deepseek-ai / DeepGEMM

DeepGEMM: clean and efficient FP8 GEMM kernels with fine-grained scaling

Repository from Github https://github.comdeepseek-ai/DeepGEMMRepository from Github https://github.comdeepseek-ai/DeepGEMM

how to change input data type as float

youzjuer opened this issue · comments

I just wanna change A and B from fp8 to fp32,but I found it will occur an error in tma desc as:

Image
why

DeepGEMM does not support fp32 input.