aparler/c-llm
1
0
Fork 0
mirror of https://github.com/karpathy/llm.c.git synced 2026-07-28 20:35:09 -04:00
Code Issues Projects Releases Packages Wiki Activity Actions

remove BF16 default from classified_fused before PR

Browse source
This commit is contained in:
ademeure 2024-05-07 00:57:01 +01:00
parent 1ea7f9bf25
commit e1f89b3043
1 changed files with 0 additions and 4 deletions
Download patch file Download diff file Expand all files Collapse all files

4
dev/cuda/classifier_fused.cu
View file

@ -23,10 +23,6 @@ nvcc -O3 --use_fast_math -lcublas -lcublasLt classifier_fused.cu -o classifier_f
// todo - this file does not properly support anything but FP32
// kernel 5 can be run in fp16/bf16 to test performance, but the outputs will be wrong
#undef ENABLE_BF16
#undef ENABLE_FP16
#define ENABLE_BF16
#if defined(ENABLE_BF16)
typedef __nv_bfloat16 floatX;
#elif defined(ENABLE_FP16)