Skip to content

Fix qnn in android demo app #15258

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

Draft
wants to merge 2 commits into
base: main
Choose a base branch
from
+33 −7
Draft

Fix qnn in android demo app #15258

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
247aecc
Fix qnn in android demo app
cccclai Oct 19, 2025
50a7cac
more fix
cccclai Oct 20, 2025
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
40 changes: 33 additions & 7 deletions extension/android/jni/jni_layer_llama.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -168,13 +168,13 @@ class ExecuTorchLlmJni : public facebook::jni::HybridClass<ExecuTorchLlmJni> {
}
#endif

std::vector<std::string> data_files_vector;
model_type_category_ = model_type_category;
if (model_type_category == MODEL_TYPE_CATEGORY_MULTIMODAL) {
multi_modal_runner_ = llm::create_multimodal_runner(
model_path->toStdString().c_str(),
llm::load_tokenizer(tokenizer_path->toStdString()));
} else if (model_type_category == MODEL_TYPE_CATEGORY_LLM) {
std::vector<std::string> data_files_vector;
if (data_files != nullptr) {
// Convert Java List<String> to C++ std::vector<string>
auto list_class = facebook::jni::findClassStatic("java/util/List");
Expand Down Expand Up @@ -202,12 +202,38 @@ class ExecuTorchLlmJni : public facebook::jni::HybridClass<ExecuTorchLlmJni> {
data_files_vector,
executorch::extension::Module::LoadMode::MmapUseMlockIgnoreErrors);
std::string decoder_model = "llama3"; // use llama3 for now
Copy link
Contributor Author

@cccclai cccclai Oct 20, 2025 •
edited
Loading

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Replace the model name here accordingly, available path is

executorch/examples/qualcomm/oss_scripts/llama/runner/runner.cpp

Line 121 in f80916e

if (decoder_model_version == "llama2") {

runner_ = std::make_unique<example::Runner<uint16_t>>( // QNN runner
std::move(module),
decoder_model.c_str(),
model_path->toStdString().c_str(),
tokenizer_path->toStdString().c_str(),
"");
example::KvBitWidth kv_bitwidth = example::KvBitWidth::kWidth8;
if (module->method_names()->count("get_kv_io_bit_width") > 0) {
kv_bitwidth = static_cast<example::KvBitWidth>(
module->get("get_kv_io_bit_width").get().toScalar().to<int64_t>());
}

if (kv_bitwidth == example::KvBitWidth::kWidth8) {
runner_ = std::make_unique<example::Runner<uint8_t>>(
std::move(module),
decoder_model.c_str(), // const std::string&
model_path->toStdString().c_str(), // const std::string&
tokenizer_path->toStdString().c_str(), // const std::string&
"", // performance_output_path
"", // dump_logits_path
temperature_ // temperature
);
} else if (kv_bitwidth == example::KvBitWidth::kWidth16) {
runner_ = std::make_unique<example::Runner<uint16_t>>(
std::move(module),
decoder_model.c_str(), // const std::string&
model_path->toStdString().c_str(), // const std::string&
tokenizer_path->toStdString().c_str(), // const std::string&
"", // performance_output_path
"", // dump_logits_path
temperature_ // temperature
);
} else {
ET_CHECK_MSG(
false,
"Unsupported kv bitwidth: %ld",
static_cast<int64_t>(kv_bitwidth));
}
model_type_category_ = MODEL_TYPE_CATEGORY_LLM;
#endif
#if defined(EXECUTORCH_BUILD_MEDIATEK)
Expand Down
Loading