fft.fftshift` always puts the
Nyquist term at the 0-index. This is the same convention used by
:func:`~torch.fft.fftfreq`.
Args:
input (Tensor): the tensor in FFT order
dim (int, Tuple[int], optional): The dimensions to rearrange.
Only dimensions specified here will be rearranged, any other dimensions
will be left in their original order.
Default: All dimensions of :attr:`input`.
Example:
>>> f = torch.fft.fftfreq(4)
>>> f
tensor([ 0.0000, 0.2500, -0.5000, -0.2500])
>>> torch.fft.fftshift(f)
tensor([-0.5000, -0.2500, 0.0000, 0.2500])
Also notice that the Nyquist frequency term at ``f[2]`` was moved to the
beginning of the tensor.
This also works for multi-dimensional transforms:
>>> x = torch.fft.fftfreq(5, d=1/5) + 0.1 * torch.fft.fftfreq(5, d=1/5).unsqueeze(1)
>>> x
tensor([[ 0.0000, 1.0000, 2.0000, -2.0000, -1.0000],
[ 0.1000, 1.1000, 2.1000, -1.9000, -0.9000],
[ 0.2000, 1.2000, 2.2000, -1.8000, -0.8000],
[-0.2000, 0.8000, 1.8000, -2.2000, -1.2000],
[-0.1000, 0.9000, 1.9000, -2.1000, -1.1000]])
>>> torch.fft.fftshift(x)
tensor([[-2.2000, -1.2000, -0.2000, 0.8000, 1.8000],
[-2.1000, -1.1000, -0.1000, 0.9000, 1.9000],
[-2.0000, -1.0000, 0.0000, 1.0000, 2.0000],
[-1.9000, -0.9000, 0.1000, 1.1000, 2.1000],
[-1.8000, -0.8000, 0.2000, 1.2000, 2.2000]])
:func:`~torch.fft.fftshift` can also be useful for spatial data. If our
data is defined on a centered grid (``[-(N//2), (N-1)//2]``) then we can
use the standard FFT defined on an uncentered grid (``[0, N)``) by first
applying an :func:`~torch.fft.ifftshift`.
>>> x_centered = torch.arange(-5, 5)
>>> x_uncentered = torch.fft.ifftshift(x_centered)
>>> fft_uncentered = torch.fft.fft(x_uncentered)
Similarly, we can convert the frequency domain components to centered
convention by applying :func:`~torch.fft.fftshift`.
>>> fft_centered = torch.fft.fftshift(fft_uncentered)
The inverse transform, from centered Fourier space back to centered spatial
data, can be performed by applying the inverse shifts in reverse order:
>>> x_centered_2 = torch.fft.fftshift(torch.fft.ifft(torch.fft.ifftshift(fft_centered)))
>>> torch.testing.assert_close(x_centered.to(torch.complex64), x_centered_2, check_stride=False)
""")
ifftshift = _add_docstr(_fft.fft_ifftshift, r"""
ifftshift(input, dim=None) -> Tensor
Inverse of :func:`~torch.fft.fftshift`.
Args:
input (Tensor): the tensor in FFT order
dim (int, Tuple[int], optional): The dimensions to rearrange.
Only dimensions specified here will be rearranged, any other dimensions
will be left in their original order.
Default: All dimensions of :attr:`input`.
Example:
>>> f = torch.fft.fftfreq(5)
>>> f
tensor([ 0.0000, 0.2000, 0.4000, -0.4000, -0.2000])
A round-trip through :func:`~torch.fft.fftshift` and
:func:`~torch.fft.ifftshift` gives the same result:
>>> shifted = torch.fft.fftshift(f)
>>> torch.fft.ifftshift(shifted)
tensor([ 0.0000, 0.2000, 0.4000, -0.4000, -0.2000])
""")
|