tj#UddlZddlZddlZddlmZddlmZddlZddlm Z m Z ej e Z GddZdeded efd Zd Zeaeed <ejd ZGddZddZdS)N)Callable) deprecated)KernelRegistrationHandlecneZdZdZdefdZedZejdZddde d ed e fd Z d S) FakeImplHolderz/A holder where one can register a fake impl to.qualnamec"||_g|_dSN)r kernels)selfr s V/var/www/html/CA-Chatbot/venv/lib/python3.11/site-packages/torch/_library/fake_impl.py__init__zFakeImplHolder.__init__s% &( cPt|jdkrdS|jdS)Nr)lenr )r s rkernelzFakeImplHolder.kernels) t|   ! !4|Brc td)NzUnable to directly set kernel. RuntimeError)r values rrzFakeImplHolder.kernel"s;<<.deregister_fake_kernelRs L   ' ' ' ' 'rrz"Failed to register fake_impl '%s':)rrr rtorch_C%_dispatch_has_kernel_for_dispatch_keyrr appendconstruct_meta_kernelimpl Exceptionloginfor#r) r rrlibrr$ meta_kernelhandlers ` @rregisterzFakeImplHolder.register&s {&"- --{)--- x==dmVTT "& &&&x== : #< <<<   f%% F### ( ( ( ( ( (,DM4@@   HHT]KH W W W W    HH4     L   ' ' '   $$:;; s 1DAEN) __name__ __module__ __qualname____doc__strrpropertyrsetterrrr1rrrrs99(((((  X  ]==]=CG;;;;&); ;;;;;;rrr fake_impl_holderrcjtdtjjjfd}|S)N(fake_impl_holder.kernel must not be Nonecjtdjjfd}t|5j|i|cdddS#1swxYwYdS)Nr<c.tdd)Nz (a): You're trying to run this operator with meta Tensors (as opposed to FakeTensors), but this operator may return an output Tensor with data-dependent shape. Meta Tensors don't support operators with outputs that have data-dependent shapes but FakeTensors do. If your operator does not return an output with data-dependent shape, make sure the FakeTensor and/or meta kernel does not call torch.library.get_ctx(). Otherwise, please use FakeTensors.r)r rsr error_on_ctxz@construct_meta_kernel..meta_kernel..error_on_ctxns8OOvOOO   r)rAssertionErrorrset_ctx_getter)argskwargsr?rr:r s @rr/z*construct_meta_kernel..meta_kernelhs  " * !KLL L!(/      L ) ) < <*#*D;F;; < < < < < < < < < < < < < < < < < >> # An operator with data-dependent output shape >>> lib = torch.library.Library("mymodule", "FRAGMENT") >>> lib.define("mymodule::custom_nonzero(Tensor x) -> Tensor") >>> >>> @torch.library.register_fake("mymodule::custom_nonzero") >>> def _(x): >>> # Number of nonzero-elements is data-dependent. >>> # Since we cannot peek at the data in a fake impl, >>> # we use the ctx object to construct a new symint that >>> # represents the data-dependent size. >>> ctx = torch.library.get_ctx() >>> nnz = ctx.new_dynamic_size() >>> shape = [nnz, x.dim()] >>> result = x.new_empty(shape, dtype=torch.int64) >>> return result >>> >>> @torch.library.impl(lib, "custom_nonzero", "CPU") >>> def _(x): >>> x_np = x.numpy() >>> res = np.stack(np.nonzero(x_np), axis=1) >>> return torch.tensor(res, device=x.device) Nzctx.new_dynamic_size(min=z, max=zZ): expected min and max to be statically known ints but got SymInt. This is not supported.rzc, ...): expected min to be greater than or equal to 0: this API can only create non-negative sizes.) rQallow_dynamic_output_shape_opsr% _subclasses fake_tensorDynamicOutputShapeExceptionrR isinstanceSymInt ValueError allocate_sizer[s rrZzFakeImplCtx.new_dynamic_sizesf O #?A $#/KKDHUU U c5< ( ( JsEL,I,I *C**s***  77'C'''  T_c3777r) r2r3r4r5rr FutureWarningr%rcr\rZr9rrrMrMs ZW-.4777EL777 7'(TF8F8F8elF8F8F8F8F8F8rrMc|}tjjj||||SrY)r\r%fx experimentalsymbolic_shapes_constrain_range_for_size)rPmin_valmax_valresults rreresG  - - / /F H)CCGD Mr)rN) contextlibrDloggingcollections.abcrtyping_extensionsrr%torch._library.utilsrr getLoggerr2r,rr6r)rGrH__annotations__contextmanagerrArMrer9rrrwsl$$$$$$(((((( ;;;;;;;;g!!RRRRRRRRjC>h8'8&&& !!!W8W8W8W8W8W8W8W8tr