Skip to content
GitLab
Explore
Sign in
Tags
Tags give the ability to mark specific points in history as being important
b2231
ba2135cc
·
gemma : allow offloading the output tensor (#5646)
·
Feb 21, 2024
b2230
89febfed
·
examples : do not assume BOS when shifting context (#5622)
·
Feb 21, 2024
b2228
1ecea255
·
server: health: fix race condition on slots data using tasks queue (#5634)
·
Feb 21, 2024
b2226
eccd7a26
·
sync : ggml (#5633)
·
Feb 21, 2024
b2223
580111d4
·
llama : add `gemma` model (#5631)
·
Feb 21, 2024
b2222
88c46cbd
·
[SYCL] conext add name (#5624)
·
Feb 21, 2024
b2221
a14679cc
·
IQ4_NL: 4-bit non-linear quants with blocks of 32 (#5590)
·
Feb 21, 2024
b2220
6560bed3
·
server : support llava 1.6 (#5553)
·
Feb 20, 2024
b2219
06bf2cf8
·
make : fix debug build with CUDA (#5616)
·
Feb 20, 2024
b2217
9c405c9f
·
Server: use llama_chat_apply_template (#5593)
·
Feb 20, 2024
b2215
8dbbd757
·
metal : add build system support for embedded metal library (#5604)
·
Feb 20, 2024
b2214
c0a8c6db
·
server : health endpoint configurable failure on no slot (#5594)
·
Feb 20, 2024
b2213
b9111bd2
·
Update ggml_sycl_op_mul_mat_vec_q (#5502)
·
Feb 20, 2024
b2212
633782b8
·
nix: now that we can do so, allow MacOS to build Vulkan binaries
·
Feb 19, 2024
b2205
40c3a6c1
·
cuda : ignore peer access already enabled errors (#5597)
·
Feb 19, 2024
b2204
f24ed14e
·
make : pass CPPFLAGS directly to nvcc, not via -Xcompiler (#5598)
·
Feb 19, 2024
b2202
1387cf60
·
llava : remove extra cont (#5587)
·
Feb 19, 2024
b2201
6fd41379
·
llava : replace ggml_cpy with ggml_cont
·
Feb 19, 2024
b2197
d0e3ce51
·
ci : enable -Werror for CUDA builds (#5579)
·
Feb 19, 2024
b2196
68a6b98b
·
make : fix CUDA build (#5580)
·
Feb 19, 2024
Prev
1
…
22
23
24
25
26
27
28
29
30
…
98
Next