thek0tyara/llama-cpp-turboquant
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions 4

Updated to use all uint32_t to avoid ambiguity on passing strings

Browse source
This commit is contained in:
Alejandro Saucedo 2021-02-20 18:09:02 +00:00
parent 5bc3ac9c06
commit 56d9a3a933
18 changed files with 65 additions and 68 deletions
Download patch file Download diff file Expand all files Collapse all files

4
examples/android/android-simple/app/src/main/cpp/KomputeModelML.cpp
View file

@ -54,11 +54,11 @@ void KomputeModelML::train(std::vector<float> yData, std::vector<float> xIData,
#ifdef KOMPUTE_ANDROID_SHADER_FROM_STRING
// Newer versions of Android are able to use shaderc to read raw string
sq->record<kp::OpAlgoBase>(
params, std::vector<char>(LR_SHADER.begin(), LR_SHADER.end()));
params, std::vector<uint32_t>(LR_SHADER.begin(), LR_SHADER.end()));
#else
// Older versions of Android require the SPIRV binary directly
sq->record<kp::OpAlgoBase>(
params, std::vector<char>(
params, std::vector<uint32_t>(
kp::shader_data::shaders_glsl_logisticregression_comp_spv,
kp::shader_data::shaders_glsl_logisticregression_comp_spv
+ kp::shader_data::shaders_glsl_logisticregression_comp_spv_len