[Tensor] Int4QTensor with quantized 4-bit integer data type

This pull request presents the class, a powerful solution for efficiently storing quantized 4-bit integer data. By packing each 4-bit integer into an 8-bit memory space, we utilize memory resources effectively—where the first four bits hold the first 4-bit value and the last four bits hold the second. 1. Build test: [X]Passed [ ]Failed [ ]Skipped 2. Run test: [X]Passed [ ]Failed [ ]Skipped Signed-off-by: Donghyeon Jeong <[email protected]>
nnstreamer · Jan 23, 2025 · d60db7a · d60db7a
1 parent 091c496
commit d60db7a
Show file tree

Hide file tree

Showing 7 changed files with 1,050 additions and 134 deletions.
diff --git a/debian/nntrainer-dev.install b/debian/nntrainer-dev.install
@@ -10,6 +10,7 @@
 /usr/include/nntrainer/memory_data.h
 /usr/include/nntrainer/tensor.h
 /usr/include/nntrainer/tensor_base.h
+/usr/include/nntrainer/int4_tensor.h
 /usr/include/nntrainer/char_tensor.h
 /usr/include/nntrainer/short_tensor.h
 /usr/include/nntrainer/uint_tensor.h