a d @sddlZddlmZddlmZddlmZmZmZmZm Z Gdddej Z Gd d d ej Z Gd d d ej Z deeeeeeeedfeejee eejfdddZdS)N) brownian_base)brownian_interval)OptionalScalarTensorTupleUnioncs^eZdZfddZdddZddZed d Zed d Zed dZ eddZ Z S)ReverseBrowniancstt|||_dSN)superr __init__ base_brownian)selfr __class__c/var/www/html/stable-diffusion-webui/venv/lib/python3.9/site-packages/torchsde/_brownian/derived.pyrszReverseBrownian.__init__NFcCs|j| | ||dSN)return_Ureturn_A)r)rtatbrrrrr__call__szReverseBrownian.__call__cCs|jjd|jdS)Nz(base_brownian=))r__name__rrrrr__repr__ szReverseBrownian.__repr__cCs|jjSr )rdtyperrrrr#szReverseBrownian.dtypecCs|jjSr )rdevicerrrrr 'szReverseBrownian.devicecCs|jjSr )rshaperrrrr!+szReverseBrownian.shapecCs|jjSr )rlevy_area_approximationrrrrr"/sz'ReverseBrownian.levy_area_approximation)NFF) r __module__ __qualname__rrrpropertyrr r!r" __classcell__rrrrr s     r csneZdZdZdeeedfdd Zddd Zd d Z e d d Z e ddZ e ddZ e ddZZS) BrownianPathaBrownian path, storing every computed value. Useful for speed, when memory isn't a concern. To use: >>> bm = BrownianPath(t0=0.0, w0=torch.zeros(4, 1)) >>> bm(0., 0.5) tensor([[ 0.0733], [-0.5692], [ 0.1872], [-0.3889]]) )t0w0 window_sizecs>|d}||_tj|||j|j|jdd|_tt| dS)zInitialize Brownian path. Arguments: t0: Initial time. w0: Initial state. window_size: Unused; deprecated. rN)r)t1sizerr Z cache_size) _w0rBrownianIntervalr!rr _intervalr r'r)rr)r*r+r,rrrrBs zBrownianPath.__init__NFcCs0|j||||d}|dur,|s,|s,||j}|Srr0r.rtrrroutrrrrOs zBrownianPath.__call__cCs|jjd|jdSNz (interval=rrrr0rrrrrVszBrownianPath.__repr__cCs|jjSr r0rrrrrrYszBrownianPath.dtypecCs|jjSr r0r rrrrr ]szBrownianPath.devicecCs|jjSr r0r!rrrrr!aszBrownianPath.shapecCs|jjSr r0r"rrrrr"esz$BrownianPath.levy_area_approximation)r()NFF)rr#r$__doc__rrintrrrr%rr r!r"r&rrrrr'4s     r'c seZdZdZdeeeeeeeeeeeeed fdd Z dd d Z d d Z e ddZ e ddZe ddZe ddZZS) BrownianTreeaBrownian tree with fixed entropy. Useful when the map from entropy -> Brownian motion shouldn't depend on the locations and order of the query points. (As the usual BrownianInterval does - note that BrownianTree is slower as a result though.) To use: >>> bm = BrownianTree(t0=0.0, w0=torch.zeros(4, 1)) >>> bm(0., 0.5) tensor([[ 0.0733], [-0.5692], [ 0.1872], [-0.3889]], device='cuda:0') Nư> ) r)r*r,w1entropytol pool_size cache_depthsafetyc sd|dur|d}|durd} n||} ||_tj|||j|j|j|||d| d |_tt| dS)aInitialize the Brownian tree. The random value generation process exploits the parallel random number paradigm and uses `numpy.random.SeedSequence`. The default generator is PCG64 (used by `default_rng`). Arguments: t0: Initial time. w0: Initial state. t1: Terminal time. w1: Terminal state. entropy: Global seed, defaults to `None` for random entropy. tol: Error tolerance before the binary search is terminated; the search depth ~ log2(tol). pool_size: Size of the pooled entropy. This parameter affects the query speed significantly. cache_depth: Unused; deprecated. safety: Unused; deprecated. NrT) r)r,r-rr rBrCrDZ halfway_treeW) r.rr/r!rr r0r r=r) rr)r*r,rArBrCrDrErFrGrrrrzs$ zBrownianTree.__init__FcCs0|j||||d}|dur,|s,|s,||j}|Srr1r2rrrrs zBrownianTree.__call__cCs|jjd|jdSr5r6rrrrrszBrownianTree.__repr__cCs|jjSr r7rrrrrszBrownianTree.dtypecCs|jjSr r8rrrrr szBrownianTree.devicecCs|jjSr r9rrrrr!szBrownianTree.shapecCs|jjSr r:rrrrr"sz$BrownianTree.levy_area_approximation)NNNr>r?r@N)NFF)rr#r$r;rrrr<floatrrrr%rr r!r"r&rrrrr=js8-    r=?.)yr)r,r-rr cKsR|dur|jn|}|dur |jn|}|dur2|jn|}tjf|||||d|S)zZReturns a BrownianInterval object with the same size, device, and dtype as a given tensor.N)r)r,r-rr )r!rr rr/)rKr)r,r-rr kwargsrrrbrownian_interval_likesrM)rIrJNNN)Ztorchrrtypesrrrr r Z BaseBrownianr r'r=r<rstrr rMrrrrs$  6Y