+
&jk ^ RI t ^ RIt^ RIt^ RIHt ^ RIHt ^ RIHt ^ RIH t H
t
R.t R t . R Ot
. R Ot. R
Ot. ROt]
].t]].tRR lt]P( P* R 4 t ! R
R]4 tR# ) NTensor) constraints) Distribution)
broadcast_all
lazy_propertyVonMisesc \ V4 pVP 4 pV' d VP 4 W, , pK' V# N)listpop)ycoefresults && u/Users/jameslopez/projects/CWCArchive/cwc-podcast/.venv/lib/python3.14/site-packages/torch/distributions/von_mises.py
_eval_polyr s0 :D
XXZF
aj( M c V^ 8w d V^8w d \ RV 24 hV R, pW", p\ V\ V, 4 pV^8X d V P 4 V, pVP 4 pRV , pV RV P 4 , ,
\ V\
V, 4 P 4 , p\ P ! V R8 W44 pV# )zL
Returns ``log(I_order(x))`` for ``x > 0``,
where `order` is either 0 or 1.
zorder must be 0 or 1, got g @ ?)AssertionErrorr _COEF_SMALLabslog _COEF_LARGEtorchwhere)xorderr smalllarger s && r _log_modified_bessel_fnr! D s
zeqj 9% ABB
DA A q+e, -E z% IIKE
qA
aeeg
1k%.@ A E E G GE
[[T5
0F Mr c \ P ! VP \ P V P R 7 pVP 4 ' EgS \ P ! RVP , V P V P R 7 pVP 4 w rgp\ P ! \ P V, 4 p ^W), , W), , p
WV
,
, p V ^V ,
, V,
^ 8 W, P 4 ^, V ,
^ 8 , p V P 4 ' g EK \ P ! WR,
P 4 V
P! 4 , V4 pWL, pEKi V\ P , V , ^\ P , , \ P ,
# )dtypedevicer ) )r zerosshapeboolr% allrandr$ unbindcosmathpir anyr signacos)
loc
concentration
proposal_rr doneuu1u2u3zfcaccepts
&&&& r _rejection_sampler? \ s% ;;qwwejj DDhhjj JJtagg~SYYszz JXXZ
IIdggl #
JN 3 !^ ,A;#q(af\\^a-?!-Cq-HI ::<< F#XOO$5$@!DA=D
K# !dgg+ . 88r c a a ] tR t^nt oRtR]P R]P /t]P t Rt
RV3R lV 3R lllt R t ]
V3R lR l4 t]
V3R
lR l4 t]
V3R lR
l4 t]P$ ! 4 ]P& ! 4 3R l4 tRV 3R llt]V3R lR l4 t]V3R lR l4 t]
V3R lR l4 tRtVtV ;t # )r a(
A circular von Mises distribution.
This implementation uses polar coordinates. The ``loc`` and ``value`` args
can be any real number (to facilitate unconstrained optimization), but are
interpreted as angles modulo 2 pi.
Example::
>>> # xdoctest: +IGNORE_WANT("non-deterministic")
>>> m = VonMises(torch.tensor([1.0]), torch.tensor([1.0]))
>>> m.sample() # von Mises distributed with loc=1 and concentration=1
tensor([1.9777])
:param torch.Tensor loc: an angle in radians.
:param torch.Tensor concentration: concentration parameter
r3 r4 Fc > < V ^8 d QhRS[ RS[ RS[R, RR/# ) r3 r4
validate_argsNreturn)r r) )format
__classdict__s "r __annotate__VonMises.__annotate__ s= B B
B B d{ B
Br c < \ W4 w V n V n V P P p\ P
! 4 p\ SV ` WEV4 R # r )r r3 r4 r( r Sizesuper__init__)selfr3 r4 rC batch_shape event_shape __class__s &&&& r rL VonMises.__init__ s@ (5S'H$$$hhnn jjl
=Ar c V V P ' d V P V4 V P \ P ! WP
,
4 , pV\ P ! ^\ P , 4 ,
\ V P ^ R7 ,
pV# )rB r )
_validate_args_validate_sampler4 r r- r3 r. r r/ r! )rM valuelog_probs && r rW VonMises.log_prob sz ! !% (%% %((2B(CC hhq477{#
$%d&8&8B
C
r c < V ^8 d QhRS[ /# rB rD r )rE rF s "r rG rH s ) )f )r c T V P P \ P 4 # r )r3 tor doublerM s &r _loc
VonMises._loc s xx{{5<<((r c < V ^8 d QhRS[ /# rZ r )rE rF s "r rG rH s 3 3 3r c T V P P \ P 4 # r )r4 r\ r r] r^ s &r _concentrationVonMises._concentration s !!$$U\\22r c < V ^8 d QhRS[ /# rZ r )rE rF s "r rG rH s J JV Jr c Z V P p^^^V^, , , P 4 , pV^V, P 4 ,
^V, , p^V^, , ^V, , p^V, V, p\ P ! VR8 WT4 # ) gh㈵>)rc sqrtr r )rM kappataurho _proposal_r_proposal_r_taylors & r rl VonMises._proposal_r s ##1q5!8|#))++a#g^^%%!e)436za#g. Y .{{54<);IIr c B V P V4 p\ P ! W P P V P
P R7 p\ V P V P V P V4 P V P
P 4 # )a
The sampling algorithm for the von Mises distribution is based on the
following paper: D.J. Best and N.I. Fisher, "Efficient simulation of the
von Mises distribution." Applied Statistics (1979): 152-157.
Sampling is always done in double precision internally to avoid a hang
in _rejection_sample() for small values of the concentration, which
starts to happen for single precision around 1e-4 (see issue #88443).
r# ) _extended_shaper emptyr_ r$ r3 r% r? rc rl r\ )rM sample_shaper( r s && r sampleVonMises.sample sl $$\2 KKYY__TXX__ M IIt**D,<,+. ." O MM - -.> ?M((//+.C . . 5 5k BM:c
N N Os A/BBc < V ^8 d QhRS[ /# rZ r )rE rF s "r rG rH s f r c V P # )z(
The provided mean is the circular one.
r3 r^ s &r mean
VonMises.mean s
xxr c < V ^8 d QhRS[ /# rZ r )rE rF s "r rG rH s f r c V P # r r r^ s &r mode
VonMises.mode s xxr c < V ^8 d QhRS[ /# rZ r )rE rF s "r rG rH s
&
r c ^\ V P ^R7 \ V P ^ R7 ,
P 4 ,
# )z,
The provided variance is the circular one.
rS )r! r4 expr^ s &r varianceVonMises.variance s>
'(:(:!D)$*<*r s + 9 B ,
~. ~. 0 9 9"o
| o
r