+ j+^RIHt^RIt^RIt^RIHt^RIHt .ROt !RR]P4t !RR]P4t]PP!]R 4R RR R/R R lltR#))AnyN) _to_dlpack)Device DLDeviceTypeReadOnlyTensorWrappercR]tRt^tRtRtRtRtRtRt R t R t R t R t R tRtRtRtRtRtR#)rN))))))) ) ) ) ) )))))__name__ __module__ __qualname____firstlineno__kDLCPUkDLCUDA kDLCUDAHost kDLOpenCL kDLVulkankDLMetalkDLVPIkDLROCM kDLROCMHost kDLExtDevkDLCUDAManaged kDLOneAPI kDLWebGPU kDLHexagonkDLMAIA__static_attributes__r S/home/luca/screen-checking/.venv/lib/python3.14/site-packages/torch/utils/dlpack.pyrrsM FGKIIH FGKINIIJGr-caa]tRt^"toRt]P P4t] !]PP]PP04t ]V3RlRl4t]R Rl4tRR/V3Rllt V3RltR tVtV;t#) ra%A zero-copy, read-only view of a tensor for DLPack interop only. Wrapping a tensor with ``ReadOnlyTensorWrapper`` declares the intent that consumers must not mutate its data. It changes only the DLPack export behavior; the wrapper shares storage with the source tensor and does not copy. Both DLPack export paths are routed to read-only variants: * the fast ``__dlpack_c_exchange_api__`` C exchange protocol (used by tvm-ffi / CuteDSL) points at the const exchange API, which exports through ``const_data_ptr()`` and sets ``DLPACK_FLAG_BITMASK_READ_ONLY``; * the ``__dlpack__()`` capsule protocol forces ``read_only=True``. Because the export uses ``const_data_ptr()``, exporting a copy-on-write tensor does not materialize it. The wrapper is export-only: every torch operation other than the DLPack protocol methods raises ``RuntimeError``. Unwrap it (e.g. via the original tensor) to operate on the data. Example:: x = torch.randn(8) ro = ReadOnlyTensorWrapper(x) cute.runtime.from_dlpack(ro, enable_tvm_ffi=True) # read-only export c8<V^8dQhRS[PRR/#)r tensorreturnr)torchTensor)format __classdict__s"r. __annotate__"ReadOnlyTensorWrapper.__annotate__Ls ''U\\'.E'r-c$VPV4#)N) as_subclass)clsr1s&&r.__new__ReadOnlyTensorWrapper.__new__Ks!!#&&r-Nc WP9dD\PP4;_uu_4T!T/T;'g/BuuRRR4#\ VP R\ VRV4 R24h +'giL6;i)Nz only supports DLPack export; 'rz;' is not allowed. Unwrap the tensor to operate on its data.)_DLPACK_ALLOWEDr3_CDisableTorchFunctionSubclass RuntimeErrorrgetattr)r;functypesargskwargss&&&&&r.__torch_function__(ReadOnlyTensorWrapper.__torch_function__Os{ && &6688T4fll498||nj$/018 8  98s A:: B max_versionc 8<VfRp\SV`!RRVRR/VB#)NrJ read_onlyTr rr )super __dlpack__)selfrJrG __class__s&$,r.rO ReadOnlyTensorWrapper.__dlpack__\s+   Kw!TkTTTVTTr-c&<V^8dQh/S[;R&#)r __dlpack_c_exchange_api__)object)r5r6s"r.r7r8"sB &MCr-r )r N)rrrr__doc__r3r@_const_dlpack_exchange_apirT frozensetr4rO__dlpack_device__r? staticmethodr< classmethodrH__annotate_func__r,__classdictcell__ __classcell__)rQr6s@@r.rr"s@).(K(K(M  %,,"@"@AO''    UUur-ato_dlpack(tensor) -> PyCapsule Returns an opaque object (a "DLPack capsule") representing the tensor. .. note:: ``to_dlpack`` is a legacy DLPack interface. The capsule it returns cannot be used for anything in Python other than use it as input to ``from_dlpack``. The more idiomatic use of DLPack is to call ``from_dlpack`` directly on the tensor object - this works when that object has a ``__dlpack__`` method, which PyTorch and most other libraries indeed have now. .. warning:: Only call ``from_dlpack`` once per capsule produced with ``to_dlpack``. Behavior when a capsule is consumed