* Return the shape of s0 op s1 with broadcast. * * compute r0, the broadcasted shape as a tensor. * s0, s1 and r0 are all integer vectors. * * Th

* Broadcast an array to a compatible shape NumPy-style. * * The tensor's shape is compared to the broadcast shape from end to beginning. * Ones are

* Returns a promise of `tf.TensorBuffer` that holds the underlying data. * * @doc {heading: 'Tensors', subheading: 'Classes'}

* Creates the layer weights. * * Must be implemented on all layers that have weights. * * Called when apply() is called to construct the w

* Build an empty cache for use with `callWithCache()`.

* Builds a map of internal node keys to node ordering. * Used in serializaion a node orderings may change as unused nodes are * dropped. Porting

* Retrieves the Container's current loss values. * * Used for regularizers during training.

* Retrieves the Layer's current loss values. * * Used for regularizers during training.

* Call the model on new inputs. * * In this case `call` just reapplies all ops in the graph to the new inputs * (e.g. build a new computation

* This is where the layer's logic lives. * * @param inputs Input tensor, or list/tuple of input tensors. * @param kwargs Additional keyword a

* Forward pass of `GPT2CausalLM` with cache. * * `callWithCache` adds an additional forward pass for the model for * autoregressive infere