Donghyeon Jeong

[Tensor] Refactorize Tensor Class to TensorV2 @open sesame 03/26 12:26

5

This pull request introduces a major refactorization of the Tensor class in our codebase by introducing the new TensorV2 class. The previous Tensor class is removed, and all instances where...

[Wait for #2500][Tensor] CharTensor class with signed 8-bit integer

3

In this PR, a new type of tensor, the CharTensor class, is designed explicitly for handling signed 8-bit integer data types that have already undergone quantization. This new tensor class...

Channel Last Tensor save/read fails occasionally

1

`nntrainer_Tensor.save_read_01_nhwc_p` unit test failure has occurred on the local machine as follows. ``` [ RUN ] nntrainer_Tensor.save_read_01_nhwc_p ../test/unittest/unittest_nntrainer_tensor_nhwc.cpp:3742: Failure Expected equality of these values: target Which is: data addr: 0x55dac0b76bf0...

CacheLoaderTest fails occasionally in GitHub Action CI

1

In GitHub Action CI, `CacheLoaderTest.load_async_03_p` in the memory unit test fails occasionally. ``` 25/40 unittest_memory FAIL 1.59s exit status 1 >>> MALLOC_PERTURB_=137 /home/runner/work/nntrainer/nntrainer/build/test/unittest/memory/unittest_memory --gtest_output=xml:/home/runner/work/nntrainer/nntrainer/build/unittest_memory.xml ... [----------] Global test environment tear-down...

Half-Precision SGEMV() and SGEMM() not support CblasColMajor

1

Currently, the implementation of half-precision SGEMV() and SGEMM() does not support CblasColMajor for the order parameter (CBLAS_ORDER). ``` /** CBLAS_ORDER order parameter is not used. */ static void sgemv_FP16(CBLAS_ORDER order,...

[PR #2897] Add arm-native build and test on Ubuntu @open sesame 12/19 09:42

9

https://github.com/nnstreamer/nntrainer/pull/2897#issue-2805918149

## Dependency of the PR None ## Commits to be reviewed in this PR Lazy Buffer Initialization and Refactor OpenCL Buffer Manager This PR implemented lazy buffer initialization for OpenCL...

Refactor Unit Tests for OpenCL Kernels

5

This pull request refactors the unit tests for OpenCL kernels. The changes involve splitting the test files into three categories based on functionality. One for INT4-related tests, another for QK_K...