multiple times is undefined. Args: tensor: a tensor to be exported The DLPack capsule shares the tensor's memory. devicecopyc\V^8dQhR\R\R,R\R,RR/#)r ext_tensorr_Nr`r2z torch.Tensor)r_Devicebool)r5s"r.r7r7|s@m-m-m- dNm- + m-  m-r-c\VR4'Edg/pRVR&TpRpRpVeW#R&VP4pVe\V\4'd\P !V4p\V\P 4'g\ R\V4 24h\PPV4pWx8gpV'gWR&V'dVRJd\R V R V R 24hV^,\P\P39do\PPR V^, 24p V^,\P8Hp V 'dV P ^8Xd^M V P p WR &Rp VP"!R/VBp V f&VP'RR4VP"!R/VBp V f(VP'RR4RpVP"!R/VBp V f%VP'RR4VP"!R/VBp \PP)V 4p VRJd!V'gV'gV P+4p V'dV P-V4p V #VfVe \ R4hTp \PP)V 4# \$dELi;i \$dELi;i \$dLi;i)a from_dlpack(ext_tensor) -> Tensor Converts a tensor from an external library into a ``torch.Tensor``. The returned PyTorch tensor will share the memory with the input tensor (which may have come from another library). Note that in-place operations will therefore also affect the data of the input tensor. This may lead to unexpected issues (e.g., other libraries may have read-only flags or immutable data structures), so the user should only do this if they know for sure that this is fine. Args: ext_tensor (object with ``__dlpack__`` attribute, or a DLPack capsule): The tensor or DLPack capsule to convert. If ``ext_tensor`` is a tensor (or ndarray) object, it must support the ``__dlpack__`` protocol (i.e., have a ``ext_tensor.__dlpack__`` method). Otherwise ``ext_tensor`` may be a DLPack capsule, which is an opaque ``PyCapsule`` instance, typically produced by a ``to_dlpack`` function or method. device (torch.device or str or None): An optional PyTorch device specifying where to place the new tensor. If None (default), the new tensor will be on the same device as ``ext_tensor``. copy (bool or None): An optional boolean indicating whether or not to copy ``self``. If None, PyTorch will copy only if necessary. Examples:: >>> import torch.utils.dlpack >>> t = torch.arange(4) # Convert a tensor directly (supported in PyTorch >= 1.10) >>> t2 = torch.from_dlpack(t) >>> t2[:2] = -1 # show that memory is shared >>> t2 tensor([-1, -1, 2, 3]) >>> t tensor([-1, -1, 2, 3]) # The old-style DLPack usage, with an intermediate capsule object >>> capsule = torch.utils.dlpack.to_dlpack(t) >>> capsule >>> t3 = torch.from_dlpack(capsule) >>> t3 tensor([-1, -1, 2, 3]) >>> t3[0] = -9 # now we're sharing memory between 3 tensors >>> t3 tensor([-9, -1, 2, 3]) >>> t2 tensor([-9, -1, 2, 3]) >>> t tensor([-9, -1, 2, 3]) rOrJTFNr`z&from_dlpack: unsupported device type: dl_devicez&cannot move DLPack tensor from device z to z- without copying. Set copy=None or copy=True.zcuda:streamzQdevice and copy kwargs not supported when ext_tensor is already a DLPack capsule.rMr )hasattrrY isinstancestrr3r_AssertionErrortyper@_torchDeviceToDLDevice ValueErrorrrr$cudacurrent_stream cuda_streamrO TypeErrorpop _from_dlpackcloneto)rbr_r`rGrequested_copyproducer_handled_copycross_device_transfer ext_devicetarget_dl_devicergis_cuda stream_ptrdlpackr1s&$$ r. from_dlpackr|s@z<(("$ &} $ %  !6N  113  &#&&f-fell33$'MdSYl^%\]] %xx>>vF &0%C !)&6{#%  JJ}d + #..88 > JJvt $$) ! #..88 > JJ{D )**4V4F&&v. T !*?H]\\^F !YYv&F   !1 c xx$$V,,Y       s6 J.5K K. J=<J= KK K K )rrr)typingrr3enumtorch._Cr to_dlpack torch.typesrrc__all__IntEnumrr4rr@ _add_docstrrr r-r.rs{ ,)  4<<&>UELL>UBY!2m-"m- m-m-r-