CS348Project/llama.cpp
5
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2025-10-27 08:21:30 +00:00
Code Issues Packages Projects Releases Wiki Activity

ggml-quants : handle zero amax for MXFP4

Browse Source
This commit is contained in:
Francis Couture-Harpin
2025-08-06 16:26:25 -04:00
parent 141cab137d
commit 2763dc8b53
1 changed files with 1 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
ggml/src/ggml-quants.c
View File

@@ -288,7 +288,7 @@ void quantize_row_mxfp4_ref(const float * GGML_RESTRICT x, block_mxfp4 * GGML_RE
} }
} }
const uint8_t e = (uint8_t) (floorf(log2f(amax)) - 2 + 127); const uint8_t e = amax > 0.0f ? (uint8_t) (floorf(log2f(amax)) - 2 + 127) : 0;
const float d = GGML_E8M0_TO_FP32_HALF(e); const float d = GGML_E8M0_TO_FP32_HALF(e);