e97f802b2d
Signed-off-by: Gregory Shtrasberg <Gregory.Shtrasberg@amd.com> Co-authored-by: Micah Williamson <micah.williamson@amd.com>
15 lines
693 B
C++
15 lines
693 B
C++
#pragma once
|
|
|
|
#include <torch/all.h>
|
|
|
|
void paged_attention(torch::Tensor& out, torch::Tensor& exp_sums,
|
|
torch::Tensor& max_logits, torch::Tensor& tmp_out,
|
|
torch::Tensor& query, torch::Tensor& key_cache,
|
|
torch::Tensor& value_cache, int64_t num_kv_heads,
|
|
double scale, torch::Tensor& block_tables,
|
|
torch::Tensor& context_lens, int64_t block_size,
|
|
int64_t max_context_len,
|
|
const std::optional<torch::Tensor>& alibi_slopes,
|
|
const std::string& kv_cache_dtype, torch::Tensor& k_scale,
|
|
torch::Tensor& v_scale);
|