Mojo function
mla_prefill_plan_kernel
mla_prefill_plan_kernel[buffer_row_offsets_layout: Layout, cache_offsets_layout: Layout, buffer_lengths_layout: Layout, input_row_offsets_layout: Layout, cache_t: KVCacheT](buffer_row_offsets: LayoutTensor[DType.uint32, buffer_row_offsets_layout, MutableAnyOrigin], cache_offsets: LayoutTensor[DType.uint32, cache_offsets_layout, MutableAnyOrigin], buffer_lengths: LayoutTensor[DType.int32, buffer_lengths_layout, MutableAnyOrigin], input_row_offsets: LayoutTensor[DType.uint32, input_row_offsets_layout, MutableAnyOrigin], k_cache: cache_t, buffer_token_size: UInt32)
Was this page helpful?
Thank you! We'll create more content like this.
Thank you for helping us improve!