Skip to content
GitLab
Explore
Sign in
Tags
Tags give the ability to mark specific points in history as being important
b2032
b2b9f025
·
format license text, restore apache license by legal suggestion (#5233)
·
Jan 31, 2024
b2031
dabcc5b4
·
ggml : limit n_threads to the max n_tasks (#5238)
·
Jan 31, 2024
b2030
f8e9140c
·
Vulkan Fixes (#5223)
·
Jan 31, 2024
b2029
d62520eb
·
Fix typos of IQ2_XXS and IQ3_XXS in llama.cpp (#5231)
·
Jan 30, 2024
b2028
01684139
·
support SYCL backend windows build (#5208)
·
Jan 31, 2024
b2027
e8dc55d0
·
kompute : llama-bench support and ggml_cpu_has_kompute() (#5226)
·
Jan 30, 2024
b2026
e0085fdf
·
Revert "server : change deps.sh xxd files to string literals (#5221)"
·
Jan 30, 2024
b2022
8f8ddfcf
·
sync : ggml (#0)
·
Jan 30, 2024
b2016
8e14e3dd
·
Faster AVX2 dot product for IQ2_XS (#5187)
·
Jan 30, 2024
b2015
f4d7e549
·
SOTA 3-bit quants (#5196)
·
Jan 30, 2024
b2014
2256f36b
·
Vulkan Windows APU Memory Handling (#5199)
·
Jan 30, 2024
b2013
7359016c
·
quantize : fix typo (#5211)
·
Jan 30, 2024
b2012
81341699
·
main : allow empty --prompt-cache file (#5176)
·
Jan 30, 2024
b2008
ceebbb5b
·
ggml alloc: Fix for null dereference on alloc failure (#5200)
·
Jan 29, 2024
b2007
6daa69ee
·
kompute : fix fallback to CPU (#5201)
·
Jan 29, 2024
b2006
fbf1ddec
·
Nomic Vulkan backend (#4456)
·
Jan 29, 2024
b2005
2aed77eb
·
fix typo "RLIMIT_MLOCK" (#5175)
·
Jan 29, 2024
b2004
c82d18e8
·
server : embeddings compatibility for OpenAI (#5190)
·
Jan 29, 2024
b2001
fbe7dfa5
·
ggml : add max buffer sizes to opencl and metal backends (#5181)
·
Jan 29, 2024
b2000
172ac826
·
cmake : fix Vulkan build (#5182)
·
Jan 29, 2024
Prev
1
…
29
30
31
32
33
34
35
36
37
…
99
Next