File: quint4x2.h

package info (click to toggle)
pytorch 1.13.1%2Bdfsg-4
  • links: PTS, VCS
  • area: main
  • in suites: bookworm
  • size: 139,252 kB
  • sloc: cpp: 1,100,274; python: 706,454; ansic: 83,052; asm: 7,618; java: 3,273; sh: 2,841; javascript: 612; makefile: 323; xml: 269; ruby: 185; yacc: 144; objc: 68; lex: 44
file content (19 lines) | stat: -rw-r--r-- 366 bytes parent folder | download | duplicates (3) 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
 
#pragma once
#include <cstdint>

#include <c10/macros/Macros.h>

namespace c10 {

/**
 * quint4x2 is for un-signed 4 bit quantized Tensors that are packed to byte
 * boundary.
 */
struct alignas(1) quint4x2 {
  using underlying = uint8_t;
  uint8_t val_;
  quint4x2() = default;
  C10_HOST_DEVICE explicit quint4x2(uint8_t val) : val_(val) {}
};

} // namespace c10