For the complete documentation index, see llms.txt. Markdown versions of all pages are available by appending .md to any URL (e.g. /get-started.md).
Mojo struct
Struct_mha_ragged_paged_rel_logits
struct Struct_mha_ragged_paged_rel_logits
Registers the mo.mha.ragged.paged.rel_logits graph op.
Adds a relative-position bias to pre-softmax scores, gathered by
rel_dist = q_pos - k_pos via RelativeLogitsMask rather than
materializing a dense (q, k) mask.
Implemented traits
Methods
execute
static def execute[out_dtype: DType, q_dtype: DType, cache_dtype: DType, //, target: StringSpan[ImmStaticOrigin], local_window_size: Int = Int(-1)](output: ManagedTensorSlice[IOSpec[_, _].Output, static_spec=output.static_spec], q: ManagedTensorSlice[IOSpec[_, _].Input, static_spec=q.static_spec], input_row_offsets: ManagedTensorSlice[IOSpec[_, _].Input, static_spec=input_row_offsets.static_spec], kv_blocks: ManagedTensorSlice[IOSpec[_, _].MutableInput, static_spec=kv_blocks.static_spec], cache_lengths: ManagedTensorSlice[IOSpec[_, _].Input, static_spec=cache_lengths.static_spec], kv_lookup_table: ManagedTensorSlice[IOSpec[_, _].Input, static_spec=kv_lookup_table.static_spec], max_prompt_length: ManagedTensorSlice[IOSpec[_, _].Input, static_spec=max_prompt_length.static_spec], max_cache_length: ManagedTensorSlice[IOSpec[_, _].Input, static_spec=max_cache_length.static_spec], layer_idx: UInt32, scale: Float32, bias: ManagedTensorSlice[IOSpec[_, _].Input, static_spec=bias.static_spec], mha_decode_dispatch_metadata: ManagedTensorSlice[IOSpec[_, _].Input, static_spec=mha_decode_dispatch_metadata.static_spec], context: DeviceContext)