Loads a fragment from memory

Returns a reference to the element at a given Coord

Replaces the managed object, deleting the old object.

Store a fragment to memory

Returns a pointer to the vector starting at the current coordinate

Returns whether access is valid or not

Returns a const iterator

Instantiate a PreTrainedBertModel from a pre-trained model file. Download and cache the pre-trained model file if needed.

Constructs a matrix from a uniform element

Returns the sum of squared elements

Possibly convert HWIO to OIHW.

Possibly convert HWIO to OIHW.

Returns the layout object's stride vector

< Layout function

//////////////////////////////////////////////////////////////////////////////////////////// Converts a cuDNN status to cutlass::Status

< rank of index space

Adds a pointer offset in units of Element

Accesses the tensor reference pointing to data

Obtains a Coord<2> from GemmCoord

Allocate a chunk of memory from the buffer to tensor and copy the values.

Return a tensor with the input `shape` as a view into the 1-D data starting at `start_index`.

Stores a fragment to memory at the location pointed to by the iterator

Return tokenizer.

Notify the iterator which k-group it is currently pointing to. This does not advance the iterator. Rather, it overrides its internal tracking with co

Default ctor

Return timers.

Returns the layout object

Overrides the internal iteration index

# Args. input_token_ids: IntTensor, (batch_size, max_input_length), input hidden state to decoder. i

Computes the grid shape

Copied from torch.optim.optimizer

< Layout function

Adds a pointer offset in units of Element

Elementwise multiply operator (1-by-2)

* Find quotient and remainder using device-side intrinsics */