 Georgi Gerganov
|
f8e58135cf
metal : move dequantize templates to beginning of MSL source (#0)
|
il y a 1 an |
 leo-pony
|
329ed914c9
CANN: adjust backend registry refactor. (#10158)
|
il y a 1 an |
|
Georgi Gerganov
|
ce027adfb3
sync : ggml
|
il y a 1 an |
 Yuri Khrustalev
|
284e5b0275
cmake : make it possible linking ggml as external lib (ggml/1003)
|
il y a 1 an |
 Plamen Minev
|
e2292aaa17
metal : fix minor string leaks (ggml/1004)
|
il y a 1 an |
 Diego Devesa
|
9f40989351
ggml : move CPU backend to a separate file (#10144)
|
il y a 1 an |
|
Georgi Gerganov
|
08828a6d7d
metal : minor fixup in FA kernel (#10143)
|
il y a 1 an |
|
Georgi Gerganov
|
1839f69130
flake.lock: Update (#10146)
|
il y a 1 an |
 Christian Köhnenkamp
|
9830b6923b
Add apple arm to presets (#10134)
|
il y a 1 an |
 sasha0552
|
42cadc74bd
server : fix slot selection by lru (#10126)
|
il y a 1 an |
|
Georgi Gerganov
|
45950415ed
server : fix endpoint checks (#10135)
|
il y a 1 an |
|
Georgi Gerganov
|
1926d6e39d
llama : adjust default context size + print warnings (#10136)
|
il y a 1 an |
|
Diego Devesa
|
b634f8a26f
simple-chat : only add bos on first prompt (#10129)
|
il y a 1 an |
 Xuan Son Nguyen
|
7554aa4655
convert-lora : make `--base` optional (#10110)
|
il y a 1 an |
|
Diego Devesa
|
a6744e43e8
llama : add simple-chat example (#10124)
|
il y a 1 an |
|
Diego Devesa
|
e991e3127f
llama : use smart pointers for ggml resources (#10117)
|
il y a 1 an |
 Shupei Fan
|
418f5eef26
vulkan : improve ggml_vk_create_buffer error handling (#9898)
|
il y a 1 an |
|
Georgi Gerganov
|
ba6f62eb79
readme : update hot topics
|
il y a 1 an |
|
sasha0552
|
d865d1478c
server : fix smart selection of available slot (#10120)
|
il y a 1 an |
|
Georgi Gerganov
|
1804adb0cf
ggml : remove ggml_scratch (#10121)
|
il y a 1 an |
|
Georgi Gerganov
|
815fe72adc
sync : ggml
|
il y a 1 an |
|
Georgi Gerganov
|
f221d56220
ggml : alloc ggml_contexts on the heap (whisper/2525)
|
il y a 1 an |
 Zhenwei Jin
|
e597e50794
build: fix build error in Windows env with OneAPI setup (#10107)
|
il y a 1 an |
|
Diego Devesa
|
85679d37f3
llama : improve output buffer type selection (#10098)
|
il y a 1 an |
|
Diego Devesa
|
1e9f94994e
quantize : fix --keep-split (#10114)
|
il y a 1 an |
|
Diego Devesa
|
c02e5ab2a6
llama : fix buffer checks for mamba and rwk (#10111)
|
il y a 1 an |
|
Zhenwei Jin
|
ab3d71f97f
loader: refactor tensor weights storage (#9935)
|
il y a 1 an |
 Kevin Gibbons
|
0a683e8088
server : include scheme when printing URL (#10106)
|
il y a 1 an |
|
Diego Devesa
|
dea5e86051
ggml : check tensor name lengths in gguf files (#10100)
|
il y a 1 an |
 Sergio López
|
1329c0a75e
kompute: add mul_mat_q4_k shader (#10097)
|
il y a 1 an |
