Mozilla
/

llm-compiler-13b-ftd-llamafile

Model card Files Files and versions Community

jartine commited on Jun 29

Commit

2cf4e67

•

1 Parent(s): 1290149

Update README.md

Files changed (1) hide show

README.md +2 -2

README.md CHANGED Viewed

@@ -150,10 +150,10 @@ appears to produce results consistent with full quality BF16/F16 weights
 except Q6 goes much faster, for both prefill and prediction, by virtue
 of consuming significantly less memory.
-### Q6\_0, F16, BF16
 The output for the first xdigit() assembly optimization example with
-Q6\_0, F16, or BF16 weights is:
 ```asm
         .globl  hiho

 except Q6 goes much faster, for both prefill and prediction, by virtue
 of consuming significantly less memory.
+### Q6\_K, F16, BF16
 The output for the first xdigit() assembly optimization example with
+Q6\_K, F16, or BF16 weights is:
 ```asm
         .globl  hiho