zhangyiss/mlx
1
0
Fork 0
mirror of https://github.com/ml-explore/mlx.git synced 2025-12-16 01:49:05 +08:00
Code Issues Actions 1 Packages Projects Releases Wiki Activity

fix (#2701)

Browse Source
This commit is contained in:
Awni Hannun
2025-10-23 14:43:41 -07:00
parent 7b34dc366e
commit 6959732915
1 changed files with 13 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

13
mlx/backend/cuda/quantized/cuda_fp4.h
View File

@@ -1,5 +1,18 @@
#pragma once #pragma once
struct __nv_fp8_e8m0 {
__device__ __nv_fp8_e8m0(uint8_t x) : __x(x) {}
__device__ operator float() {
if (__x == 0xFF) {
return std::numeric_limits<float>::quiet_NaN();
}
return std::ldexp(1.0f, static_cast<int>(__x) - 127);
}
uint8_t __x{0};
};
struct __nv_fp4_e2m1 { struct __nv_fp4_e2m1 {
__device__ __nv_fp4_e2m1(float x) { __device__ __nv_fp4_e2m1(float x) {
if (std::isnan(x)) { if (std::isnan(x)) {