saymrwulf/onnxruntime
1
0
Fork 0
mirror of https://github.com/saymrwulf/onnxruntime.git synced 2026-07-07 17:15:29 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

Fix moe tests to run on supported arch (#20872)

Browse source 
### Description
<!-- Describe your changes. -->

https://github.com/microsoft/onnxruntime/issues/20788

Will do sm70 validation separately. 

### Motivation and Context
<!-- - Why is this change required? What problem does it solve?
- If it fixes an open issue, please link to the issue here. -->
This commit is contained in:
Ye Wang 2024-05-30 13:26:38 -07:00 committed by GitHub
parent 65ef270e06
commit 2200a0b3dd
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
1 changed files with 7 additions and 3 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

10
onnxruntime/test/contrib_ops/moe_test.cc
View file

@ -16,9 +16,10 @@ static void RunMoETest(const std::vector<float>& input, const std::vector<float>
const std::vector<float>& fc2_experts_bias, const std::vector<float>& output_data, int num_rows,
int num_experts, int hidden_size, int inter_size, std::string activation_type,
int normalize_routing_weights = 0, int top_k = 1, bool use_float16 = false) {
int min_cuda_architecture = use_float16 ? 700 : 0;
constexpr int min_cuda_arch = 700;
constexpr int max_cuda_arch = 900;
bool enable_cuda = HasCudaEnvironment(min_cuda_architecture);
bool enable_cuda = HasCudaEnvironment(min_cuda_arch) && !NeedSkipIfCudaArchGreaterEqualThan(max_cuda_arch);
if (enable_cuda) {
OpTester tester("MoE", 1, onnxruntime::kMSDomain);
tester.AddAttribute<int64_t>("k", static_cast<int64_t>(top_k));
@ -91,7 +92,10 @@ static void RunQMoETest(const std::vector<float>& input, const std::vector<float
const std::vector<float>& fc2_scales, const std::vector<float>& fc3_scales,
const std::vector<float>& output_data, int num_rows, int num_experts, int hidden_size,
int inter_size, std::string activation_type, int normalize_routing_weights = 0, int top_k = 1) {
bool enable_cuda = HasCudaEnvironment(700);
constexpr int min_cuda_arch = 700;
constexpr int max_cuda_arch = 900;
bool enable_cuda = HasCudaEnvironment(min_cuda_arch) && !NeedSkipIfCudaArchGreaterEqualThan(max_cuda_arch);
if (enable_cuda) {
OpTester tester("QMoE", 1, onnxruntime::kMSDomain);
tester.AddAttribute<int64_t>("k", static_cast<int64_t>(top_k));