ilotoki_thu/alpaca_lora_4bit
2
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

fix bug

Browse Source
This commit is contained in:
John Smith
2023-03-23 23:37:39 +08:00
committed by GitHub
parent 94851cec68
commit 4906961bf1
1 changed files with 1 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
GPTQ-for-LLaMa/quant_cuda_kernel.cu
View File

@@ -631,7 +631,7 @@ __global__ void VecQuant4TransposeMatMulHalfKernel(
#if __CUDA_ARCH__ < 700 && __CUDA_ARCH__ > 600 #if __CUDA_ARCH__ < 700 && __CUDA_ARCH__ > 600
atomicAddHalf(&mul2[n_cols * height * 8 + n_rows], res); atomicAddHalf(&mul2[n_cols * height * 8 + n_rows], res);
#else #else
atomicAddHalf(&mul2[n_cols * height * 8 + n_rows], res); atomicAdd(&mul2[n_cols * height * 8 + n_rows], res);
#endif #endif
#endif #endif
} }