 Matvey Soloviev
|
904d2a8d6a
Q4_1 quantization (#193)
|
il y a 2 ans |
 Nebula
|
9b4a15b17d
Fix RMS norm in GGML (#191)
|
il y a 2 ans |
 hoangmit
|
6eac39ba95
Add RMS norm and use it (#187)
|
il y a 2 ans |
|
hoangmit
|
113e685d18
inline -> static inline for "bytesFromNibbles" (#161)
|
il y a 2 ans |
 Ronsor
|
47857e564c
Don't use vdotq_s32 if it's not available (#139)
|
il y a 2 ans |
 Thomas Klausner
|
41be0a3b3d
Add NetBSD support. (#90)
|
il y a 2 ans |
 Georgi Gerganov
|
84d9015c4a
Use vdotq_s32 to improve performance (#67)
|
il y a 2 ans |
|
Georgi Gerganov
|
c80e2a8f2a
Revert "10% performance boost on ARM"
|
il y a 2 ans |
|
Georgi Gerganov
|
54a0e66ea0
Check for vdotq_s32 availability
|
il y a 2 ans |
|
Georgi Gerganov
|
543c57e991
Ammend to previous commit - forgot to update non-QRDMX branch
|
il y a 2 ans |
|
Georgi Gerganov
|
113a9e83eb
10% performance boost on ARM
|
il y a 2 ans |
 Sebastián A
|
eb062bb012
Windows fixes (#31)
|
il y a 2 ans |
|
Georgi Gerganov
|
f1eaff4721
Add AVX2 support for x86 architectures thanks to @Const-me !
|
il y a 2 ans |
|
Georgi Gerganov
|
007a8f6f45
Support all LLaMA models + change Q4_0 quantization storage
|
il y a 2 ans |
|
Georgi Gerganov
|
26c0846629
Initial release
|
il y a 2 ans |
