o bi0@s\ddlmZddlmZddlmZedGdddZddd Zed d gd d ZdS))tree) keras_export) auto_namezkeras.KerasTensorc@s2eZdZdZ     dtddZedd Zejd d Zed d Zejd d ZeddZ e jddZ eddZ e jddZ eddZ e jddZ eddZ e jddZ eddZ ddZduddZd d!Zd"d#Zd$d%Zd&d'Zdvd(d)Zd*d+Zd,d-Zd.d/Zd0d1Zd2d3Zd4d5Zd6d7Zd8d9Zd:d;Zdd?Zd@dAZ dBdCZ!dDdEZ"dFdGZ#dHdIZ$dJdKZ%dLdMZ&dNdOZ'dPdQZ(dRdSZ)dTdUZ*dVdWZ+dXdYZ,dZd[Z-d\d]Z.d^d_Z/d`daZ0dbdcZ1dddeZ2dfdgZ3dhdiZ4djdkZ5dldmZ6dndoZ7dpdqZ8dudrdsZ9dS)w KerasTensoraSymbolic tensor -- encapsulates a shape and a dtype. You can use `KerasTensor` instances to build computation graphs of Keras operations, such as `keras.Function` objects or Functional `keras.models.Model` objects. Example: >>> x = keras.KerasTensor(shape=(3, 4), dtype="float32") >>> x.shape (3, 4) >>> x.dtype float32 Calling a Keras operation (including a layer or a model) on a `KerasTensor` instance will return another `KerasTensor` instance with the appropriate shape and dtype. This is called a "symbolic call" (since there is no actual data involved). The computation of the correct output shape and dtype is called "static shape inference". float32FTNc Ksddlm}|dd} |dd} |r tdd||||_|||_ t ||_ t ||_ |j r@|j r@t d| durHt| nd|_| durT|| nd|_|p^t|jj|_||_dS)Nr)backend ragged_rankrow_splits_dtypezUnexpected keyword arguments: z, zIKerasTensor cannot have `sparse=True` and `ragged=True` at the same time.) keras.srcrpop TypeErrorjoinkeysstandardize_shape_shapestandardize_dtype_dtypebool_sparse_ragged ValueErrorint _ragged_rank_row_splits_dtyper __class____name__namerecord_history) selfshapedtypesparseraggedrrkwargsrrr r$Y/home/ubuntu/.local/lib/python3.10/site-packages/keras/src/backend/common/keras_tensor.py__init__s.         zKerasTensor.__init__cC|jSN)rrr$r$r%rEzKerasTensor.shapecCtd)NzlThe `shape` attribute of KerasTensor is immutable. One should create a new instance of KerasTensor for this.AttributeErrorrvaluer$r$r%rIcCr'r()rr)r$r$r%r Pr*zKerasTensor.dtypecCr+)NzlThe `dtype` attribute of KerasTensor is immutable. One should create a new instance of KerasTensor for this.r,r.r$r$r%r Tr0cCr'r()rr)r$r$r%r![r*zKerasTensor.sparsecCr+)NzmThe `sparse` attribute of KerasTensor is immutable. One should create a new instance of KerasTensor for this.r,r.r$r$r%r!_r0cCr'r()rr)r$r$r%rfr*zKerasTensor.ragged_rankcCr+)NzrThe `ragged_rank` attribute of KerasTensor is immutable. One should create a new instance of KerasTensor for this.r,r.r$r$r%rjr0cCr'r()rr)r$r$r%r qr*zKerasTensor.row_splits_dtypecCr+)NzwThe `row_splits_dtype` attribute of KerasTensor is immutable. One should create a new instance of KerasTensor for this.r,r.r$r$r%r ur0cCr'r()rr)r$r$r%r"|r*zKerasTensor.raggedcCr+)NzmThe `ragged` attribute of KerasTensor is immutable. One should create a new instance of KerasTensor for this.r,r.r$r$r%r"r0cCs t|jSr()lenrr)r$r$r%ndims zKerasTensor.ndimcCddlm}|||SNrops)r r6Reshape)rnewshaper6r$r$r%reshape zKerasTensor.reshapecCr3r4)r r6Squeeze)raxisr6r$r$r%squeezer:zKerasTensor.squeezecCr+)NzA KerasTensor is symbolic: it's a placeholder for a shape an a dtype. It doesn't have any actual numerical value. You cannot convert it to an int.rr)r$r$r%__int__zKerasTensor.__int__cCr+)NzA KerasTensor is symbolic: it's a placeholder for a shape an a dtype. It doesn't have any actual numerical value. You cannot convert it to a float.r>r)r$r$r% __float__r@zKerasTensor.__float__cCr+)NzA KerasTensor is symbolic: it's a placeholder for a shape an a dtype. It doesn't have any actual numerical value. You cannot convert it to a NumPy array.r>r)r$r$r% __array__r@zKerasTensor.__array__cCr+)Na%A KerasTensor cannot be used as input to a JAX function. A KerasTensor is a symbolic placeholder for a shape and dtype, used when constructing Keras Functional models or Keras Functions. You can only use it as input to a Keras layer or a Keras operation (from the namespaces `keras.layers` and `keras.ops`). You are likely doing something like: ``` x = Input(...) ... jax_fn(x) # Invalid. ``` What you should do instead is wrap `jax_fn` in a layer: ``` class MyLayer(Layer): def call(self, x): return jax_fn(x) x = MyLayer()(x) ``` r>r)r$r$r% __jax_array__r@zKerasTensor.