youngkingdom/vllm
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

[MISC] Replace c10::optional with std::optional (#11730)

Browse Source 
Signed-off-by: Lu Fang <lufang@fb.com>
This commit is contained in:
Lu Fang
2025-01-04 17:20:34 -08:00
committed by GitHub
parent 47831430cc
commit 4068f4b5b5
24 changed files with 136 additions and 136 deletions
Download Patch File Download Diff File Expand all files Collapse all files

10
csrc/cpu/quant.cpp
View File

@ -359,7 +359,7 @@ void int8_scaled_mm(torch::Tensor& c, // [M, OC], row-major
const torch::Tensor& b, // [IC, OC], column-major
const torch::Tensor& a_scales, // [1] or [M]
const torch::Tensor& b_scales, // [1] or [OC]
const c10::optional<torch::Tensor>& bias // [OC]
const std::optional<torch::Tensor>& bias // [OC]
) {
CPU_KERNEL_GUARD_IN(cutlass_scaled_mm)
// Checks for conformality
@ -442,8 +442,8 @@ void int8_scaled_mm_azp(torch::Tensor& c, // [M, OC], row-major
const torch::Tensor& a_scales, // [1] or [M]
const torch::Tensor& b_scales, // [1] or [OC]
const torch::Tensor& azp_adj, // [OC]
const c10::optional<torch::Tensor>& azp, // [1] or [M]
const c10::optional<torch::Tensor>& bias // [OC]
const std::optional<torch::Tensor>& azp, // [1] or [M]
const std::optional<torch::Tensor>& bias // [OC]
) {
CPU_KERNEL_GUARD_IN(cutlass_scaled_mm_azp)
// Checks for conformality
@ -561,7 +561,7 @@ void int8_scaled_mm_azp(torch::Tensor& c, // [M, OC], row-major
void static_scaled_int8_quant(torch::Tensor& out, // [..., hidden_size]
const torch::Tensor& input, // [..., hidden_size]
const torch::Tensor& scale,
c10::optional<torch::Tensor> const& azp) {
std::optional<torch::Tensor> const& azp) {
CPU_KERNEL_GUARD_IN(static_scaled_int8_quant)
TORCH_CHECK(input.is_contiguous());
TORCH_CHECK(out.is_contiguous());
@ -590,7 +590,7 @@ void dynamic_scaled_int8_quant(
torch::Tensor& out, // [..., hidden_size]
const torch::Tensor& input, // [..., hidden_size]
torch::Tensor& scale, // [..., 1]
c10::optional<torch::Tensor> const& azp) {
std::optional<torch::Tensor> const& azp) {
CPU_KERNEL_GUARD_IN(dynamic_scaled_int8_quant)
TORCH_CHECK(input.is_contiguous());
TORCH_CHECK(out.is_contiguous());