ggerganov/llama.cpp

Pull Requests Commits
workflows: unbreak nix-build-aarch64, and split it out (#4915)
SomeoneSerge committed 1 year ago
Verified 6b48ed08
main : add parameter --no-display-prompt (#4541)
YannFollet committed 1 year ago
Verified 722d33f3
gguf : fix potential infinite for-loop (#4600)
texmex76 committed 1 year ago
Verified c30b1ef3
metal : refactor kernel loading code (#4794)
ggerganov committed 1 year ago
Verified b38b5e93
compare-llama-bench: tweak output format (#4910)
JohannesGaessler committed 1 year ago
Verified 7dc78764
server : fix deadlock that occurs in multi-prompt scenarios (#4905)
ziedbha committed 1 year ago
Verified 356327fe
server : fix crash with multimodal models without BOS token (#4904)
makomk committed 1 year ago
Verified ee8243ad
convert : update phi-2 to latest HF repo (#4903)
ggerganov committed 1 year ago
Verified 15ebe592
sync : ggml
ggerganov committed 1 year ago
Verified de473f5f
ggml : fix 32-bit ARM compat for IQ2_XS (whisper/1758)
ggerganov committed 1 year ago
Verified f2384612
backend_sched : fix assignments
slaren committed 1 year ago
Verified fa5c1fb4
examples : add pydantic models to GBNF grammar generator (#4883)
Maximilian-Winter committed 1 year ago
Verified 52ee4540
CUDA: faster q8_0 -> f16 dequantization (#4895)
JohannesGaessler committed 1 year ago
Verified 3fe81781
llama : ggml-backend integration (#4766)
slaren committed 1 year ago
Verified e7e4df03
llama : remove redundant assert for StableLM (#4901)
ggerganov committed 1 year ago
Verified 584d674b
export-lora : use LLAMA_FILE_MAGIC_GGLA (#4894)
danbev committed 1 year ago
Verified 930f907d
llama.swiftui : update models layout (#4826)
isaiahbjork committed 1 year ago
Verified e790eef2
gitignore : imatrix
ggerganov committed 1 year ago
Verified 5537d9d3
CUDA: fix softmax compile for old CUDA versions (#4862)
JohannesGaessler committed 1 year ago
Verified 1b280c9f
llama : fix typo "imp_embd" -> "inp_embd"
ggerganov committed 1 year ago
Verified 3cabe806
common : streamline the formatting of help (#4890)
howlger committed 1 year ago
Verified 4315a943
py : fix lint (#4889)
ggerganov committed 1 year ago
Verified 2d00741e
llama : fix llm_build_k_shift to use correct n_rot (#4889)
ggerganov committed 1 year ago
Verified f445c0e6
Importance Matrix calculation (#4861)
ikawrakow committed 1 year ago
Verified 326b418b
server : fix infill when prompt is empty (#4833)
ggerganov committed 1 year ago
Verified 1d118386
main : better name for variable n_print (#4874)
ggerganov committed 1 year ago
Verified 7edefbd7
main : disable token count by default (#4874)
ggerganov committed 1 year ago
Verified 3ca63b45
swift : track ggml release branch (#4867)
ggerganov committed 1 year ago
Verified b0377875
llama : restore intended k-quants mixes for MoE models (#4872)
ikawrakow committed 1 year ago
Verified 469e75d0
ggml : SOTA 2-bit quants (add IQ2_XS) (#4856)
ikawrakow committed 1 year ago
Verified 49662cbe
Terms Of Use Privacy Policy Refund Policy Impressum