Mojo function
mla_decode_sm100_dispatch
mla_decode_sm100_dispatch[q_type: DType, k_t: MHAOperand, output_type: DType, mask_t: MHAMask, config: MHAConfig[config.dtype], depth: Int, num_heads: Int, group: Int = 1, *, ragged: Bool = False, _is_cache_length_accurate: Bool = False, decoding_warp_split_k: Bool = False, per_token_scale_rope_aware: Bool = False](q: TileTensor[q_type, q.LayoutType, q.origin, linear_idx_type=q.linear_idx_type, element_size=q.element_size], k: k_t, output: TileTensor[output_type, output.LayoutType, output.origin, linear_idx_type=output.linear_idx_type, element_size=output.element_size], scale: Float32, valid_length: TileTensor[DType.uint32, valid_length.LayoutType, valid_length.origin, linear_idx_type=valid_length.linear_idx_type, element_size=valid_length.element_size], mask: mask_t, scalar_args_buf: TileTensor[DType.int64, scalar_args_buf.LayoutType, scalar_args_buf.origin, linear_idx_type=scalar_args_buf.linear_idx_type, element_size=scalar_args_buf.element_size], batch_size: Int, q_max_seq_len: Int, max_cache_valid_length: Int, ctx: DeviceContext, q_scale_ptr: UnsafePointer[Float32, MutAnyOrigin] = UnsafePointer())
Was this page helpful?
Thank you! We'll create more content like this.
Thank you for helping us improve!