Skip to content
GitLab
Explore
Sign in
Tags
Tags give the ability to mark specific points in history as being important
This project is mirrored from
https://github.com/ggerganov/llama.cpp
. Pull mirroring updated
Sep 16, 2024
.
b3772
23e0d70b
·
ggml : move common CPU backend impl to new header (#9509)
·
Sep 16, 2024
b3771
acb2c32c
·
llama : rename n_embed to n_embd in rwkv6_time_mix (#9504)
·
Sep 16, 2024
b3770
a6a3a5c5
·
ggml : link MATH_LIBRARY not by its full path (#9339)
·
Sep 16, 2024
b3767
5c3d0f18
·
ggml : IQ4_NL sgemm + Q4_0 AVX optimization (#9422)
·
Sep 16, 2024
b3766
0aadac10
·
llama : support OLMoE (#9462)
·
Sep 16, 2024
b3765
95ca8516
·
llama : support MiniCPM3 (#9322)
·
Sep 16, 2024
b3764
441b72b9
·
main : option to disable context shift (#9484)
·
Sep 16, 2024
b3763
c4965a64
·
metal : handle zero-sized allocs (#9466)
·
Sep 16, 2024
b3761
6262d13e
·
common : reimplement logging (#9418)
·
Sep 15, 2024
b3760
e6deac31
·
gguf-split : add basic checks (#9499)
·
Sep 15, 2024
b3759
6988da94
·
cmake : correct order of sycl flags (#9497)
·
Sep 15, 2024
b3756
7596487b
·
cmake : try to fix sycl+intel build (#9487)
·
Sep 15, 2024
b3755
822b6322
·
ggml : ggml_type_name return "NONE" for invalid values (#9458)
·
Sep 14, 2024
b3754
dcdcee3a
·
server: add data: [DONE] to /chat/completions stream response (#9459)
·
Sep 14, 2024
b3753
1f4111e5
·
cmake : use list(APPEND ...) instead of set() + dedup linker (#9463)
·
Sep 14, 2024
b3752
befaf119
·
llama : make cell_id const in inp_s_mask block (#9470)
·
Sep 14, 2024
b3751
feff4aa8
·
server : add loading html page while model is loading (#9468)
·
Sep 13, 2024
b3750
0abc6a2c
·
llama : llama_perf + option to disable timings during decode (#9355)
·
Sep 13, 2024
b3749
bd35cb0a
·
feat: remove a sampler from a chain (#9445)
·
Sep 13, 2024
b3747
e6b7801b
·
cann: Add host buffer type for Ascend NPU (#9406)
·
Sep 12, 2024
Prev
1
2
3
4
5
…
122
Next