llama.cpp/common
Georgi Gerganov c4f496648c
metal : fix kernel_norm (fixes Falcon on Metal) (#3057)
* metal : fix kernel_norm

ggml-ci

* metal : put warning in kernel_norm to not combine the loops

* metal : restore original F16 mat-vec multiplication

It works after the norm fixes

* common : don't do warm-up with more than n_batch tokens (close #3058)

ggml-ci

* metal : minor
2023-09-07 15:49:09 +03:00
..
CMakeLists.txt gguf : new file format with flexible meta data (beta) (#2398) 2023-08-21 23:07:43 +03:00
common.cpp metal : fix kernel_norm (fixes Falcon on Metal) (#3057) 2023-09-07 15:49:09 +03:00
common.h build : on Mac OS enable Metal by default (#2901) 2023-09-04 22:26:24 +03:00
console.cpp build : fix most gcc and clang warnings (#2861) 2023-09-01 16:34:50 +03:00
console.h gguf : new file format with flexible meta data (beta) (#2398) 2023-08-21 23:07:43 +03:00
grammar-parser.cpp gguf : new file format with flexible meta data (beta) (#2398) 2023-08-21 23:07:43 +03:00
grammar-parser.h gguf : new file format with flexible meta data (beta) (#2398) 2023-08-21 23:07:43 +03:00
log.h examples : replace fprintf to stdout with printf (#3017) 2023-09-05 15:10:27 -04:00