__jax_array__cCr+)Na)A KerasTensor cannot be used as input to a TensorFlow function. A KerasTensor is a symbolic placeholder for a shape and dtype, used when constructing Keras Functional models or Keras Functions. You can only use it as input to a Keras layer or a Keras operation (from the namespaces `keras.layers` and `keras.ops`). You are likely doing something like: ``` x = Input(...) ... tf_fn(x) # Invalid. ``` What you should do instead is wrap `tf_fn` in a layer: ``` class MyLayer(Layer): def call(self, x): return tf_fn(x) x = MyLayer()(x) ``` r>)rr rr$r$r% __tf_tensor__r@zKerasTensor.__tf_tensor__c Cs.d|jd|jd|jd|jd|jd S)Nz)rr r!r"rr)r$r$r%__repr__szKerasTensor.__repr__cCr+)Nz7Iterating over a symbolic KerasTensor is not supported.)NotImplementedErrorr)r$r$r%__iter__r@zKerasTensor.__iter__cCr+)Nz3A symbolic KerasTensor cannot be used as a boolean.)r r)r$r$r%__bool__szKerasTensor.__bool__cCddlm}|||Sr4r r6Add symbolic_callrotherr6r$r$r%__add__ zKerasTensor.__add__cCddlm}|||Sr4rKrNr$r$r%__radd__rQzKerasTensor.__radd__cCrJr4r r6SubtractrMrNr$r$r%__sub__rQzKerasTensor.__sub__cCrRr4rTrNr$r$r%__rsub__rQzKerasTensor.__rsub__cCrJr4r r6MultiplyrMrNr$r$r%__mul__rQzKerasTensor.__mul__cCrRr4rXrNr$r$r%__rmul__rQzKerasTensor.__rmul__cCrJr4r r6MatmulrMrNr$r$r% __matmul__rQzKerasTensor.__matmul__cCrRr4r\rNr$r$r% __rmatmul__ rQzKerasTensor.__rmatmul__cCrJr4r r6DividerMrNr$r$r%__div__rQzKerasTensor.__div__cCrRr4r`rNr$r$r%__rdiv__rQzKerasTensor.__rdiv__cCrJr4r r6 TrueDividerMrNr$r$r% __truediv__rQzKerasTensor.__truediv__cCrRr4rdrNr$r$r% __rtruediv__rQzKerasTensor.__rtruediv__cCddlm}||Sr4)r r6NegativerMrr6r$r$r%__neg__"r:zKerasTensor.__neg__cCrhr4)r r6AbsoluterMrjr$r$r%__abs__'r:zKerasTensor.__abs__cCrJr4r r6PowerrMrNr$r$r%__pow__,rQzKerasTensor.__pow__cCrRr4rnrNr$r$r%__rpow__1rQzKerasTensor.__rpow__cCrJr4r r6 FloorDividerMrNr$r$r% __floordiv__6rQzKerasTensor.__floordiv__cCrRr4rrrNr$r$r% __rfloordiv__;rQzKerasTensor.__rfloordiv__cCrJr4r r6ModrMrNr$r$r%__mod__@rQzKerasTensor.__mod__cCrRr4rvrNr$r$r%__rmod__ErQzKerasTensor.__rmod__cCrJr4)r r6LessrMrNr$r$r%__lt__JrQzKerasTensor.__lt__cCrJr4)r r6 LessEqualrMrNr$r$r%__le__OrQzKerasTensor.__le__cCrJr4)r r6GreaterrMrNr$r$r%__gt__TrQzKerasTensor.__gt__cCrJr4)r r6 GreaterEqualrMrNr$r$r%__ge__YrQzKerasTensor.__ge__cCrJr4)r r6NotEqualrMrNr$r$r%__ne__^rQzKerasTensor.__ne__cCrJr4r r6 LogicalAndrMrNr$r$r%__and__crQzKerasTensor.__and__cCrRr4rrNr$r$r%__rand__hrQzKerasTensor.__rand__cCrJr4r r6 LogicalOrrMrNr$r$r%__or__mrQzKerasTensor.__or__cCrRr4rrNr$r$r%__ror__rrQzKerasTensor.__ror__cCrhr4)r r6 LogicalNotrMrjr$r$r% __invert__wr:zKerasTensor.__invert__cCrJr4r r6 LogicalXorrMrNr$r$r%__xor__|rQzKerasTensor.__xor__cCrRr4rrNr$r$r%__rxor__rQzKerasTensor.__rxor__cCrJr4)r r6GetItemrM)rkeyr6r$r$r% __getitem__rQzKerasTensor.__getitem__cCs&ddlm}|p d}|j|d|S)Nrr5)decimals)r r6RoundrM)rndigitsr6rr$r$r% __round__s zKerasTensor.__round__)rFFTNr(NN):r __module__ __qualname____doc__r&propertyrsetterr r!rr r"r2r9r=r?rArBrCrDrFrHrIrPrSrVrWrZr[r^r_rbrcrfrgrkrmrprqrtrurxryr{r}rrrrrrrrrrrrr$r$r$r%rs '               rNcCs8|pd}|pi}t||fD] }t|trdSqdS)Nr$TF)rflatten isinstancer)argsr#xr$r$r%any_symbolic_tensorss rzkeras.utils.is_keras_tensorzkeras.backend.is_keras_tensorcCs t|tS)a|Returns whether `x` is a Keras tensor. A "Keras tensor" is a *symbolic tensor*, such as a tensor that was created via `Input()`. A "symbolic tensor" can be understood as a placeholder -- it does not contain any actual numerical data, only a shape and dtype. It can be used for building Functional models, but it cannot be used in actual computations. )rr)rr$r$r%is_keras_tensors rr) r rkeras.src.api_exportrkeras.src.utils.namingrrrrr$r$r$r%s