Skip to content

feat: add x_attention_v2 operator #46

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Closed
hot-water-fish wants to merge 1 commit into from
Closed

feat: add x_attention_v2 operator #46

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
29 changes: 29 additions & 0 deletions test/python_test/RegisterOps.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -89,6 +89,34 @@ at::Tensor x_attention_impl_npu(const at::Tensor& query,
return attnOut;
}

at::Tensor x_attention_v2_impl_npu(const at::Tensor& query,
const at::Tensor& key_cache,
const at::Tensor& value_cache,
const at::Tensor& unshared_key,
const at::Tensor& unshared_value,
const c10::optional<at::Tensor>& shared_block_tables,
const c10::optional<at::Tensor>& unshared_block_tables,
const at::Tensor& actual_shared_kvlen,
const at::Tensor& decode_step,
double scale_value = 0.0) {
at::Tensor attnOut = at::empty_like(query);

EXEC_NPU_CMD(aclnnXAttentionV2,
query,
key_cache,
value_cache,
unshared_key,
unshared_value,
unshared_block_tables,
actual_shared_kvlen,
decode_step,
shared_block_tables,
scale_value,
attnOut);

return attnOut;
}

std::tuple<at::Tensor, at::Tensor, at::Tensor> beam_search_impl_npu(
const at::Tensor& log_probs,
const at::Tensor& top_tokens,
Expand Down Expand Up @@ -1987,6 +2015,7 @@ PYBIND11_MODULE(TORCH_EXTENSION_NAME, m) {
m.def("select_unshared_kv", &select_unshared_kv_impl_npu, "select_unshared_kv");
m.def("cache_unshared_kv", &cache_unshared_kv_impl_npu, "cache_unshared_kv");
m.def("x_attention", &x_attention_impl_npu, "x_attention");
m.def("x_attention_v2", &x_attention_v2_impl_npu, "x_attention_v2");
m.def("beam_search", &beam_search_impl_npu, "beam_search");
m.def("beam_search_group", &beam_search_group_impl_npu, "beam_search_group");
m.def("beam_search_rec_final_select",
Expand Down
11 changes: 11 additions & 0 deletions test/python_test/custom_ops.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -15,6 +15,17 @@ def x_attention_npu(query, key_cache, value_cache, unshared_key, unshared_value,
return custom_ops_lib.x_attention(query, key_cache, value_cache, unshared_key, unshared_value,
shared_block_tables, unshared_block_tables, actual_shared_kvlen, decode_step, scale_value)


# x_attention_v2
def x_attention_v2_npu(query, key_cache, value_cache, unshared_key, unshared_value, actual_shared_kvlen, decode_step,
shared_block_tables = None,
unshared_block_tables = None,
scale_value = None):
if scale_value is None:
scale_value = 0.0
return custom_ops_lib.x_attention_v2(query, key_cache, value_cache, unshared_key, unshared_value,
shared_block_tables, unshared_block_tables, actual_shared_kvlen, decode_step, scale_value)

# reshape cache kv
def cache_unshared_kv_npu(x_key_block, x_value_block, curr_key, curr_value, block_table, decode_step):
return custom_ops_lib.cache_unshared_kv(x_key_block, x_value_block, curr_key, curr_value, block_table, decode_step)
Expand Down
Loading